From ulf.wiger@REDACTED  Wed Dec  1 09:57:31 2004
From: ulf.wiger@REDACTED (Ulf Wiger (AL/EAB))
Date: Wed, 1 Dec 2004 09:57:31 +0100 
Subject: emulate target node names on a workstation
Message-ID: <37FB7AA6F5F9814FB634A7BF4C35A6F54029B8@ESEALNT442.al.sw.ericsson.se>


I just thought I'd mention that I sent a set of patches to OTP 
(based on OTP R10B_0) that make it possible to run several 
Erlang nodes on the same machine using the convention 

  'erl -sname <name>@<host>'. 

This wasn't possible before, unless perhaps if the <name> part 
happened to be unique for all nodes involved. My patch makes 
epmd aware of the whole node name -- not just the first half.

I've attached an archive with a set of diffs (for epmd_srv.c, 
epmd.h, epmd_int.h, net_kernel.erl, erl_epmd.[eh]rl, 
inet_tcp_dist.erl). If anyone wants copies of the full source, 
please email me.

Note that I've only dealt with the Erlang parts -- not C nodes 
et al. Here's a chance to contribute. ;-)

AFAIK, the new epmd is backward compatible, but the <host> part 
known to epmd will default to the original host name (this is 
done since older erlang nodes don't pass the <host> part when 
they register with epmd.)

Using this feature also requires a custom inetrc file. This 
is all described in the ERTS User Guide.

(At AXD 301, we've used a mapping layer on top of the node names 
to allow for them changing depending on environment. While there 
may be some additional point to doing this, it has occured to me 
that the Erlang node name is a logical abstraction in itself, and 
should be sufficient in most cases. Using the node names directly 
would cut down on the amount of glue code needed in the system. 
Note, though, that we're not pushing this as a priority item. OTP 
may choose to include it, or not. We don't care that much. I still 
think it's a nice idea, though. Finally - fair warning: this was 
my first "significant" c hack in 12-13 years -- absolutely no 
warranty! :-)

/Uffe
 
 <<epmd_diffs.tgz>> 
-------------- next part --------------
A non-text attachment was scrubbed...
Name: epmd_diffs.tgz
Type: application/octet-stream
Size: 2888 bytes
Desc: not available
URL: <http://erlang.org/pipermail/erlang-questions/attachments/20041201/03be8fa5/attachment.obj>

From bjorn@REDACTED  Wed Dec  1 10:04:09 2004
From: bjorn@REDACTED (Bjorn Gustavsson)
Date: 01 Dec 2004 10:04:09 +0100
Subject: parallel LC
In-Reply-To: <37FB7AA6F5F9814FB634A7BF4C35A6F54029B5@ESEALNT442.al.sw.ericsson.se>
References: <37FB7AA6F5F9814FB634A7BF4C35A6F54029B5@ESEALNT442.al.sw.ericsson.se>
Message-ID: <o1fz2q1jau.fsf@erix.ericsson.se>

"Ulf Wiger (AL/EAB)" <ulf.wiger@REDACTED> writes:
[...]
> which happens to work just like (the undocumented) lists:zip([1,2,3,4], [a,b,c,d]) in Erlang.

lists:zip/2 and friends are documented in OTP-R10B-1.

/Bjorn

-- 
Bj?rn Gustavsson, Erlang/OTP, Ericsson AB


From ulf.wiger@REDACTED  Wed Dec  1 10:22:37 2004
From: ulf.wiger@REDACTED (Ulf Wiger (AL/EAB))
Date: Wed, 1 Dec 2004 10:22:37 +0100 
Subject: parallel LC
Message-ID: <37FB7AA6F5F9814FB634A7BF4C35A6F54029B9@ESEALNT442.al.sw.ericsson.se>


> lists:zip/2 and friends are documented in OTP-R10B-1.
> 
> /Bjorn

OK, great... but then the "Erlang/OTP R10B documentation"
link at http://www.erlang.org/doc.html doesn't point to 
OTP R10B_1.

/Uffe


From ulf.wiger@REDACTED  Wed Dec  1 11:11:18 2004
From: ulf.wiger@REDACTED (Ulf Wiger (AL/EAB))
Date: Wed, 1 Dec 2004 11:11:18 +0100 
Subject: risky_shell
Message-ID: <37FB7AA6F5F9814FB634A7BF4C35A6F54029BA@ESEALNT442.al.sw.ericsson.se>


Am I just blind, or is the command-line flag -risky_shell undocumented?

The stdlib release notes for OTP R10B_1 mention that there is a possibility to start the Erlang shell in parallel with the rest of the system, but there's no explanation of how this is done (fortunately, there's always the source code... :-) 

Perhaps at least the release notes could mention the name of the flag that makes this possible, if you don't want the 'erl' documentation in ERTS to mention it?

The idea behind re-introducing this "bug" was that it is sometimes very important to be able to trace on the application start sequence. One convenient way of doing this is to simply enter calls to e.g. dbg in the shell.

/Uffe


From bjarne@REDACTED  Wed Dec  1 11:59:34 2004
From: bjarne@REDACTED (=?Windows-1252?Q?Bjarne_D=E4cker?=)
Date: Wed, 1 Dec 2004 11:59:34 +0100
Subject: Joe's promotion
Message-ID: <001701c4d794$e4f660a0$b10e69d4@segeltorp>

Hello

A year ago, Joe successfully defended his
thesis "Making reliable distributed systems
in the presence of software errors" and on
December 19 this year, he received his
diploma (together with > 200 other new
doctors) at a formal occassion in the town
hall in Stockholm.

Some photos from this happening can
be view at

http://www.cs-lab.org/joes_promotion/

Bjarne



-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://erlang.org/pipermail/erlang-questions/attachments/20041201/856c6ef3/attachment.htm>

From richardc@REDACTED  Wed Dec  1 12:57:16 2004
From: richardc@REDACTED (Richard Carlsson)
Date: Wed, 1 Dec 2004 12:57:16 +0100 (MET)
Subject: New test release of EDoc - feedback wanted
In-Reply-To: <41AC8056.70303@csd.uu.se>
References: <37FB7AA6F5F9814FB634A7BF4C35A6F54029B4@ESEALNT442.al.sw.ericsson.se>
 <41AC8056.70303@csd.uu.se>
Message-ID: <Pine.SOL.4.53.0412011249410.6748@meg.it.uu.se>


On Tue, 30 Nov 2004, Richard Carlsson wrote:

> Ulf Wiger (AL/EAB) wrote:
> >
> > How do the syntax_tools and xmerl archives in that
> > same location compare with the ones in OTP R10B?
>
> The R10B version did not have the new verbatim-quoting
> (at least not officially and in the current form), nor
> the improved scanning, nor proper try-expression support.

Wow - I obviously did not read the question properly. Sorry.
At least I answered the part about the syntax tools (scanning
and try-expressions). The quoting stuff is in EDoc.

The xmerl tarball is just there for completeness. It should
not differ from the OTP R10B version.

	/yxskaft


Richard Carlsson (richardc@REDACTED)   (This space intentionally left blank.)
E-mail: Richard.Carlsson@REDACTED	WWW: http://user.it.uu.se/~richardc/
 "Having users is like optimization: the wise course is to delay it."
   -- Paul Graham


From matthias@REDACTED  Wed Dec  1 13:01:15 2004
From: matthias@REDACTED (Matthias Lang)
Date: Wed, 1 Dec 2004 13:01:15 +0100
Subject: Joe's promotion
In-Reply-To: <001701c4d794$e4f660a0$b10e69d4@segeltorp>
References: <001701c4d794$e4f660a0$b10e69d4@segeltorp>
Message-ID: <16813.45707.741223.507816@antilipe.corelatus.se>

Bjarne D?cker writes:

 > on December 19 this year, he received his diploma (together 
 > with > 200 other new doctors) at a formal occassion in the town
 > hall in Stockholm.

And on the 26th. January 2005, Joe invented the time machine which
allowed him to accomplish the above ;-)

(Maybe it was the 19th. of November)

Matthias


From tobbe@REDACTED  Wed Dec  1 13:15:34 2004
From: tobbe@REDACTED (Torbjorn Tornkvist)
Date: Wed, 01 Dec 2004 13:15:34 +0100
Subject: Joe's promotion
In-Reply-To: <001701c4d794$e4f660a0$b10e69d4@segeltorp>
References: <001701c4d794$e4f660a0$b10e69d4@segeltorp>
Message-ID: <41ADB5E6.4040507@nortelnetworks.com>


Nice photos!

I wonder what Joe told Jan Uddenfeldt...

/Tobbe


Bjarne D?cker wrote:

> Hello
>  
> A year ago, Joe successfully defended his
> thesis "Making reliable distributed systems
> in the presence of software errors" and on
> December 19 this year, he received his
> diploma (together with > 200 other new
> doctors) at a formal occassion in the town
> hall in Stockholm.
>  
> Some photos from this happening can
> be view at
>  
> http://www.cs-lab.org/joes_promotion/
>  
> Bjarne
>  
>  
>  




From erlang@REDACTED  Wed Dec  1 13:55:33 2004
From: erlang@REDACTED (Peter-Henry Mander)
Date: Wed, 1 Dec 2004 12:55:33 +0000
Subject: Joe's promotion
In-Reply-To: <16813.45707.741223.507816@antilipe.corelatus.se>
References: <001701c4d794$e4f660a0$b10e69d4@segeltorp>
	<16813.45707.741223.507816@antilipe.corelatus.se>
Message-ID: <20041201125533.3add6ba3.erlang@manderp.freeserve.co.uk>


Hang on, it {must have been,will be} Bjarne who invent{ed,s} the time machine...

How does English grammar express this tense?

Pete.


On Wed, 1 Dec 2004 13:01:15 +0100
Matthias Lang <matthias@REDACTED> wrote:

> 
>  > on December 19 this year, he received his diploma (together 
>  > with > 200 other new doctors) at a formal occassion in the town
>  > hall in Stockholm.
> 
> And on the 26th. January 2005, Joe invented the time machine which
> allowed him to accomplish the above ;-)
> 
> (Maybe it was the 19th. of November)
> 
> Matthias


-- 
"The Tao of Programming
 flows far away 
 and returns 
 on the wind of morning."



From joe@REDACTED  Wed Dec  1 15:01:13 2004
From: joe@REDACTED (Joe Armstrong)
Date: Wed, 1 Dec 2004 15:01:13 +0100 (CET)
Subject: Joe's promotion
In-Reply-To: <41ADB5E6.4040507@nortelnetworks.com>
References: <001701c4d794$e4f660a0$b10e69d4@segeltorp> <41ADB5E6.4040507@nortelnetworks.com>
Message-ID: <Pine.LNX.4.61.0412011500110.31096@enfield.sics.se>



> Nice photos!
>
> I wonder what Joe told Jan Uddenfeldt...
>

    That Erlang was great and was going to earn oodels of $'s for
Ericsson :-)

   /Joe


> /Tobbe
>
>
> Bjarne D?cker wrote:
>
>> Hello
>>  A year ago, Joe successfully defended his
>> thesis "Making reliable distributed systems
>> in the presence of software errors" and on
>> December 19 this year, he received his
>> diploma (together with > 200 other new
>> doctors) at a formal occassion in the town
>> hall in Stockholm.
>>  Some photos from this happening can
>> be view at
>>  http://www.cs-lab.org/joes_promotion/
>>  Bjarne
>> 
>
>

From tobbe@REDACTED  Wed Dec  1 15:08:55 2004
From: tobbe@REDACTED (Torbjorn Tornkvist)
Date: Wed, 01 Dec 2004 15:08:55 +0100
Subject: Joe's promotion
In-Reply-To: <Pine.LNX.4.61.0412011500110.31096@enfield.sics.se>
References: <001701c4d794$e4f660a0$b10e69d4@segeltorp> <41ADB5E6.4040507@nortelnetworks.com> <Pine.LNX.4.61.0412011500110.31096@enfield.sics.se>
Message-ID: <41ADD077.4030003@nortelnetworks.com>

Joe Armstrong wrote:

>
>
>> Nice photos!
>>
>> I wonder what Joe told Jan Uddenfeldt...
>>
>
> That Erlang was great and was going to earn oodels of $'s for
> Ericsson :-)

You should have said: "....is earning oodels of $'s for Nortel"

;-) Tobbe

>
> /Joe
>
>
>> /Tobbe
>>
>>
>> Bjarne D?cker wrote:
>>
>>> Hello
>>> A year ago, Joe successfully defended his
>>> thesis "Making reliable distributed systems
>>> in the presence of software errors" and on
>>> December 19 this year, he received his
>>> diploma (together with > 200 other new
>>> doctors) at a formal occassion in the town
>>> hall in Stockholm.
>>> Some photos from this happening can
>>> be view at
>>> http://www.cs-lab.org/joes_promotion/
>>> Bjarne
>>>
>>
>
>>> Bjarne www.cs-lab.org/joes_promotion/ own 
>>> oodels of $'s for 
>>> ?A??Ex??@??E,??@H?E???@?E?@??E'??@0?E9??@x?E??A??E???A?E??@I(E???A?E=??@??E
>>



From carsten@REDACTED  Wed Dec  1 15:18:09 2004
From: carsten@REDACTED (Carsten Schultz)
Date: Wed, 1 Dec 2004 15:18:09 +0100
Subject: String search in a binary
Message-ID: <20041201141808.GB16118@penne.localnet>

Hi!

I have the following piece of code:

extract_contents(Bin) when binary(Bin) ->
    Str = binary_to_list(Bin),
    Start = string:str(Str, "<a name=\"datei")-1,
    End = string:rstr(Str, "</body>"),
    Len = End-Start-1,
    <<_:Start/binary, C:Len/binary, _/binary>> = Bin,
    C.

The binary_to_list is bothering me.  Is there a short and efficient
version doing the same?

Thanks,

Carsten

-- 
Carsten Schultz (2:38, 33:47), FB Mathematik, FU Berlin
http://carsten.codimi.de/
PGP/GPG key on the pgp.net key servers, 
fingerprint on my home page.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: not available
URL: <http://erlang.org/pipermail/erlang-questions/attachments/20041201/0783661d/attachment.bin>

From magnus.thoang@REDACTED  Wed Dec  1 15:56:32 2004
From: magnus.thoang@REDACTED (=?ISO-8859-1?Q?Magnus_Tho=E4ng?=)
Date: Wed, 01 Dec 2004 15:56:32 +0100
Subject: String search in a binary
In-Reply-To: <20041201141808.GB16118@penne.localnet>
References: <20041201141808.GB16118@penne.localnet>
Message-ID: <41ADDBA0.5020608@ericsson.com>

Carsten Schultz wrote:
...
> I have the following piece of code:
> 
> extract_contents(Bin) when binary(Bin) ->
>     Str = binary_to_list(Bin),
>     Start = string:str(Str, "<a name=\"datei")-1,
>     End = string:rstr(Str, "</body>"),
>     Len = End-Start-1,
>     <<_:Start/binary, C:Len/binary, _/binary>> = Bin,
>     C.
> 
> The binary_to_list is bothering me.  Is there a short and efficient
> version doing the same?
...

IIRC, the string:str function is a brute-force text search, so you don't 
gain any performance by using it.

You might do something like...

extract_contents(Bin) when binary(Bin) ->
     extract_contents(Bin, 0).

extract_contents(Bin, Offset) ->
     case Bin of
         <<_:Offset/binary,"<a name=\"datei", Rest/binary>> ->
             extract_body(Rest,0);
         _ ->
             extract_contents(Bin, Offset + 1)
     end.

extract_body(Bin, Offset) ->
     case Bin of
         <<Body:Offset/binary,"</body>",_/binary>> ->
             Body;
         _ ->
             extract_body(Bin, Offset + 1)
     end.

...but since you are not parameterizing the start and end strings, you 
could probably hard-code some clever advancing of the Offset when not 
hitting the strings.

The reason for using case clauses instead of matching with function 
clauses is simply that (for unknown reasons) the syntax is not allowed 
in the function head.

-- 
Magnus Tho?ng



From bjorn@REDACTED  Wed Dec  1 16:25:26 2004
From: bjorn@REDACTED (Bjorn Gustavsson)
Date: 01 Dec 2004 16:25:26 +0100
Subject: parallel LC
In-Reply-To: <37FB7AA6F5F9814FB634A7BF4C35A6F54029B9@ESEALNT442.al.sw.ericsson.se>
References: <37FB7AA6F5F9814FB634A7BF4C35A6F54029B9@ESEALNT442.al.sw.ericsson.se>
Message-ID: <o14qj6kpll.fsf@erix.ericsson.se>

"Ulf Wiger (AL/EAB)" <ulf.wiger@REDACTED> writes:

> OK, great... but then the "Erlang/OTP R10B documentation"
> link at http://www.erlang.org/doc.html doesn't point to 
> OTP R10B_1.

We have fixed it now.

Thanks for pointing it out!

/Bjorn

-- 
Bj?rn Gustavsson, Erlang/OTP, Ericsson AB


From bjorn@REDACTED  Wed Dec  1 16:36:14 2004
From: bjorn@REDACTED (Bjorn Gustavsson)
Date: 01 Dec 2004 16:36:14 +0100
Subject: crypto and ssl modules under OSX
In-Reply-To: <75597826-39B5-11D9-AF62-000A95927CCE@mac.com>
References: <Pine.GSO.3.96.1041118123305.8716B-100000@radagast>
	<75597826-39B5-11D9-AF62-000A95927CCE@mac.com>
Message-ID: <o1wtw2c9ox.fsf@erix.ericsson.se>

Actually, we DID get that part correct in R10B (getting rid of -R
for Mac OS X).

Unfortunately, there is another problem... which I think I have
solved now. If my change doesn't break linking on some other platform,
it shold appear in R10B-2.

/Bjorn

Sean Hinde <sean.hinde@REDACTED> writes:

> I have submitted a patch for this more times than I care to remember.
> 
> In addition the correct mechanism is described in Erlang/OTPs very own
> README file.
> 
> Please try to get this right for OS X next time.
> 
> Sean
> 
> On 18 Nov 2004, at 11:34, Peter H|gfeldt wrote:
> 
> >
> > Just remove the -R switch from the Makefile. In R10B the -R switch is
> > not included for OSX (or is it called Darwin?).
> >
> > /Peter
> >
> >
> > On Wed, 17 Nov 2004, Reto Kramer wrote:
> >
> >> I'm trying to build the R9C-2 ssl and crypto under OSX 10.3.4. Can
> >> some
> >> BSD/Mac guru out there help me figure out how to change the ld
> >> switches
> >> (under OSX, the R option of the -Wl switch is not available).
> >>
> >> Thanks,
> >> - Reto
> >>
> >> from: otp_src_R9C-2/otp_src_R9C-2/lib/crypto/c_src/Makefile.in
> >>
> >> $(LIBDIR)/crypto_drv.so: $(OBJS)
> >> 	$(INSTALL_DIR) $(LIBDIR)
> >> 	$(CC) $(LDFLAGS) -L$(SSL_LIBDIR) -Wl,-R$(SSL_LIBDIR) \
> >> 			-o $@ $^ $(LDLIBS) -lcrypto
> >>
> >>
> >> === Entering application crypto
> >> make[3]: Nothing to be done for `opt'.
> >> make -f powerpc-apple-darwin7.5.0/Makefile TYPE=opt
> >> /usr/bin/install -c -d ../priv/lib/powerpc-apple-darwin7.5.0
> >> gcc -bundle -flat_namespace -undefined suppress -lbundle1.o -L/usr/lib
> >> -Wl,-R/usr/lib \
> >> 		 -o ../priv/lib/powerpc-apple-darwin7.5.0/crypto_drv.so
> >> ../priv/obj/powerpc-apple-darwin7.5.0/crypto_drv.o  -lcrypto
> >> ld: unknown flag: -R/usr/lib
> >> === Skipping subdir doc/src, it is missing
> >> === Leaving application crypto
> >>
> >> #
> >>
> >
> 

-- 
Bj?rn Gustavsson, Erlang/OTP, Ericsson AB


From jamesh@REDACTED  Wed Dec  1 17:47:38 2004
From: jamesh@REDACTED (James Hague)
Date: Wed, 1 Dec 2004 10:47:38 -0600 
Subject: Finding unique elements of a list
Message-ID: <F7698C6BAED2D41189E00002A50A625F06A07CD4@v-exchange.volition.net>

Is there a standard Erlang library function or idiom for removing duplicate
items from an unsorted list?  Ideally the elements in the final list should
be in the order they were passed in, except without duplicates:

	[3,5,2,2,3] -> [3,5,2].

If order doesn't matter there are some easy options:

1. lists:usort(L)
2. gb_sets:to_list(gb_sets:from_list(L))

If order does matter, then the only thing I've come up with is to build an
ets table or gb_set iteratively, adding each element not previously in the
table or set to a list of unique items.  Is there an obvious library
function I'm not seeing?  Feels like there should be lists:unique/1.


From ulf.wiger@REDACTED  Wed Dec  1 18:10:23 2004
From: ulf.wiger@REDACTED (Ulf Wiger (AL/EAB))
Date: Wed, 1 Dec 2004 18:10:23 +0100 
Subject: Finding unique elements of a list
Message-ID: <37FB7AA6F5F9814FB634A7BF4C35A6F54029C5@ESEALNT442.al.sw.ericsson.se>


To my knowledge, there's no such function.
Not that I doubt your ability to write one yourself,
but here's a suggestion:

u(L) -> u(L, []).
u([H|T], Acc) ->
    case lists:member(H,Acc) of
	true ->  u(T, Acc);
	false -> u(T, [H|Acc])
    end;
u([], Acc) ->
    lists:reverse(Acc).


Here's another one, shorter, O(1), and, most importantly, 
cuter (functional purists may retch at will):

u2(L) ->
    T = ets:new(temp,[set]),
    L1 = lists:filter(fun(X) -> ets:insert_new(T, {X,1}) end, L),
    ets:delete(T),
    L1.

Works only on newer erlangs (newest R9C and R10, I believe).

/Uffe

> -----Original Message-----
> From: owner-erlang-questions@REDACTED
> [mailto:owner-erlang-questions@REDACTED]On Behalf Of James Hague
> Sent: den 1 december 2004 17:48
> To: 'erlang-questions@REDACTED'
> Subject: Finding unique elements of a list
> 
> 
> Is there a standard Erlang library function or idiom for 
> removing duplicate
> items from an unsorted list?  Ideally the elements in the 
> final list should
> be in the order they were passed in, except without duplicates:
> 
> 	[3,5,2,2,3] -> [3,5,2].
> 
> If order doesn't matter there are some easy options:
> 
> 1. lists:usort(L)
> 2. gb_sets:to_list(gb_sets:from_list(L))
> 
> If order does matter, then the only thing I've come up with 
> is to build an
> ets table or gb_set iteratively, adding each element not 
> previously in the
> table or set to a list of unique items.  Is there an obvious library
> function I'm not seeing?  Feels like there should be lists:unique/1.
> 


From carsten@REDACTED  Wed Dec  1 18:29:56 2004
From: carsten@REDACTED (Carsten Schultz)
Date: Wed, 1 Dec 2004 18:29:56 +0100
Subject: Finding unique elements of a list
In-Reply-To: <F7698C6BAED2D41189E00002A50A625F06A07CD4@v-exchange.volition.net>
References: <F7698C6BAED2D41189E00002A50A625F06A07CD4@v-exchange.volition.net>
Message-ID: <20041201172955.GA23307@penne.localnet>

Hi!

On Wed, Dec 01, 2004 at 10:47:38AM -0600, James Hague wrote:
> Is there a standard Erlang library function or idiom for removing duplicate
> items from an unsorted list?  Ideally the elements in the final list should
> be in the order they were passed in, except without duplicates:
> 
> 	[3,5,2,2,3] -> [3,5,2].
> 
> If order doesn't matter there are some easy options:
> 
> 1. lists:usort(L)
> 2. gb_sets:to_list(gb_sets:from_list(L))
> 
> If order does matter,

and you do not care for performance, then a quick solution would be

unique([]) ->
    [];
unique([H|T]) ->
    [H|unique(lists:filter(fun(X) -> X =/= H
			   end, T))].

But, yes, this is more reasonable in Haskell and stolen from its
library :-)

> then the only thing I've come up with is to build an ets table or
> gb_set iteratively, adding each element not previously in the table
> or set to a list of unique items.

Definitely the better option.

Sorry, that probably was not really helpful...

Greetings,

Carsten

-- 
Carsten Schultz (2:38, 33:47), FB Mathematik, FU Berlin
http://carsten.codimi.de/
PGP/GPG key on the pgp.net key servers, 
fingerprint on my home page.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: not available
URL: <http://erlang.org/pipermail/erlang-questions/attachments/20041201/3f6a04ce/attachment.bin>

From jamesh@REDACTED  Wed Dec  1 20:21:02 2004
From: jamesh@REDACTED (James Hague)
Date: Wed, 1 Dec 2004 13:21:02 -0600 
Subject: Finding unique elements of a lis
Message-ID: <F7698C6BAED2D41189E00002A50A625F06A07CE1@v-exchange.volition.net>

Ulf Wiger wrote:

>u2(L) ->
>    T = ets:new(temp,[set]),
>    L1 = lists:filter(
>        fun(X) -> ets:insert_new(T, {X,1}) end, L),
>    ets:delete(T),
>    L1.

Oooh, that's pretty.  My implementation did essentially the same thing, but
I didn't use filter.  I suggest this function get added to stdlib as
lists:unique/1.

James


From ulf.wiger@REDACTED  Wed Dec  1 21:14:27 2004
From: ulf.wiger@REDACTED (Ulf Wiger)
Date: Wed, 01 Dec 2004 21:14:27 +0100
Subject: Finding unique elements of a list
In-Reply-To: <37FB7AA6F5F9814FB634A7BF4C35A6F54029C5@ESEALNT442.al.sw.ericsson.se>
References: <37FB7AA6F5F9814FB634A7BF4C35A6F54029C5@ESEALNT442.al.sw.ericsson.se>
Message-ID: <opsiciidpgmk9l4x@mail1.telia.com>

On Wed, 1 Dec 2004 18:10:23 +0100, Ulf Wiger (AL/EAB)  
<ulf.wiger@REDACTED> wrote:

> Here's another one, shorter, O(1), [...]

Now wouldn't that have been neat?
I meant O(N), of course.

/Uffe
-- 
Ulf Wiger



From vances@REDACTED  Wed Dec  1 21:30:02 2004
From: vances@REDACTED (Vance Shipley)
Date: Wed, 1 Dec 2004 15:30:02 -0500
Subject: Finding unique elements of a list
In-Reply-To: <37FB7AA6F5F9814FB634A7BF4C35A6F54029C5@ESEALNT442.al.sw.ericsson.se>
References: <37FB7AA6F5F9814FB634A7BF4C35A6F54029C5@ESEALNT442.al.sw.ericsson.se>
Message-ID: <20041201203002.GA24653@frogman.motivity.ca>

On Wed, Dec 01, 2004 at 06:10:23PM +0100, Ulf Wiger (AL/EAB) wrote:
}  
}  
}  Here's another one, shorter, O(1), and, most importantly, 
}  cuter (functional purists may retch at will):
}  
}  u2(L) ->
}      T = ets:new(temp,[set]),
}      L1 = lists:filter(fun(X) -> ets:insert_new(T, {X,1}) end, L),
}      ets:delete(T),
}      L1.

Those that retched may prefer this:  :)

u2(L) ->
	u2(L, [], gb_sets:empty()).
u2([H|T], Acc, S) ->
	case catch gb_sets:insert(H, S) of
		{'EXIT', _} ->
			u2(T, Acc, S);
		N ->
			u2(T, [H|Acc], N);
	end;
u2([], Acc, _S) ->
	lists:reverse(Acc).



	-Vance


From jamesh@REDACTED  Wed Dec  1 21:35:55 2004
From: jamesh@REDACTED (James Hague)
Date: Wed, 1 Dec 2004 14:35:55 -0600 
Subject: dict and set variants
Message-ID: <F7698C6BAED2D41189E00002A50A625F06A07CEA@v-exchange.volition.net>

stdlib has three modules each for sets and dicts (key/value pairs):

sets: sets, ordsets, gb_sets
dicts: dict, orddict, gb_trees (and ets, I suppose :)

(BTW, notice the naming discrepancy.  Most are plural, some aren't.)

So, what's the rationale behind three versions of each?  This doesn't appear
to be explained anywhere.  To the best of my knowledge, sets/dict have been
superceded by gb_sets/gb_trees (am I alone in thinking that the names
gb_sets and gb_trees are horrible?).  I'm not sure if there are reasons to
use sets/dict any more (should they go away?).  ordsets/orddict are
lighter-weight modules that just store everything in a big list.  Great for
small amounts of data, not so good when things get huge.

Am I close?



From erlang-list@REDACTED  Wed Dec  1 22:40:53 2004
From: erlang-list@REDACTED (Dominic Williams)
Date: Wed, 01 Dec 2004 22:40:53 +0100
Subject: Finding unique elements of a list
In-Reply-To: <F7698C6BAED2D41189E00002A50A625F06A07CD4@v-exchange.volition.net>
References: <F7698C6BAED2D41189E00002A50A625F06A07CD4@v-exchange.volition.net>
Message-ID: <41AE3A65.8000200@dominicwilliams.net>

FWIW, the version I have in my own lib is:

unique(List) ->
     lists:reverse(
       lists:foldl(
         fun(X,Acc) ->
                 accumulate_unless(lists:member(X,Acc),X,Acc)
         end,
         [],
         List)).

accumulate_unless(true, _X, Acc) ->
     Acc;
accumulate_unless(false, X, Acc) ->
     [X|Acc].


Regards,

Dominic Williams
http://www.dominicwilliams.net

----


From matthias@REDACTED  Thu Dec  2 10:50:45 2004
From: matthias@REDACTED (Matthias Lang)
Date: Thu, 2 Dec 2004 10:50:45 +0100
Subject: dict and set variants
In-Reply-To: <F7698C6BAED2D41189E00002A50A625F06A07CEA@v-exchange.volition.net>
References: <F7698C6BAED2D41189E00002A50A625F06A07CEA@v-exchange.volition.net>
Message-ID: <16814.58741.791545.541634@antilipe.corelatus.se>

James Hague writes:

 > stdlib has three modules each for sets and dicts (key/value pairs):
 > 
 > sets: sets, ordsets, gb_sets
 > dicts: dict, orddict, gb_trees (and ets, I suppose :)
 > 
 > (BTW, notice the naming discrepancy.  Most are plural, some aren't.)
 > 
 > So, what's the rationale behind three versions of each?  

There isn't any, it's just various people letting code escape from
the lab before it got cleaned up and then writing some half-hearted
documentation before going on to something more interesting.

My favourite is this bit in the 'ordsets' documentation:

   "An ordered list is more efficient than an unordered list."

--------------------

Anyone care to extend and correct this table?

           implementation   insert  find    traverse:random/in-order
----------------------------------------------------------------------
   sets    hash?            1 ?     1 ?     N              /not-possible
ordsets    sorted list      N       N       N              /N
   list    list             1       N       N              /not-possible
gb_sets    tree             lg N    lg N    N ?            /N lg N
    ets    hash             1       1       N              /N

'N' means O(N), 'lg N' means O(log N), etc.

I haven't tried to indicate whether the bound is worst-case or
on-average because, in general, I don't know. I think a random
traverse of a gb_set is O(N), but the documentation doesn't actually
promise that, so maybe it's O(log N).

(Keep in mind that the overhead isn't shown above and that it often
does matter. There's little point using a tree when your set has ten
members.)

Matthias


From thomasl_erlang@REDACTED  Thu Dec  2 11:22:42 2004
From: thomasl_erlang@REDACTED (Thomas Lindgren)
Date: Thu, 2 Dec 2004 02:22:42 -0800 (PST)
Subject: Finding unique elements of a list
In-Reply-To: <opsiciidpgmk9l4x@mail1.telia.com>
Message-ID: <20041202102242.38694.qmail@web41902.mail.yahoo.com>


Here is a straightforward version with good O(.):
first sort the list, then drop consecutive duplicates.

Complexity is given by the sort (assuming unit cost of
comparing elements). O(N log N), presumably? 

unique(Xs) ->
    rem_dups(lists:sort(Xs)).

rem_dups([X|Xs]) ->
    [X|drop(X, Xs)];
rem_dups([]) ->
    [].

drop(X, [X|Xs]) ->
    drop(X, Xs);
drop(X, Ys) ->
    rem_dups(Ys).

The code above seems to work, but could be optimized
further; which is left as an exercise to the reader.

Best,
Thomas



		
__________________________________ 
Do you Yahoo!? 
Meet the all-new My Yahoo! - Try it today! 
http://my.yahoo.com 
 



From ulf.wiger@REDACTED  Thu Dec  2 11:34:34 2004
From: ulf.wiger@REDACTED (Ulf Wiger (AL/EAB))
Date: Thu, 2 Dec 2004 11:34:34 +0100 
Subject: Finding unique elements of a list
Message-ID: <37FB7AA6F5F9814FB634A7BF4C35A6F54029CC@ESEALNT442.al.sw.ericsson.se>


> From: Vance Shipley [mailto:vances@REDACTED]
> Those that retched may prefer this:  :)
> 
> u2(L) ->
> 	u2(L, [], gb_sets:empty()).
> u2([H|T], Acc, S) ->
> 	case catch gb_sets:insert(H, S) of
> 		{'EXIT', _} ->
> 			u2(T, Acc, S);
> 		N ->
> 			u2(T, [H|Acc], N);
> 	end;
> u2([], Acc, _S) ->
> 	lists:reverse(Acc).


This has the desireable property that the elements
of the list are not duplicated (since inserting them
in a heap-based structure only means copying the
pointer. My ets-based version would duplicate all 
elements in the final list, which could of course be
very costly if the elements are large.

/Uffe


From carsten@REDACTED  Thu Dec  2 12:10:37 2004
From: carsten@REDACTED (Carsten Schultz)
Date: Thu, 2 Dec 2004 12:10:37 +0100
Subject: Finding unique elements of a list
In-Reply-To: <20041202102242.38694.qmail@web41902.mail.yahoo.com>
References: <opsiciidpgmk9l4x@mail1.telia.com> <20041202102242.38694.qmail@web41902.mail.yahoo.com>
Message-ID: <20041202111036.GA24436@penne.localnet>

Hi Thomas!

On Thu, Dec 02, 2004 at 02:22:42AM -0800, Thomas Lindgren wrote:
> 
> Here is a straightforward version with good O(.):
> first sort the list, then drop consecutive duplicates.

Please read the original question and also note the existence of
lists:usort :-)

Take care,

Carsten
-- 
Carsten Schultz (2:38, 33:47), FB Mathematik, FU Berlin
http://carsten.codimi.de/
PGP/GPG key on the pgp.net key servers, 
fingerprint on my home page.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: not available
URL: <http://erlang.org/pipermail/erlang-questions/attachments/20041202/8c6c62eb/attachment.bin>

From carsten@REDACTED  Thu Dec  2 12:22:26 2004
From: carsten@REDACTED (Carsten Schultz)
Date: Thu, 2 Dec 2004 12:22:26 +0100
Subject: Finding unique elements of a list
In-Reply-To: <37FB7AA6F5F9814FB634A7BF4C35A6F54029C5@ESEALNT442.al.sw.ericsson.se>
References: <37FB7AA6F5F9814FB634A7BF4C35A6F54029C5@ESEALNT442.al.sw.ericsson.se>
Message-ID: <20041202112225.GB24436@penne.localnet>

Hi Uffe!

On Wed, Dec 01, 2004 at 06:10:23PM +0100, Ulf Wiger (AL/EAB) wrote:
> Here's another one, shorter, O(1), and, most importantly, 
> cuter 

Hmm...

> (functional purists may retch at will):

I have just learned a new word ;-)

Ok, since you asked for it...

> u2(L) ->
>     T = ets:new(temp,[set]),
>     L1 = lists:filter(fun(X) -> ets:insert_new(T, {X,1}) end, L),
>     ets:delete(T),
>     L1.

The documentation of lists:filter does not say anything about the
order in which the predicate function is called for the various list
elements.  It does not even state if it is called several times if an
element appears several times.

This is nasty.  But nastyness and cuteness go together often enough,
so who cares.

Take care,

Carsten

-- 
Carsten Schultz (2:38, 33:47), FB Mathematik, FU Berlin
http://carsten.codimi.de/
PGP/GPG key on the pgp.net key servers, 
fingerprint on my home page.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: not available
URL: <http://erlang.org/pipermail/erlang-questions/attachments/20041202/9db2b53b/attachment.bin>

From magnus.thoang@REDACTED  Thu Dec  2 12:53:27 2004
From: magnus.thoang@REDACTED (=?ISO-8859-1?Q?Magnus_Tho=E4ng?=)
Date: Thu, 02 Dec 2004 12:53:27 +0100
Subject: String search in a binary
In-Reply-To: <41ADDBA0.5020608@ericsson.com>
References: <20041201141808.GB16118@penne.localnet> <41ADDBA0.5020608@ericsson.com>
Message-ID: <41AF0237.7070508@ericsson.com>

Magnus Tho?ng wrote:
> IIRC, the string:str function is a brute-force text search, so you don't 
> gain any performance by using it.
 > ...

Ooops. This was not true.

Some testing reveals that traversing a list (as in string:str/2) is 
faster than traversing a binary in the way I suggested...

Sorry :-)

-- 
Magnus Tho?ng


From thomasl_erlang@REDACTED  Thu Dec  2 15:35:11 2004
From: thomasl_erlang@REDACTED (Thomas Lindgren)
Date: Thu, 2 Dec 2004 06:35:11 -0800 (PST)
Subject: Finding unique elements of a list
In-Reply-To: <16814.61552.715963.431243@harpo.it.uu.se>
Message-ID: <20041202143511.98326.qmail@web41906.mail.yahoo.com>


--- Sven-Olof Nystr|m <svenolof@REDACTED> wrote:

> Also, the original poster required that the
> order of the input
> list should be preserved. Besides that, nice try :-)

Yes, I somehow forgot that requirement ... sorry about
that :-) Oh well.

Best,
Thomas



		
__________________________________ 
Do you Yahoo!? 
Read only the mail you want - Yahoo! Mail SpamGuard. 
http://promotions.yahoo.com/new_mail 


From heinrich@REDACTED  Thu Dec  2 15:33:27 2004
From: heinrich@REDACTED (Heinrich Venter)
Date: Thu, 2 Dec 2004 16:33:27 +0200
Subject: Http_response funny
Message-ID: <CE32AABA6936B0459B3EBE99331FC75D09E059@server.Cellfind.local>

Hi

I am using http:request (inets 4.0) to do a SOAP request.  The status
line has a type of "100 Continue".  I get an error because the function
http_response:status_continue returns only ok when the
httpc_handler:handle_response case statements expect something like {ok,
""}

Simply replacing the ok with {ok, ""} breaks rest of the resonses
because now the next status line is deemed as part of the body of the
"100 Continue" response.

Stil looking for a solution to this.  Any one seen this before?

-]-[einrich


Click here for more information.
http://asp1.rocketseed.com/RocketSeed/mail/433a32353a3839353032313a33303639373a2d323a313831

Bought to you by Launchpad.
http://asp1.rocketseed.com/RocketSeed/mail/433a32353a3839353032313a33303639373a2d323a313438
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://erlang.org/pipermail/erlang-questions/attachments/20041202/a028cdbd/attachment.htm>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: MTk2LjMwLjc5LjE1NQ==
Type: image/gif
Size: 557 bytes
Desc: not available
URL: <http://erlang.org/pipermail/erlang-questions/attachments/20041202/a028cdbd/attachment.gif>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: MTk2LjMwLjc5LjE1NQ==
Type: image/gif
Size: 16459 bytes
Desc: not available
URL: <http://erlang.org/pipermail/erlang-questions/attachments/20041202/a028cdbd/attachment-0001.gif>

From thomasl_erlang@REDACTED  Thu Dec  2 16:30:57 2004
From: thomasl_erlang@REDACTED (Thomas Lindgren)
Date: Thu, 2 Dec 2004 07:30:57 -0800 (PST)
Subject: Finding unique elements of a list
In-Reply-To: <16815.11933.395850.484405@harpo.it.uu.se>
Message-ID: <20041202153057.64471.qmail@web41901.mail.yahoo.com>


--- Sven-Olof Nystr|m <svenolof@REDACTED> wrote:

> Removing duplicates with sort assumes that < is a
> total
> ordering. Unfortunately it is not: We have neither 1
> < 1.0 nor 1.0 < 1
> Thus,  trying to remove the duplicates of a list of
> floats and integers
> will give strange results.
> 
> 3> tl:unique([1,1.0,1,1.0]).
> [1,1.00000,1,1.00000]

As far as I can tell from lists.erl, sort seems to use
==/2 to actually get a total ordering; since 1 == 1.0,
we get the effect of "mixed" floats and integers.
Perhaps someone closer acquainted with sort can tell
us the real story.

Rewriting drop/2 of my original program to use ==/2
instead of pattern matching yields [1.00000] for the
example above, eliminating all duplicates (as far as
==/2 is concerned). At this point, those of a
legalistic bent may perk up and begin to ask what
defines a "unique" element ... :-)

Moral of the story: stay away from floats?

Best,
Thomas



		
__________________________________ 
Do you Yahoo!? 
Yahoo! Mail - Helps protect you from nasty viruses. 
http://promotions.yahoo.com/new_mail


From ulf.wiger@REDACTED  Thu Dec  2 17:17:58 2004
From: ulf.wiger@REDACTED (Ulf Wiger (AL/EAB))
Date: Thu, 2 Dec 2004 17:17:58 +0100 
Subject: risky_shell
Message-ID: <37FB7AA6F5F9814FB634A7BF4C35A6F54029D3@ESEALNT442.al.sw.ericsson.se>



> -----Original Message-----
> From: peppe@REDACTED [mailto:peppe@REDACTED]

> We have not documented the details because we don't want 
> people to use this flag and, like you, become dependent on
> the old shell startup behaviour. So don't push me or I'll
> take away this flag of yours! :-)

I thought my suggestion to mention the name of the flag in the 
release notes only was quite reasonable.  (-:


> The old parallell startup behaviour *could* be useful to
> start tracing early.

*Is* useful.
Feel free to provide a list of *more* useful ways to 
accomplish the same thing.  ;-)
(I believe this was the main reason why you agreed to 
put it back in...)


> But since it is indeed parallell and the time to start 
> the shell process depends on e.g. io speed, it must be an
> unpredictable way to start tracing an application start
> sequence.

Indeed. Depending on how quickly the situation occurs that 
you want to trace, it can be difficult to get the commands
entered quickly enough. People here have learned to keep 
such commands in a text editor window, and then select 
a bunch of lines at a time, and copy and paste into the 
erlang shell. Also, a system the size of the AXD 301 doesn't
start *that* quickly. Code loading alone takes nearly one
minute (actually, this happens before even the old shell
appears, when embedded loading is used), and loading the 
Mnesia database from disk takes another minute (and we have
code that makes sure that no "interesting" applications 
start before Mnesia is done loading). For many of the 
sequences that one might want to trace on, there's plenty 
of time.

OK, I didn't really mean to push you to make the feature 
"documented" in the sense that it would encourage people
to depend on it. 

Tracing early on a system *is* difficult, and there is
no single good way to do it. I welcome initiatives to 
relieve people of the challenge of pasting in shell
commands quickly enough to catch the bug, but not 
so quickly that they break something. (-:
In the meantime, one might even want to have a 
"tricky fault finding" tutorial, where anything goes,
of course with all the appropriate disclaimers.

If I may, there are some documented functions where 
users are *strongly discouraged* from even using them 
at all. I think this is appropriate. Some examples perhaps:

"erlang:bump_reductions(Reductions)
[...]This BIF might be removed in a future version of the 
Beam machine without prior warning. It is unlikely to be 
implemented in other Erlang implementations. If you think 
that you must use it, encapsulate it your own wrapper module, 
and/or wrap it in a catch."

"erlang:suspend_process(Pid)
Suspends the process Pid. 

 <WARNING/> This BIF is intended for debugging only."


"[application:]set_env(Application, Par, Val) -> ok
[...] <WARNING/> Use this function only if you know what 
you are doing, that is, on your own applications. It is 
very application and configuration parameter dependent when 
and how often the value is read by the application, and 
careless use of this function may put the application in 
a weird, inconsistent, and malfunctioning state. "


/Uffe


From peppe@REDACTED  Thu Dec  2 18:51:28 2004
From: peppe@REDACTED (UAB L/K Peter Andersson)
Date: Thu, 02 Dec 2004 18:51:28 +0100
Subject: risky_shell
References: <37FB7AA6F5F9814FB634A7BF4C35A6F54029D3@ESEALNT442.al.sw.ericsson.se>
Message-ID: <41AF5620.A085B982@erix.ericsson.se>


[...]
> 
> I thought my suggestion to mention the name of the flag in the
> release notes only was quite reasonable.  (-:
> 

Fair enough, but I'd rather not mention it at all. ;-) 


> > The old parallell startup behaviour *could* be useful to
> > start tracing early.
> 
> *Is* useful.
> Feel free to provide a list of *more* useful ways to
> accomplish the same thing.  ;-)
> (I believe this was the main reason why you agreed to
> put it back in...)
> 

Maybe the application could start in two steps: First start a tracer
process (which evaluates a trace config file, for example), then
initiate the main application start sequence.

You could maybe use the -s flag to call a main tracer start function
(again which perhaps evaluates a trace config file). That should switch
on tracing right after erts & kernel, but before your own applications
start. 

If it's not possible for you to have tracing start automatically, but
need the option to call these functions manually from the shell, I agree
there's no simple straightforward way to sync your application with the
shell.

I thought the reason you insisted we put the old behaviour back was that
it was not feasible for you to change the way you are tracing today in
AXD-301, not that this was necessarily the most useful way to accomplish
it.

With my statement I wanted to say to everyone else: "please find other
ways to start tracing of your applications than go experimenting with
the old error prone shell start behaviour, which may even give you
unpredictable results". Not documenting the flag in question is of
course a way of saying the same thing.


[...]
> 
> OK, I didn't really mean to push you to make the feature
> "documented" in the sense that it would encourage people
> to depend on it.
> 
[...]

I see your point. I could - and maybe I should - have documented the
flag as one of those *strongly discouraged* things. But you know,
reintroducing an error (even under somewhat controlled forms) isn't
something to be proud of. I thought I could get away with not saying
anything to anyone, except you. Then you, of all people, give me away!
You'll never get another undocumented flag from me again! :-)

 /Peppe


From heinrich@REDACTED  Thu Dec  2 18:59:12 2004
From: heinrich@REDACTED (Heinrich Venter)
Date: Thu, 2 Dec 2004 19:59:12 +0200
Subject: Inets 4.0 and HTTP/1.1 "100 Continue" responses
Message-ID: <CE32AABA6936B0459B3EBE99331FC75D09E06A@server.Cellfind.local>

It looks like I have managed to fix the problem with the 100 Continue
response lines

It's a bit messy and I certainly wouldn't recommend any one else use it
until the powers that be ok'd it.

I am only going to post the bits I changed

The idea is that if the "100 Continue" response is recognised, the Body
is re-parsed from scratch.  I am not sure if all the extra response
handling in the parsing case is neccesary.
I also don't know if the behavior of the header "expect=100-continue"
will stil be correct.

-]-[einrich


-- httpc_handler.erl

handle_response(State = #state{request = Request = #request{id = ID,
							    from =
Client},
			       session = Session, 
			       status_line = StatusLine,
			       headers = Headers, body = Body}) ->
    
    case lists:member(ID, State#state.canceled) of
	true ->
	    handle_pipeline(
	      State#state{canceled = lists:delete(ID, 
	
State#state.canceled)});
	false ->
	    case http_response:result({StatusLine, Headers, Body}, 
				       Request, Session) of
		{ok, ""} -> % redirect
                    handle_pipeline(State);
                %% Added this
                {ok, continue} ->
                    
                    {Module, Function, Args} = {http_response, parse,
	
[State#state.max_header_size]},
                    Data = State#state.body,
                    Method = (State#state.request)#request.method,
      		    case Module:Function([Data | Args]) of
                        {ok, Result} ->
                            handle_http_msg(Result, State); 
                        {_, whole_body, _} when Method == head ->
                            handle_response(State#state{body = <<>>}); 
                        NewMFA ->
 
http_transport:setopts(Session#tcp_session.scheme, 
 
Session#tcp_session.socket, 
                                                  [{active, once}]),
                            {noreply, State#state{mfa = NewMFA}}
                    end;
                    %% End addition
		{ok, Msg} ->
		    http_response:send(Client, Msg),
		    handle_pipeline(State);
		{stop, Msg} ->
		    http_response:send(Client, Msg),
		    {stop, normal, State}
	    end
    end.

-- http_response.erl

%%% RFC2616, Section 10.1.1
%%% Note:
%%% - Only act on the 100 status if the request included the
%%%   "Expect:100-continue" header, otherwise just ignore this response.
status_continue(Req = #http_request_h{expect="100-continue"}, Session)
->
    {_, Body} = Req#request.content,
    http_transport:send(Session#tcp_session.scheme, 
			Session#tcp_session.socket, Body),
    {ok, continue};

status_continue(_Req, _Session) ->
    error_logger:info_msg("Found Continue"),
    {ok, continue}.




Look 4 Help - Click here for more info
http://asp1.rocketseed.com/RocketSeed/mail/433a32353a3839353138393a33303639373a2d323a313933

 
http://asp1.rocketseed.com/RocketSeed/mail/433a32353a3839353138393a33303639373a2d323a313836
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://erlang.org/pipermail/erlang-questions/attachments/20041202/ca571d95/attachment.htm>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: MTk2LjMwLjc5LjE1NQ==
Type: image/gif
Size: 19652 bytes
Desc: not available
URL: <http://erlang.org/pipermail/erlang-questions/attachments/20041202/ca571d95/attachment.gif>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: MTk2LjMwLjc5LjE1NQ==
Type: image/gif
Size: 556 bytes
Desc: not available
URL: <http://erlang.org/pipermail/erlang-questions/attachments/20041202/ca571d95/attachment-0001.gif>

From parault2@REDACTED  Thu Dec  2 19:59:33 2004
From: parault2@REDACTED (Pat29)
Date: Thu, 2 Dec 2004 19:59:33 +0100
Subject: newby questions
Message-ID: <005f01c4d8a1$0fe66c30$3c6efea9@BLUESKY>

Hello,

It's my last post about this subject because i solved my problem
even it's certainely not the best solution.
Use REPOS1.0 to test some Windows configurations with a modified start.bat

rem set 
EJABBERD_CONFIG_PATH=../../../other_apps/ejabberd/src/win32/ejabberd.cfg
set EJABBERD_CONFIG_PATH=../../../other_apps/ejabberd/ejabberd.cfg
set MNESIA_DIR='"%TMP%/mnesia"'

rem To allow more than 1000 opened sockets:
set ERL_MAX_PORTS=32000

rem JEAI
set JEAI_PATH=../../../other_apps/j-eai
set JEAI_CONFIG_PATH=%JEAI_PATH%/priv/ejabberd-jeai.cfg
set CREATE_JEAI_DEMO_USERS=true

rem explorer index.html
start index.html
cd erlang\erts-5.3-Windows\bin
rem in case of boot problem werl -init_debug
echo %EJABBERD_SO_PATH%
start werl -gs backend_comm port %* -pa "%EJABBERD_SO_PATH%" -pa 
"%JEAI_PATH%/yaws/ebin" -mnesia dir %MNESIA_DIR% -yaws embedded 'true'

WINXP PRO SP2
--------------------
repos:check() and repos:start(wings) passed but other tests which use 
sockets failed.

Erlang (BEAM) emulator version 5.3.6.3 [source] [threads:0]

Eshell V5.3.6.3  (abort with ^G)
1> repos:check().

=INFO REPORT==== 30-Nov-2004::06:08:03 ===
{"eunit.erl(80)","Running unit tests in module [repos_check]"}

=INFO REPORT==== 30-Nov-2004::06:08:03 ===
{"eunit.erl(118)","Running test repos_check:xmerl_test/0 - passed."}

=INFO REPORT==== 30-Nov-2004::06:08:06 ===
{"eunit.erl(118)","Running test repos_check:ssl_test/0 - passed."}

=INFO REPORT==== 30-Nov-2004::06:08:07 ===
{"eunit.erl(118)","Running test repos_check:crypto_test/0 - passed."}

=INFO REPORT==== 30-Nov-2004::06:08:12 ===
{"eunit.erl(118)","Running test repos_check:gs_test/0 - passed."}
{4,
 0,
 [{repos_check,{gs_test,0}},
  {repos_check,{crypto_test,0}},
  {repos_check,{ssl_test,0}},
  {repos_check,{xmerl_test,0}}],
 []}

=INFO REPORT==== 30-Nov-2004::06:08:12 ===
{"eunit.erl(77)","Test finished\n"}

=INFO REPORT==== 30-Nov-2004::06:08:12 ===
{"eunit.erl(123)","Summary: 4 passed, 0failed"}

WINXP PRO SP1 with some KBs
------------------------------------
Like WINXP PRO SP2.

WIN2000 PRO
----------------
All test passed but repos:start(wings) failed (OpenGL not installed).

WINXP PRO without any SP and KBs
----------------------------------------
All test passed and  the ping pong example given in the Getting Start worked 
fine.
Now i can enjoy fully with Erlang/OTP.

Thanks again for all peoples for your helps.
Pat. 





From ernie.makris@REDACTED  Thu Dec  2 20:25:14 2004
From: ernie.makris@REDACTED (ernie.makris@REDACTED)
Date: Thu, 02 Dec 2004 19:25:14 +0000
Subject: qlc debug statements
Message-ID: <120220041925.1393.41AF6C1A0001D7CC0000057122070029539C079D050E03D20A07029D0A@comcast.net>

Hello fellow erlangers,

I was playing around with qlc against an ets table and saw the following debug statements being printed:

prep match_spec 2 Opt = {qlc_opt,false,false,-1}

I looked inside qlc.erl and found: 

prep_qlc_lc({single_v1, QFun, CodeF, Qdata0, _, MS, PosFun}, Opt, GOpt) ->
    %% Filter optional:
    [?qual_data(QNum, GoI_G, SI_G, {gen, LE_fun}) | Filter] = Qdata0, 
    #prepared{qh = LE0} = Prep = eval_le(LE_fun, GOpt),
    LuV = find_const_positions(LE0, PosFun, Opt),
    case LE0 of
        #qlc_table{trav_MS = true} when MS =/= no_match_spec, LuV =/= false ->
            io:format("prep match_spec 0 Opt = ~p~n", [Opt]),
            io:format("    lookup values ~p~n", [LuV]),
            io:format("    match spec ~p~n", [MS]),
            LE = LE0#qlc_table{lu_vals = LuV, ms = MS},
            may_create_simple(Opt, Prep#prepared{qh = LE});
        #qlc_table{} when LuV =/= false ->
            io:format("prep match_spec 1 Opt = ~p~n", [Opt]),
            io:format("    lookup values ~p~n", [LuV]),
            Prep1 = Prep#prepared{qh = LE0#qlc_table{lu_vals = LuV}},
            Qdata = [?qual_data(QNum, GoI_G, SI_G, {gen, Prep1}) | Filter],
            prep_qlc(QFun, CodeF, Qdata, Opt);
        #qlc_table{trav_MS = true} when MS =/= no_match_spec ->
            io:format("prep match_spec 2 Opt = ~p~n", [Opt]),
            may_create_simple(Opt, Prep#prepared{qh = LE0#qlc_table{ms = MS}});
        #qlc_list{l = []} ->
            io:format("pre match_spec empty list~n"),
            may_create_simple(Opt, Prep);
        #qlc_list{ms = no_match_spec} when MS =/= no_match_spec ->
            io:format("pre match_spec list no_match_spec~n"),
            may_create_simple(Opt, Prep#prepared{qh = LE0#qlc_list{ms = MS}});
        #qlc_list{} when MS =/= no_match_spec ->
            io:format("pre match_spec list match_spec~n"),
            ListMS = #qlc_list{l = Prep, ms = MS},
            may_create_simple(Opt, #prepared{qh = ListMS, is_cached = true});
        _ ->
            io:format("pre match_spec qlc~n"),
            Qdata = [?qual_data(QNum, GoI_G, SI_G, {gen, Prep}) | Filter],
            prep_qlc(QFun, CodeF, Qdata, Opt)
    end;

Can this be removed in the next version?

Thanks
Ernie


From sean.hinde@REDACTED  Thu Dec  2 23:32:15 2004
From: sean.hinde@REDACTED (Sean Hinde)
Date: Thu, 2 Dec 2004 22:32:15 +0000
Subject: crypto and ssl modules under OSX
In-Reply-To: <o1wtw2c9ox.fsf@erix.ericsson.se>
References: <Pine.GSO.3.96.1041118123305.8716B-100000@radagast> <75597826-39B5-11D9-AF62-000A95927CCE@mac.com> <o1wtw2c9ox.fsf@erix.ericsson.se>
Message-ID: <04B471F6-44B2-11D9-A09D-000A95927CCE@mac.com>

Hi Bjorn,

I hope your change gets rid of the undocumented -lbundle1.o and reverts 
to using the mechanism in the README - namely -bundle, and using the cc 
program to link rather than ld

Quote from README:

If you develop linked in drivers (shared library) you need to link
using "cc" and the flags "-bundle -flat_namespace -undefined
suppress". You also include "-fno-common" in CFLAGS when
compiling. Use ".so" as the library suffix.

Sean


On 1 Dec 2004, at 15:36, Bjorn Gustavsson wrote:

> Actually, we DID get that part correct in R10B (getting rid of -R
> for Mac OS X).
>
> Unfortunately, there is another problem... which I think I have
> solved now. If my change doesn't break linking on some other platform,
> it shold appear in R10B-2.
>
> /Bjorn
>
> Sean Hinde <sean.hinde@REDACTED> writes:
>
>> I have submitted a patch for this more times than I care to remember.
>>
>> In addition the correct mechanism is described in Erlang/OTPs very own
>> README file.
>>
>> Please try to get this right for OS X next time.
>>
>> Sean
>>
>> On 18 Nov 2004, at 11:34, Peter H|gfeldt wrote:
>>
>>>
>>> Just remove the -R switch from the Makefile. In R10B the -R switch is
>>> not included for OSX (or is it called Darwin?).
>>>
>>> /Peter
>>>
>>>
>>> On Wed, 17 Nov 2004, Reto Kramer wrote:
>>>
>>>> I'm trying to build the R9C-2 ssl and crypto under OSX 10.3.4. Can
>>>> some
>>>> BSD/Mac guru out there help me figure out how to change the ld
>>>> switches
>>>> (under OSX, the R option of the -Wl switch is not available).
>>>>
>>>> Thanks,
>>>> - Reto
>>>>
>>>> from: otp_src_R9C-2/otp_src_R9C-2/lib/crypto/c_src/Makefile.in
>>>>
>>>> $(LIBDIR)/crypto_drv.so: $(OBJS)
>>>> 	$(INSTALL_DIR) $(LIBDIR)
>>>> 	$(CC) $(LDFLAGS) -L$(SSL_LIBDIR) -Wl,-R$(SSL_LIBDIR) \
>>>> 			-o $@ $^ $(LDLIBS) -lcrypto
>>>>
>>>>
>>>> === Entering application crypto
>>>> make[3]: Nothing to be done for `opt'.
>>>> make -f powerpc-apple-darwin7.5.0/Makefile TYPE=opt
>>>> /usr/bin/install -c -d ../priv/lib/powerpc-apple-darwin7.5.0
>>>> gcc -bundle -flat_namespace -undefined suppress -lbundle1.o 
>>>> -L/usr/lib
>>>> -Wl,-R/usr/lib \
>>>> 		 -o ../priv/lib/powerpc-apple-darwin7.5.0/crypto_drv.so
>>>> ../priv/obj/powerpc-apple-darwin7.5.0/crypto_drv.o  -lcrypto
>>>> ld: unknown flag: -R/usr/lib
>>>> === Skipping subdir doc/src, it is missing
>>>> === Leaving application crypto
>>>>
>>>> #
>>>>
>>>
>>
>
> -- 
> Bj?rn Gustavsson, Erlang/OTP, Ericsson AB



From richardc@REDACTED  Fri Dec  3 01:30:12 2004
From: richardc@REDACTED (Richard Carlsson)
Date: Fri, 03 Dec 2004 01:30:12 +0100
Subject: Jungerl edoc/syntax_tools
Message-ID: <41AFB394.4030200@csd.uu.se>

I've finally updated the versions of edoc and syntax_tools
in the jungerl (the code is the same that you can find at
http://user.it.uu.se/~richardc/edoc/).

	/Richard


From ok@REDACTED  Fri Dec  3 01:52:38 2004
From: ok@REDACTED (Richard A. O'Keefe)
Date: Fri, 3 Dec 2004 13:52:38 +1300 (NZDT)
Subject: Finding unique elements of a list
Message-ID: <200412030052.iB30qc3u392567@atlas.otago.ac.nz>


If you want to remove duplicate elements from a list but otherwise
preserve the order of the elements, so that, e.g., 

    remove_dups([3,1,4,1,5,9,2,6,5,3,5,8,9]) -> [3,1,4,5,9,2,6,8]

then there is an O(N.lgN) method which goes like this:
    (1) pair each element with its position number in the list
	O(N)
    (2) sort the list of pairs (using whole pairs as keys)
	O(N.lgN)
    (3) remove adjacent duplicates
	O(N)
    (4) switch the {Elem,Posn} pairs to {Posn,Elem} pairs
	O(U) where 1 <= U <= N, U is the number of unique elements
    (5) sort the list of {Posn,Elem} pairs
	O(U.lgU)
    (6) strip off the positions, {Posn,Elem} -> Elem

In Haskell notation, to keep this message short,

    remove_dups :: Ord a => [a] -> [a]
    
    remove_dups xs = [x | (_,x) <- sort (remdups (sort (zip xs [1..])))]
      where remdups []            = []
            remdups ((x,n) : xns) = (n,x) : remdup2 xns x

            remdup2 ((y,_) : xns) x | y == x = remdup2 xns x
            remdup2 xns           _          = remdups xns

This is a little over twice as slow as simply sorting the elements and
removing adjacent duplicates, so it's only something you would use if
the original order of the elements mattered.



From thomas.xa.johnsson@REDACTED  Fri Dec  3 08:45:49 2004
From: thomas.xa.johnsson@REDACTED (Thomas Johnsson XA (LN/EAB))
Date: Fri, 3 Dec 2004 08:45:49 +0100 
Subject: Finding unique elements of a list
Message-ID: <A943FD84BD9ED41193460008C79180500CE0BFAC@ESEALNT419.al.sw.ericsson.se>

In other functional languages I've used (Haskell, (Lazy) ML,..), I've often used the following variation of quicksort when representing sets as sorted lists without duplicates:
(This is Erlang syntax, not Haskell..:-)

mkset([M|Xs]) ->
    mkset([X||X<-Xs,X<M])++[M]++mkset([X||X<-Xs,X>M]);
mkset([]) -> 
    [].

Cute, if I may say so myself (:-), but unfortunately O(n^2) when the argument
already is a sorted list :-(

Erlang has lists:usort, which presumably does the same thing,
hopefully more efficiently. (?)

-- Thomas




From hasse@REDACTED  Fri Dec  3 09:02:01 2004
From: hasse@REDACTED (Hans Bolinder)
Date: Fri, 3 Dec 2004 09:02:01 +0100
Subject: qlc debug statements
In-Reply-To: <120220041925.1393.41AF6C1A0001D7CC0000057122070029539C079D050E03D20A07029D0A@comcast.net>
References: <120220041925.1393.41AF6C1A0001D7CC0000057122070029539C079D050E03D20A07029D0A@comcast.net>
Message-ID: <16816.7545.356360.485212@gargle.gargle.HOWL>

> I was playing around with qlc against an ets table and saw the
> following debug statements being printed:
> 
> prep match_spec 2 Opt = {qlc_opt,false,false,-1}

> Can this be removed in the next version?

It will be fixed in R10B-2 which is due within a couple of weeks.

Thanks for pointing it out.

Best regards,

Hans Bolinder, Erlang/OTP


From bjorn@REDACTED  Fri Dec  3 10:01:06 2004
From: bjorn@REDACTED (Bjorn Gustavsson)
Date: 03 Dec 2004 10:01:06 +0100
Subject: crypto and ssl modules under OSX
In-Reply-To: <04B471F6-44B2-11D9-A09D-000A95927CCE@mac.com>
References: <Pine.GSO.3.96.1041118123305.8716B-100000@radagast>
	<75597826-39B5-11D9-AF62-000A95927CCE@mac.com>
	<o1wtw2c9ox.fsf@erix.ericsson.se>
	<04B471F6-44B2-11D9-A09D-000A95927CCE@mac.com>
Message-ID: <o14qj369il.fsf@erix.ericsson.se>

Sean Hinde <sean.hinde@REDACTED> writes:

> Hi Bjorn,
> 
> I hope your change gets rid of the undocumented -lbundle1.o and
> reverts to using the mechanism in the README - namely -bundle, and
> using the cc program to link rather than ld

No, sorry.

Our Makefiles must work on all platforms (including other Unixes,
Windows, VxWorks and other embedded system). Currently, the Makefiles
use the C compiler for compiling and ld for linking.

Starting to use the C compiler for linking will probably not work
for some platform, so that is not an option.

The best way to solve the problem would of course be to choose whether to use
gcc or ld for linking on a platform-by-platform basis, but that would
require changes to configure.in and many Makefiles in source code and a lot of
time for testing.

This is not something we can do for a patch release. Maybe for R11B,
if it gets high enough priority.

/Bjorn

> 
> Quote from README:
> 
> If you develop linked in drivers (shared library) you need to link
> using "cc" and the flags "-bundle -flat_namespace -undefined
> suppress". You also include "-fno-common" in CFLAGS when
> compiling. Use ".so" as the library suffix.
> 
> Sean
> 
> 
> On 1 Dec 2004, at 15:36, Bjorn Gustavsson wrote:
> 
> > Actually, we DID get that part correct in R10B (getting rid of -R
> > for Mac OS X).
> >
> > Unfortunately, there is another problem... which I think I have
> > solved now. If my change doesn't break linking on some other platform,
> > it shold appear in R10B-2.
> >
> > /Bjorn
> >
> > Sean Hinde <sean.hinde@REDACTED> writes:
> >
> >> I have submitted a patch for this more times than I care to remember.
> >>
> >> In addition the correct mechanism is described in Erlang/OTPs very own
> >> README file.
> >>
> >> Please try to get this right for OS X next time.
> >>
> >> Sean
> >>
> >> On 18 Nov 2004, at 11:34, Peter H|gfeldt wrote:
> >>
> >>>
> >>> Just remove the -R switch from the Makefile. In R10B the -R switch is
> >>> not included for OSX (or is it called Darwin?).
> >>>
> >>> /Peter
> >>>
> >>>
> >>> On Wed, 17 Nov 2004, Reto Kramer wrote:
> >>>
> >>>> I'm trying to build the R9C-2 ssl and crypto under OSX 10.3.4. Can
> >>>> some
> >>>> BSD/Mac guru out there help me figure out how to change the ld
> >>>> switches
> >>>> (under OSX, the R option of the -Wl switch is not available).
> >>>>
> >>>> Thanks,
> >>>> - Reto
> >>>>
> >>>> from: otp_src_R9C-2/otp_src_R9C-2/lib/crypto/c_src/Makefile.in
> >>>>
> >>>> $(LIBDIR)/crypto_drv.so: $(OBJS)
> >>>> 	$(INSTALL_DIR) $(LIBDIR)
> >>>> 	$(CC) $(LDFLAGS) -L$(SSL_LIBDIR) -Wl,-R$(SSL_LIBDIR) \
> >>>> 			-o $@ $^ $(LDLIBS) -lcrypto
> >>>>
> >>>>
> >>>> === Entering application crypto
> >>>> make[3]: Nothing to be done for `opt'.
> >>>> make -f powerpc-apple-darwin7.5.0/Makefile TYPE=opt
> >>>> /usr/bin/install -c -d ../priv/lib/powerpc-apple-darwin7.5.0
> >>>> gcc -bundle -flat_namespace -undefined suppress -lbundle1.o
> >>>> -L/usr/lib
> >>>> -Wl,-R/usr/lib \
> >>>> 		 -o ../priv/lib/powerpc-apple-darwin7.5.0/crypto_drv.so
> >>>> ../priv/obj/powerpc-apple-darwin7.5.0/crypto_drv.o  -lcrypto
> >>>> ld: unknown flag: -R/usr/lib
> >>>> === Skipping subdir doc/src, it is missing
> >>>> === Leaving application crypto
> >>>>
> >>>> #
> >>>>
> >>>
> >>
> >
> > -- 
> > Bj?rn Gustavsson, Erlang/OTP, Ericsson AB
> 

-- 
Bj?rn Gustavsson, Erlang/OTP, Ericsson AB


From erik.ej.reitsma@REDACTED  Fri Dec  3 10:27:53 2004
From: erik.ej.reitsma@REDACTED (Erik Reitsma EJ (RY/ETM))
Date: Fri, 3 Dec 2004 10:27:53 +0100 
Subject: Finding unique elements of a list
Message-ID: <440A2703A54A8F4FB2AC2AE34F27129D05495B02@ESEALNT889.al.sw.ericsson.se>

So what you are saying is, that it is a simple one-liner?

remove_dups(L) ->
 [F||{_,F}<-lists:sort(lists:foldl(fun({H,K},[{_,H}|_]=OL) -> OL; ({H,K},[{K1,H1}|T]) -> [{K,H},{K1,H1}|T]; ({H,K},[]) -> [{K,H}] end,[],lists:sort(element(1,lists:foldl(fun(E,{ResN,CountN}) -> {[{E,CountN}|ResN],CountN+1} end,{[],0},L)))))].

or, not on one line:

remove_dups(L) -> 
   [F||{_,F}<-
      lists:sort(
	   lists:foldl(
	      fun({H,K},[{_,H}|_]=OL) -> 
               OL;
            ({H,K},[{K1,H1}|T]) -> 
               [{K,H},{K1,H1}|T]; 
            ({H,K},[]) -> 
               [{K,H}] 
            end,
            [],
            lists:sort(
               element(
                  1,
                  lists:foldl(
                     fun(E,{ResN,CountN}) -> 
                        {[{E,CountN}|ResN],CountN+1}
                     end,
                     {[],0},
                     L)
               )
            )
         )
      )
   ].


:)

> If you want to remove duplicate elements from a list but otherwise
> preserve the order of the elements, so that, e.g., 
> 
>     remove_dups([3,1,4,1,5,9,2,6,5,3,5,8,9]) -> [3,1,4,5,9,2,6,8]
> 
> then there is an O(N.lgN) method which goes like this:
>     (1) pair each element with its position number in the list
> 	O(N)
>     (2) sort the list of pairs (using whole pairs as keys)
> 	O(N.lgN)
>     (3) remove adjacent duplicates
> 	O(N)
>     (4) switch the {Elem,Posn} pairs to {Posn,Elem} pairs
> 	O(U) where 1 <= U <= N, U is the number of unique elements
>     (5) sort the list of {Posn,Elem} pairs
> 	O(U.lgU)
>     (6) strip off the positions, {Posn,Elem} -> Elem



*Erik.               
> 


From ingela@REDACTED  Fri Dec  3 11:18:11 2004
From: ingela@REDACTED (Ingela Anderton)
Date: Fri, 3 Dec 2004 11:18:11 +0100
Subject: Inets 4.0 and HTTP/1.1 "100 Continue" responses
References: <CE32AABA6936B0459B3EBE99331FC75D09E06A@server.Cellfind.local>
Message-ID: <16816.15715.204225.428616@gargle.gargle.HOWL>


Heinrich Venter wrote:
> It looks like I have managed to fix the problem with the 100 Continue
> response lines
Well not really! However I will fix it for the next open source release.

> The idea is that if the "100 Continue" response is recognised, the Body
> is re-parsed from scratch.  I am not sure if all the extra response
> handling in the parsing case is neccesary.
> I also don't know if the behavior of the header "expect=100-continue"
> will stil be correct.
I do not agree with this idea! When you receive a 100-continue
response the body should be sent if the request had a
expect:100-continue header and then we wait for the next response from
the server or if not we ignore this response and wait for the next
response.
 
> %%% RFC2616, Section 10.1.1
> %%% Note:
> %%% - Only act on the 100 status if the request included the
> %%%   "Expect:100-continue" header, otherwise just ignore this response.
> status_continue(Req = #http_request_h{expect="100-continue"}, Session)
> ->
>     {_, Body} = Req#request.content,
>     http_transport:send(Session#tcp_session.scheme, 
> 			Session#tcp_session.socket, Body),
>     {ok, continue};
> 
> status_continue(_Req, _Session) ->
>     error_logger:info_msg("Found Continue"),
>     {ok, continue}.
You are however right that this function did not have a correct return
value. In fact this could not ever have worked as intended as the first
argument is not of type http_request_h! (I think it might be an
unfortunate mistake made when evolving the the old unsupported
client). Alas testing is of the nature that it can always be improved!

-- 
/Ingela

Ericsson AB - OTP team








From peppe@REDACTED  Thu Dec  2 16:33:02 2004
From: peppe@REDACTED (UAB L/K Peter Andersson)
Date: Thu, 02 Dec 2004 16:33:02 +0100
Subject: risky_shell
References: <37FB7AA6F5F9814FB634A7BF4C35A6F54029BA@ESEALNT442.al.sw.ericsson.se>
Message-ID: <41AF35AE.9EF07908@erix.ericsson.se>


Uffe,

The release note says: "The possibility to start the erlang shell in
parallell with the rest of the system has been reintroduced for
backwards compatibility. Note that this old behaviour is error prone and
should not be used unless for some reason necessary."

By being able to access the system from the shell early during startup,
it is easy to crash the system by mistake (since you can't tell what
state the system is in). We consider this an error and so in R10 we sync
the shell process with init so that it doesn't start until the system is
in a safe state.

You (i.e. AXD-301) requested we reintroduce the parallell pre-R10
behaviour (not as default, but to be activated by means of a flag). We
really didn't want to do this, but agreed because you insisted. Our
condition was that we could keep this an *undocumented* "feature" for
AXD-301 use only, hopefully (to be removed later when we'd implemented
better support for "early tracing"). You agreed you'd use this on your
own risk. ;-)

We have not documented the details because we don't want people to use
this flag and, like you, become dependent on the old shell startup
behaviour. So don't push me or I'll take away this flag of yours! :-)

The old parallell startup behaviour *could* be useful to start tracing
early. But since it is indeed parallell and the time to start the shell
process depends on e.g. io speed, it must be an unpredictable way to
start tracing an application start sequence.

  /Peppe


"Ulf Wiger (AL/EAB)" wrote:
> 
> Am I just blind, or is the command-line flag -risky_shell undocumented?
> 
> The stdlib release notes for OTP R10B_1 mention that there is a possibility to start the Erlang shell in parallel with the rest of the system, but there's no explanation of how this is done (fortunately, there's always the source code... :-)
> 
> Perhaps at least the release notes could mention the name of the flag that makes this possible, if you don't want the 'erl' documentation in ERTS to mention it?
> 
> The idea behind re-introducing this "bug" was that it is sometimes very important to be able to trace on the application start sequence. One convenient way of doing this is to simply enter calls to e.g. dbg in the shell.
> 
> /Uffe


From keymon@REDACTED  Fri Dec  3 12:49:45 2004
From: keymon@REDACTED (Hector Rivas Gandara)
Date: Fri, 03 Dec 2004 12:49:45 +0100
Subject: invalid data on distribution channel, offending packet
Message-ID: <41B052D9.8030108@wanadoo.es>


Hi,

I'm working with EI library, using it in C++, but I have some problems:

I'm using ei_rpc function family, and I'm trying to call this funcion in 
the erlang node:

-module(replyserv).
-export[imprime/1]).
imprime(X) ->
   io:format("print: ~w~n",[X]),
   {echo, X}.

I send the rpc (using ei_rpc_to or ei_rpc), but it does not call the 
function (No print), and fails in ei_rpc_from (if using ei_rpc_to), 
printint the following message in the erlang console:

Got invalid data on distribution channel, offending packet is: 
<<112,131,104,4,97,6,103,100,0,24,112,101,112,105,116,111,64,116,114,97,110,116,111,114,46,108,102,99,105,97,46,112,114,105,0,0,0,3,0,0,0,0,0,100,0,0,100,0,3,114,101,120,131,104,2,103,100,0,24,112,101,112,105,116,111,64,116,114,97,110,116,111,114,46,108,102,99,105,97,46,112,114,105,0,0,0,3,0,0,0,0,0,104,5,100,0,4,99,97,108,108,100,0,9,114,101,112,108,121,115,101,114,118,100,0,7,105,109,112,114,105,109,101,131,108,0,0,0,1,100,0,4,72,111,108,97,106,100,0,4,117,115,101,114>>

Setting the EI_TRACELEVEL=100 the output is:
ei_xconnect: Fri Dec  3 12:38:48 2004: -> CONNECT attempt to connect to 
pepita
ei_epmd_r4_port: Fri Dec  3 12:38:48 2004: -> PORT2_REQ alive=pepita 
ip=192.168.41.68
ei_epmd_r4_port: Fri Dec  3 12:38:48 2004: <- PORT2_RESP result=0 (ok)
ei_epmd_r4_port: Fri Dec  3 12:38:48 2004:    port=46537 ntype=77 
proto=0 dist-high=5 dist-low=5
ei_xconnect: Fri Dec  3 12:38:49 2004: -> CONNECT connected to remote
recv_status: Fri Dec  3 12:38:49 2004: <- RECV_STATUS (ok)
recv_challenge: Fri Dec  3 12:38:49 2004: <- RECV_CHALLENGE (ok) node = 
pepita@REDACTED, version = 5, flags = 254, challenge = -1457110599
send_challenge_reply: Fri Dec  3 12:38:49 2004: -> SEND_CHALLENGE_REPLY 
(ok) challenge = -1143731390, digest = 81600edc
recv_challenge_ack: Fri Dec  3 12:38:49 2004: <- RECV_CHALLENGE_ACK (ok) 
digest = 7c0557fb
ei_xconnect: Fri Dec  3 12:38:49 2004: -> CONNECT (ok) remote = pepita
-> REG_SEND From: #Pid<pepito@REDACTED> To: rex
    {#Pid<pepito@REDACTED>, {call,  replyserv, imprime, 
[hola], user}}
<-- failed ei_rpc_from -->

What does that mean? Where is the problem?

I've only found a reference a google, and the binary stream seems to be 
the serialized message.

I also get that message **sometimes** with simple send/receive, but that 
is not deterministic: If I change the name of a variable, it works :?

-- 
Saudos
H?ctor


From sean.hinde@REDACTED  Fri Dec  3 13:59:40 2004
From: sean.hinde@REDACTED (Sean Hinde)
Date: Fri, 3 Dec 2004 12:59:40 +0000
Subject: crypto and ssl modules under OSX
In-Reply-To: <o14qj369il.fsf@erix.ericsson.se>
References: <Pine.GSO.3.96.1041118123305.8716B-100000@radagast> <75597826-39B5-11D9-AF62-000A95927CCE@mac.com> <o1wtw2c9ox.fsf@erix.ericsson.se> <04B471F6-44B2-11D9-A09D-000A95927CCE@mac.com> <o14qj369il.fsf@erix.ericsson.se>
Message-ID: <31DE51EA-452B-11D9-A09D-000A95927CCE@mac.com>


On 3 Dec 2004, at 09:01, Bjorn Gustavsson wrote:

> Sean Hinde <sean.hinde@REDACTED> writes:
>
>> Hi Bjorn,
>>
>> I hope your change gets rid of the undocumented -lbundle1.o and
>> reverts to using the mechanism in the README - namely -bundle, and
>> using the cc program to link rather than ld
>
> No, sorry.
>
> Our Makefiles must work on all platforms (including other Unixes,
> Windows, VxWorks and other embedded system). Currently, the Makefiles
> use the C compiler for compiling and ld for linking.
>
> Starting to use the C compiler for linking will probably not work
> for some platform, so that is not an option.
>
> The best way to solve the problem would of course be to choose whether 
> to use
> gcc or ld for linking on a platform-by-platform basis, but that would
> require changes to configure.in and many Makefiles in source code and 
> a lot of
> time for testing.

I don't understand this - there is already an explicit rule done in 
configure.in for darwin. You already have a DED_LD substitution for 
precisely this purpose. The patch is trivial:

--- configure.in.orig   Fri Dec  3 10:17:19 2004
+++ configure.in        Fri Dec  3 10:18:13 2004
@@ -1162,8 +1162,9 @@
                 darwin*)
                         # Mach-O linker: a shared lib and a loadable
                         # object file is not the same thing.
-                       DED_LDFLAGS="-bundle -flat_namespace -undefined 
suppress -lbundle1.o"
+                       DED_LDFLAGS="-bundle -flat_namespace -undefined 
suppress"
                         DED_CFLAGS="$DED_CFLAGS -fno-common"
+                        DED_LD="$CC"
                 ;;
                 *)
                         # assume GNU linker and ELF

So long as DED_LD is used for linking the various drivers then nothing 
will be broken - everything else will use the default value of ld. I 
provided a complete patch set sometime ago to do just this.

Equally I don't get the reluctance to tidy up or make any changes to 
the build procedure. You have a set of Officially Supported platforms, 
which you presumably can and do test every night in a fully automated 
way.

Unsupported platforms are just that - if you break the build for some 
ancient version of some weird and wonderful UNIX toolchain as part of a 
tidy up then the community will provide patches if they care enough. If 
it becomes broken and no-one notices then that suggests that no-one is 
using that platform anyway.

I see this in action most releases - the *BSD crew submit patches, and 
Redhat Latest always seems to break something. No-one minds too much, 
or if they do they either provide a fix it or use a supported platform.

You guys should be able to make more use of the open source community, 
and support of "non-supported" platforms would seem to be a good start.

<end_soapbox/>

Sean



From ingela@REDACTED  Fri Dec  3 14:09:41 2004
From: ingela@REDACTED (Ingela Anderton)
Date: Fri, 3 Dec 2004 14:09:41 +0100
Subject: Inets 4.0 and HTTP/1.1 "100 Continue" responses
References: <CE32AABA6936B0459B3EBE99331FC75D09E091@server.Cellfind.local>
Message-ID: <16816.26005.832595.884704@gargle.gargle.HOWL>


Heinrich Venter wrote:
> >When you receive a 100-continue response the body should be sent if the
> request had a
> >expect:100-continue header and then we wait for the next response from
> the server
> Do I understand correctly that the received Body is sent back to the
> server?  If this is the correct behavior, then my suggested change would
> be wrong for this case.
No, not the received body. The body of the request that was not sent
in the first place, instead the client sent an except header so that
the server will send a 100-continue response if it will accept the
request. This can be handy if the client has a big body that it wants
to send and it does not want to send it if the server will reject the
request. In this case the next response will not be sent by the
server until it receives the body from the client....

> >or if not we ignore this response and wait for the next response.
> Agreed, but here is the snag.  What I saw in live tests is the Body of a
> 100-continue response actually contained the next complete received
> response, which was a 200.  The reason for this is simply that the
> buffer received from the socket contained more than one complete
> response (100 and 200).
...however in the case that the server sends a 100-continue and
there was no expect-header the next response will come right away.

> The assumption that every response will have its own socket event seems
> to be wrong.  If the 100-continue (and its supposedly empty Body) is
I have made no such assumption! (However I missed to think about the
condition mentioned above)

-- 
/Ingela

Ericsson AB - OTP team







From joe@REDACTED  Fri Dec  3 14:37:27 2004
From: joe@REDACTED (Joe Armstrong)
Date: Fri, 3 Dec 2004 14:37:27 +0100 (CET)
Subject: Finding unique elements of a list
In-Reply-To: <20041202102242.38694.qmail@web41902.mail.yahoo.com>
References: <20041202102242.38694.qmail@web41902.mail.yahoo.com>
Message-ID: <Pine.LNX.4.61.0412031429010.26034@enfield.sics.se>


The optimization leads to (with renaming)

unique(Xs) ->
     remove_consequative(lists:sort(Xs)).

remove_consequative([H,H|T]) ->
     remove_consequative([H|T]);
remove_consequative([H|T]) ->
     [H|remove_consequative(T)];
remove_consequative([]) ->
     [].


   (Now this is  the way I wrote  this years ago - seems  like my brain
has an optimizing partical evaluator built-in).

   << extra prize - can you write a program which automatically transforms
the latter program to the former? >>

<<aside - question is this way of writing things even better
   or is the compiler smart enough to recognise that [H|T] does
not need rebuilding?

remove_consequative([H,X=[H|T]]) ->
     remove_consequative(X);
remove_consequative([H|T]) ->
     [H|remove_consequative(T)];
remove_consequative([]) ->
     [].

>>

   /Joe



On Thu, 2 Dec 2004, Thomas Lindgren wrote:

>
> Here is a straightforward version with good O(.):
> first sort the list, then drop consecutive duplicates.
>
> Complexity is given by the sort (assuming unit cost of
> comparing elements). O(N log N), presumably?
>
> unique(Xs) ->
>    rem_dups(lists:sort(Xs)).
>
> rem_dups([X|Xs]) ->
>    [X|drop(X, Xs)];
> rem_dups([]) ->
>    [].
>
> drop(X, [X|Xs]) ->
>    drop(X, Xs);
> drop(X, Ys) ->
>    rem_dups(Ys).
>
> The code above seems to work, but could be optimized
> further; which is left as an exercise to the reader.
>
> Best,
> Thomas
>
>
>
>
> __________________________________
> Do you Yahoo!?
> Meet the all-new My Yahoo! - Try it today!
> http://my.yahoo.com
>
>


From bjorn@REDACTED  Fri Dec  3 15:06:56 2004
From: bjorn@REDACTED (Bjorn Gustavsson)
Date: 03 Dec 2004 15:06:56 +0100
Subject: crypto and ssl modules under OSX
In-Reply-To: <31DE51EA-452B-11D9-A09D-000A95927CCE@mac.com>
References: <Pine.GSO.3.96.1041118123305.8716B-100000@radagast>
	<75597826-39B5-11D9-AF62-000A95927CCE@mac.com>
	<o1wtw2c9ox.fsf@erix.ericsson.se>
	<04B471F6-44B2-11D9-A09D-000A95927CCE@mac.com>
	<o14qj369il.fsf@erix.ericsson.se>
	<31DE51EA-452B-11D9-A09D-000A95927CCE@mac.com>
Message-ID: <o1oehbwk5b.fsf@erix.ericsson.se>

OK. You are right that there already exists a mechanism
for choosing whether to link using the compiler or the linker.
I had forgotten that.

I'll try to have another look before we release R10B-2.

Sean Hinde <sean.hinde@REDACTED> writes:
[...]
> 
> You guys should be able to make more use of the open source community,
> and support of "non-supported" platforms would seem to be a good start.

Actually, we have spent more time lately on getting the non-supported
platforms to build cleaner and work better.

In R10B-2 crypto will build on Mac OS X, although it may still use the
old ugly, undocument bundle1.o trick. At the very latest, I hope to eliminate
that trick before Tiger is out.

/Bjorn
-- 
Bj?rn Gustavsson, Erlang/OTP, Ericsson AB


From sean.hinde@REDACTED  Fri Dec  3 15:37:44 2004
From: sean.hinde@REDACTED (Sean Hinde)
Date: Fri, 3 Dec 2004 14:37:44 +0000
Subject: crypto and ssl modules under OSX
In-Reply-To: <o1oehbwk5b.fsf@erix.ericsson.se>
References: <Pine.GSO.3.96.1041118123305.8716B-100000@radagast> <75597826-39B5-11D9-AF62-000A95927CCE@mac.com> <o1wtw2c9ox.fsf@erix.ericsson.se> <04B471F6-44B2-11D9-A09D-000A95927CCE@mac.com> <o14qj369il.fsf@erix.ericsson.se> <31DE51EA-452B-11D9-A09D-000A95927CCE@mac.com> <o1oehbwk5b.fsf@erix.ericsson.se>
Message-ID: <E4EABD8E-4538-11D9-A09D-000A95927CCE@mac.com>


On 3 Dec 2004, at 14:06, Bjorn Gustavsson wrote:

> I'll try to have another look before we release R10B-2.

Thanks - we have quite a few people in development and test who use 
macs now and I'm forever being asked to get things working for them.

> In R10B-2 crypto will build on Mac OS X, although it may still use the
> old ugly, undocument bundle1.o trick. At the very latest, I hope to 
> eliminate
> that trick before Tiger is out.

I think this is easy to get rid of - it is simply unnecessary if 
-bundle is used with cc :-)

Sean



From tony@REDACTED  Sat Dec  4 10:19:56 2004
From: tony@REDACTED (Tony Rogvall)
Date: Sat, 4 Dec 2004 10:19:56 +0100
Subject: invalid data on distribution channel, offending packet
In-Reply-To: <41B052D9.8030108@wanadoo.es>
References: <41B052D9.8030108@wanadoo.es>
Message-ID: <AA2623F0-45D5-11D9-BFCB-000A95EB4F88@rogvall.com>


2004-12-03 kl. 12.49 skrev Hector Rivas Gandara:

>
> Hi,
>
> I'm working with EI library, using it in C++, but I have some problems:
>
> I'm using ei_rpc function family, and I'm trying to call this funcion  
> in the erlang node:
>
> -module(replyserv).
> -export[imprime/1]).
> imprime(X) ->
>   io:format("print: ~w~n",[X]),
>   {echo, X}.
>
> I send the rpc (using ei_rpc_to or ei_rpc), but it does not call the  
> function (No print), and fails in ei_rpc_from (if using ei_rpc_to),  
> printint the following message in the erlang console:
>
> Got invalid data on distribution channel, offending packet is:  
> <<112,131,104,4,97,6,103,100,0,24,112,101,112,105,116,111,64,116,114,97 
> ,110,116,111,114,46,108,102,99,105,97,46,112,114,105,0,0,0,3,0,0,0,0,0, 
> 100,0,0,100,0,3,114,101,120,131,104,2,103,100,0,24,112,101,112,105,116, 
> 111,64,116,114,97,110,116,111,114,46,108,102,99,105,97,46,112,114,105,0 
> ,0,0,3,0,0,0,0,0,104,5,100,0,4,99,97,108,108,100,0,9,114,101,112,108,12 
> 1,115,101,114,118,100,0,7,105,109,112,114,105,109,101,131,108,0,0,0,1,1 
> 00,0,4,72,111,108,97,106,100,0,4,117,115,101,114>>
>

The message is not properly aligned 131 is the "version magic"  and  
should be the first byte,
this mens that there is a problem with the message sending, I think I  
answered to a similar problem
recently. Check you C code (are you using threads?)

Good luck

/Tony



From ernie.makris@REDACTED  Sat Dec  4 23:24:43 2004
From: ernie.makris@REDACTED (Ernie Makris)
Date: Sat, 4 Dec 2004 17:24:43 -0500
Subject: rb:show/1 problem
Message-ID: <000501c4da50$0e472940$4601a8c0@matrix>

>Folks,
>
>There is a problem with rb:show/1 in R10B which appears after you
>run it with an integer() argument.  It prints the report the first
>time but the next time it prints the error below.
>
> -Vance
>
>rb: ERROR! A report on bad form was encountered. It can not be printed
>to the log.
>
>Details:
>"2004-11-30 05:37:41" {{{2004,11,30},{5,37,41}},
>                       {info_report,<0.33.0>,
>                                    {<0.36.0>,
>                                     progress,
>                                     [{supervisor,{local,sasl_safe_sup}},
>                                     {started,
>                                         [{pid,<0.37.0>},
>                                           {name,alarm_handler},
>
{mfa,{alarm_handler,start_link,[]}},
>                                           {restart_type,permanent},
>                                           {shutdown,2000},
>                                           {child_type,worker}]}]}}}
>
>rb: Can not reopen standard_io. Logging aborted.

 Hello,

I too am having the same problem. R9C-1 seems to work fine. Calling rb:show
on the second time
produces the same error message as quoted above.
I looked at the code inside rb.erl and found that when handle_call is
serviced:
handle_call({show_number, Number}, _From, State) ->
    #state{dir = Dir, data = Data, device = Device, abort = Abort, log =
Log} = State,
    NewDevice = print_report(Dir, Data, Number, Device, Abort, Log),
    {reply, ok, State#state{device = NewDevice}};

The device is replaced with {proceed, standard_io}. In R9C-1 it is:
handle_call({show_number, Number}, _From, State) ->
    #state{dir = Dir, data = Data, device = Device} = State,
    catch print_report(Dir, Data, Number, Device),
    {reply, ok, State};

Notice in R9C-1 there is no Device replacement.

The problem arises when in format_lib_supp.erl:

print_header2(Device, Line, Header) ->
    Format1 = lists:concat(["~n~", Line, ".", Line, "s~n"]),
    Format2 = lists:concat(["~", Line, "c~n"]),
    io:format(Device, Format1, [Header]),
    io:format(Device, Format2, [$=]).

a call to print_header2 is made and io:format is passed {proceed,
standard_io} instead of
standard_io.

I have a fix, but I'm not sure if its proper. Here it is:
handle_call({show_number, Number}, _From, State) ->
    #state{dir = Dir, data = Data, device = Device, abort = Abort, log =
Log} = State,
    {Next, NewDevice} = print_report(Dir, Data, Number, Device, Abort, Log),
    {reply, ok, State#state{device = NewDevice}};

Since print_report now returns {XXXSOMETHINGXXX, Device}, I do the proper
match. However, I'm not
sure what should happen when {abort, Device} is returned?
If this is a bug, can it be fixed in the next release?

Thanks
Ernie



From ernie.makris@REDACTED  Sat Dec  4 23:38:18 2004
From: ernie.makris@REDACTED (Ernie Makris)
Date: Sat, 4 Dec 2004 17:38:18 -0500
Subject: rb:show/1 again
Message-ID: <000e01c4da51$f45f50f0$4601a8c0@matrix>

Hello erlangers,

One more thing about rb:show. If you try to print a non-existent report
by number, you get a badmatch since handle_call({show_number.... 
no longer calls print_report via a catch.

Thanks
Ernie
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://erlang.org/pipermail/erlang-questions/attachments/20041204/cff00892/attachment.htm>

From peppe@REDACTED  Mon Dec  6 09:32:47 2004
From: peppe@REDACTED (UAB L/K Peter Andersson)
Date: Mon, 06 Dec 2004 09:32:47 +0100
Subject: rb:show/1 again
References: <000e01c4da51$f45f50f0$4601a8c0@matrix>
Message-ID: <41B4192F.CA9BBD2F@erix.ericsson.se>


Vance & Ernie,

Thanks for reporting these problems. We'll have a look at them promptly!

Regards, Peppe

Erlang/OTP Team


> Ernie Makris wrote:
> 
> Hello erlangers,
> 
> One more thing about rb:show. If you try to print a non-existent
> report
> by number, you get a badmatch since handle_call({show_number....
> no longer calls print_report via a catch.
> 
> Thanks
> Ernie


From olivier@REDACTED  Mon Dec  6 11:22:20 2004
From: olivier@REDACTED (olivier)
Date: Mon, 06 Dec 2004 11:22:20 +0100
Subject: Broken debian package ?
Message-ID: <41B432DC.9030901@dolphian.com>

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1


	Hi,

	I've been getting a segfault for some days now, using the
debian-packaged erlang:

[...]
stat64("/usr/lib/erlang/lib/kernel-2.10.2/ebin/erl_distribution.beam",
{st_mode=S_IFREG|0644, st_size=3720, ...}) = 0
open("/usr/lib/erlang/lib/kernel-2.10.2/ebin/erl_distribution.beam",
O_RDONLY) = 7
read(7, "FOR1\0\0\16\200BEAMAtom\0\0\1\252\0\0\0,\20erl_dis"..., 3720) =
3720
close(7)                                = 0
stat64("/usr/lib/erlang/lib/kernel-2.10.2/ebin/global_group.beam",
{st_mode=S_IFREG|0644, st_size=34496, ...}) = 0
open("/usr/lib/erlang/lib/kernel-2.10.2/ebin/global_group.beam",
O_RDONLY) = 7
read(7, "FOR1\0\0\206\270BEAMAtom\0\0\7\223\0\0\0\255\fglobal_"...,
34496) = 34496
close(7)                                = 0
- --- SIGCHLD (Child exited) @ 0 (0) ---
- --- SIGSEGV (Segmentation fault) @ 0 (0) ---
+++ killed by SIGSEGV (core dumped) +++

	Today's package is: erlang-base    10.b.1a-2

	Just wondering if this happens to someone else..

	Cheers,

- --
Olivier
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.5 (GNU/Linux)
Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org

iD8DBQFBtDLcpqVXaJzJYNIRAnEGAJwMlKP7elWZLf1TOkAj4TOfSopKWACeMo9R
1/AniATrYai8Q/PSsx7JSKM=
=IKJ4
-----END PGP SIGNATURE-----


From bry@REDACTED  Mon Dec  6 12:03:22 2004
From: bry@REDACTED (bry@REDACTED)
Date: Mon,  6 Dec 2004 12:03:22 +0100
Subject: repos download slow
Message-ID: <1102331002.41b43c7aac419@horde.scannet.dk>



Whenever I go to download Repos it is so damn slow I decide to forego the
experience, this is over ADSL. Anybody else been complaining about this?




From mickael.remond@REDACTED  Mon Dec  6 12:20:55 2004
From: mickael.remond@REDACTED (Mickael Remond)
Date: Mon, 06 Dec 2004 12:20:55 +0100
Subject: repos download slow
In-Reply-To: <1102331002.41b43c7aac419@horde.scannet.dk>
References: <1102331002.41b43c7aac419@horde.scannet.dk>
Message-ID: <41B44097.1000104@erlang-fr.org>

bry@REDACTED wrote:
> 
> Whenever I go to download Repos it is so damn slow I decide to forego the
> experience, this is over ADSL. Anybody else been complaining about this?

It seems slow right now.
The interest in Erlang-projects and repos have increased recently.
We will try to provide a bittorrent download soon to overcome those 
problems. Would this be ok for you ?

-- 
Micka?l R?mond
  http://www.erlang-projects.org/


From ernie.makris@REDACTED  Mon Dec  6 12:31:16 2004
From: ernie.makris@REDACTED (Ernie Makris)
Date: Mon, 6 Dec 2004 06:31:16 -0500
Subject: repos download slow
References: <1102331002.41b43c7aac419@horde.scannet.dk>
Message-ID: <001501c4db87$1a868d30$4601a8c0@matrix>

Hello,

It has been slow for me also, from multiple locations.
It takes about an hour and a half to download from either a cable modem or a
T1 connection
that I have access to.

Thanks
Ernie
----- Original Message ----- 
From: <bry@REDACTED>
To: <erlang-questions@REDACTED>
Sent: Monday, December 06, 2004 6:03 AM
Subject: repos download slow


>
>
> Whenever I go to download Repos it is so damn slow I decide to forego the
> experience, this is over ADSL. Anybody else been complaining about this?
>
>



From erlang@REDACTED  Mon Dec  6 13:57:55 2004
From: erlang@REDACTED (Peter L)
Date: Mon, 6 Dec 2004 13:57:55 +0100 (CET)
Subject: OS X patch for run_erl.c
Message-ID: <51370.193.15.249.14.1102337875.squirrel@www69.webcows.se>

This Sunday I discovered that my FreeBSD 4.3 had exactly the same
problem as Sean had in his OS X, with mknod/mkfifo in the run_erl.c
file. When I applied the same patch as Sean did I was able to start
erlang as a daemon, which I had not been able to do before on FreeBSD.

Looking at this file run_erl.c in the latest R10B-1a version of
OTP/Erlang, I was still unable to see any fix for this problem,
despite it being well over a year since Sean reported the problem
the first time.

Is it in Erlang or in FreeBSD / OS X, where the problem needs to be
fixed? Because I guess that Erlang is supposed to support also OS X
and FreeBSD...

A maybe related problem I had, was that I was not able to attach
properly to the erlang daemon with the supplied bin/to_erl. Is there
perhaps another fix that has to be applied for this to work under
FreeBSD or OS X? Sean, any idea?

/Peter

%% ============ From the erlang list archive ===============

    * To: Bengt Kleberg
    * Subject: Re: OS X patch for run_erl.c
    * From: Sean Hinde
    * Date: Fri, 19 Sep 2003 23:46:54 +0100
    * Cc: Erlang Questions

On Monday, September 15, 2003, at 09:07  am, Bengt Kleberg wrote:

> Sean Hinde wrote:
>> Hi,
>> The following patch to run_erl.c is required for Erlang to run in
>> embedded mode under OS X. Without this patch run_erl fails to create
>> the
>
> ...deleted
>> +#ifdef HAVE_MACH_O_DYLD_H
>> +  if ((mkfifo(name, perm) < 0) && (errno != EEXIST))
>> +#else
>> +    if ((mknod(name, S_IFIFO | perm, 0) < 0) && (errno != EEXIST))
>> +#endif
>
> perhaps mkfifo() is deprecated (or something even worse :-), but would
> it not have been better to test like this:
>
> #ifdef HAVE_MKFIFO
> 	if ((mkfifo(name, perm) < 0) && (errno != EEXIST))
> #else
> 	if ((mknod(name, S_IFIFO | perm, 0) < 0) && (errno != EEXIST))
> #endif
>

No doubt. I leave it to the real maintainers of Erlang to take care of
elegance in the build process. All this difficult cross platform C
stuff is light years from anything I claim competence in, beyond the
annoying necessity to occasionally waste many hours making Erlang work
on my favourite laptop :)

Sean


-- 
Peter Lund
Web: http://lundata.se
Tel: +46 70 543 9416


From bjorn@REDACTED  Mon Dec  6 14:14:23 2004
From: bjorn@REDACTED (Bjorn Gustavsson)
Date: 06 Dec 2004 14:14:23 +0100
Subject: OS X patch for run_erl.c
In-Reply-To: <51370.193.15.249.14.1102337875.squirrel@www69.webcows.se>
References: <51370.193.15.249.14.1102337875.squirrel@www69.webcows.se>
Message-ID: <o1653fh8ls.fsf@erix.ericsson.se>

Thanks!

We have look at this issue. I can't promise we'll be able to do that
for R10B-2.

/Bjorn

"Peter L" <erlang@REDACTED> writes:

> This Sunday I discovered that my FreeBSD 4.3 had exactly the same
> problem as Sean had in his OS X, with mknod/mkfifo in the run_erl.c
> file. When I applied the same patch as Sean did I was able to start
> erlang as a daemon, which I had not been able to do before on FreeBSD.
> 
> Looking at this file run_erl.c in the latest R10B-1a version of
> OTP/Erlang, I was still unable to see any fix for this problem,
> despite it being well over a year since Sean reported the problem
> the first time.
> 
> Is it in Erlang or in FreeBSD / OS X, where the problem needs to be
> fixed? Because I guess that Erlang is supposed to support also OS X
> and FreeBSD...
> 
> A maybe related problem I had, was that I was not able to attach
> properly to the erlang daemon with the supplied bin/to_erl. Is there
> perhaps another fix that has to be applied for this to work under
> FreeBSD or OS X? Sean, any idea?
> 
> /Peter
> 
> %% ============ From the erlang list archive ===============
> 
>     * To: Bengt Kleberg
>     * Subject: Re: OS X patch for run_erl.c
>     * From: Sean Hinde
>     * Date: Fri, 19 Sep 2003 23:46:54 +0100
>     * Cc: Erlang Questions
> 
> On Monday, September 15, 2003, at 09:07  am, Bengt Kleberg wrote:
> 
> > Sean Hinde wrote:
> >> Hi,
> >> The following patch to run_erl.c is required for Erlang to run in
> >> embedded mode under OS X. Without this patch run_erl fails to create
> >> the
> >
> > ...deleted
> >> +#ifdef HAVE_MACH_O_DYLD_H
> >> +  if ((mkfifo(name, perm) < 0) && (errno != EEXIST))
> >> +#else
> >> +    if ((mknod(name, S_IFIFO | perm, 0) < 0) && (errno != EEXIST))
> >> +#endif
> >
> > perhaps mkfifo() is deprecated (or something even worse :-), but would
> > it not have been better to test like this:
> >
> > #ifdef HAVE_MKFIFO
> > 	if ((mkfifo(name, perm) < 0) && (errno != EEXIST))
> > #else
> > 	if ((mknod(name, S_IFIFO | perm, 0) < 0) && (errno != EEXIST))
> > #endif
> >
> 
> No doubt. I leave it to the real maintainers of Erlang to take care of
> elegance in the build process. All this difficult cross platform C
> stuff is light years from anything I claim competence in, beyond the
> annoying necessity to occasionally waste many hours making Erlang work
> on my favourite laptop :)
> 
> Sean
> 
> 
> -- 
> Peter Lund
> Web: http://lundata.se
> Tel: +46 70 543 9416
> 

-- 
Bj?rn Gustavsson, Erlang/OTP, Ericsson AB


From ingela@REDACTED  Mon Dec  6 15:53:14 2004
From: ingela@REDACTED (Ingela Anderton)
Date: Mon, 6 Dec 2004 15:53:14 +0100
Subject: Inets 4.0 and HTTP/1.1 "100 Continue" responses
References: <CE32AABA6936B0459B3EBE99331FC75D09E0A5@server.Cellfind.local>
Message-ID: <16820.29274.895801.726023@gargle.gargle.HOWL>


Heinrich Venter wrote:
> Ingela
> 
> >This can be handy if the client has a big body that it wants to send
> and it does not 
> >want to send it if the server will reject the request. In this case the
> next response
> >will not be sent by the server until it receives the body from the
> client....
> Ahhh, I see.  Thanks for clearing this up for me :)
> 
> >> The assumption that every response will have its own socket event 
> >> seems to be wrong.  If the 100-continue (and its supposedly empty 
> >> Body) is
> >I have made no such assumption!
> Its only the 100-continue without the expect header that breaks the
> pattern realy, because there is more than one response to one request.
> Is there a better way to handle this than re-parsing the Body?
Well after some consideration I will have to partyl change my mind ;)
and agree with you that in the case will need to reparse the body.

-- 
/Ingela

Ericsson AB - OTP team








From sean.hinde@REDACTED  Tue Dec  7 01:25:50 2004
From: sean.hinde@REDACTED (Sean Hinde)
Date: Tue, 7 Dec 2004 00:25:50 +0000
Subject: crypto and ssl modules under OSX
In-Reply-To: <o1oehbwk5b.fsf@erix.ericsson.se>
References: <Pine.GSO.3.96.1041118123305.8716B-100000@radagast> <75597826-39B5-11D9-AF62-000A95927CCE@mac.com> <o1wtw2c9ox.fsf@erix.ericsson.se> <04B471F6-44B2-11D9-A09D-000A95927CCE@mac.com> <o14qj369il.fsf@erix.ericsson.se> <31DE51EA-452B-11D9-A09D-000A95927CCE@mac.com> <o1oehbwk5b.fsf@erix.ericsson.se>
Message-ID: <8C082C72-47E6-11D9-A37E-000A95927CCE@mac.com>


On 3 Dec 2004, at 14:06, Bjorn Gustavsson wrote:
>
> In R10B-2 crypto will build on Mac OS X, although it may still use the
> old ugly, undocument bundle1.o trick. At the very latest, I hope to 
> eliminate
> that trick before Tiger is out.

Interresting..

The way I got crypto to compile and run with R10B-1a was simply to 
remove the offending -lbundle1.o from the generated Makefile. 
Everything else was correct.

This strongly suggests that the only remaining problem is the presence 
of the very thing you are trying to remove.

The solution would appear to be quite straightforward from this point.

Sean



From danie@REDACTED  Mon Dec  6 23:00:53 2004
From: danie@REDACTED (Danie Schutte)
Date: Tue, 07 Dec 2004 00:00:53 +0200
Subject: Mnesia Related
Message-ID: <1102370453.7725.12.camel@blackbox.erlfinsys.net>

Is there anywhere I can find more information regarding the following
message (or how to avoid it) :)

Mnesia('acb_application@REDACTED
acard.net'): {unpaids_bin,dcd2ets} repaired, lost 21 bad bytes

Thank you

-- 
Danie Schutte <danie@REDACTED>
Erlang Financial Systems
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://erlang.org/pipermail/erlang-questions/attachments/20041207/255da6f1/attachment.htm>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: smiley-3.png
Type: image/png
Size: 819 bytes
Desc: not available
URL: <http://erlang.org/pipermail/erlang-questions/attachments/20041207/255da6f1/attachment.png>

From luke@REDACTED  Tue Dec  7 08:48:54 2004
From: luke@REDACTED (Luke Gorrie)
Date: 07 Dec 2004 08:48:54 +0100
Subject: Finding unique elements of a list
References: <F7698C6BAED2D41189E00002A50A625F06A07CD4@v-exchange.volition.net>
Message-ID: <l3653ea6qh.fsf@gant.netcamp.se>

James Hague <jamesh@REDACTED> writes:

> Is there a standard Erlang library function or idiom for removing duplicate
> items from an unsorted list?  Ideally the elements in the final list should
> be in the order they were passed in, except without duplicates:

I've always liked this one that Magnus Fr?berg showed me years ago:

  remove_dups([])    -> [];
  remove_dups([H|T]) -> [H | [X || X <- T, X /= H]].

That's O(n^2) on basic list operations, but my rule-of-thumb is that's
fine if it's not used in a loop and the inputs are a thousand elements
or less.

To be faster I think Vance-style with a dictionary or set to get O(N)
or O(N log N) is the way, and I think this is reasonably nice:

  remove_dups_linear(L) -> remove_dups_linear(L, gb_sets:new()).

  remove_dups_linear([], _) -> [];
  remove_dups_linear([H|T], S) ->
      case gb_sets:is_member(H, S) of
          true  -> remove_dups_linear(T, S);
          false -> [H|remove_dups_linear(T, gb_sets:insert(H, S))]
      end.

Cheers,
Luke




From bjorn@REDACTED  Tue Dec  7 10:07:46 2004
From: bjorn@REDACTED (Bjorn Gustavsson)
Date: 07 Dec 2004 10:07:46 +0100
Subject: crypto and ssl modules under OSX
In-Reply-To: <8C082C72-47E6-11D9-A37E-000A95927CCE@mac.com>
References: <Pine.GSO.3.96.1041118123305.8716B-100000@radagast>
	<75597826-39B5-11D9-AF62-000A95927CCE@mac.com>
	<o1wtw2c9ox.fsf@erix.ericsson.se>
	<04B471F6-44B2-11D9-A09D-000A95927CCE@mac.com>
	<o14qj369il.fsf@erix.ericsson.se>
	<31DE51EA-452B-11D9-A09D-000A95927CCE@mac.com>
	<o1oehbwk5b.fsf@erix.ericsson.se>
	<8C082C72-47E6-11D9-A37E-000A95927CCE@mac.com>
Message-ID: <o1hdmybhnh.fsf@erix.ericsson.se>

If you link using gcc, -lbundle1.o must NOT be present
because ggc automatically links bundle1.o and it gets
linked TWICE (and causes a duplicate symbol error).

If you link using ld, -lbundle1.o MUST be present.

The crypto Makefile was wrong. It explicitly linked
using gcc, but it still included the other DED variables,
so the -lbundle1.o flag was included too. Clearly wrong.

Now the crypto Makefile links using @DED_LD@ (which is
currently set to 'ld') and it works.


The next step, to change DED_LD to gcc for Mac OS X and remove
the -lbundle1.o flag is easy and should work, PROVIDED that all
Makefiles in all applications that link dynamic drivers are
correct (in particular, they should not use any explicit options
that only work in either the compiler or linker). I am not sure
they are correct. Yet.

/Bjorn

Sean Hinde <sean.hinde@REDACTED> writes:

> On 3 Dec 2004, at 14:06, Bjorn Gustavsson wrote:
> >
> > In R10B-2 crypto will build on Mac OS X, although it may still use the
> > old ugly, undocument bundle1.o trick. At the very latest, I hope to
> > eliminate
> > that trick before Tiger is out.
> 
> Interresting..
> 
> The way I got crypto to compile and run with R10B-1a was simply to
> remove the offending -lbundle1.o from the generated
> Makefile. Everything else was correct.
> 
> This strongly suggests that the only remaining problem is the presence
> of the very thing you are trying to remove.
> 
> The solution would appear to be quite straightforward from this point.
> 
> Sean
> 

-- 
Bj?rn Gustavsson, Erlang/OTP, Ericsson AB


From hakan.stenholm@REDACTED  Tue Dec  7 10:10:55 2004
From: hakan.stenholm@REDACTED (=?ISO-8859-1?Q?Administrat=F6r?=)
Date: Tue, 7 Dec 2004 10:10:55 +0100
Subject: sending emails with erlang (for use with YXA/CPL)
Message-ID: <E66A629F-482F-11D9-BEC7-00039378D5B8@it.su.se>

Is there some easy way to get erlang to send emails or should I simply 
use os:cmd/1 and call some appropriate unix command ?
The only thing I found that might do the job (I haven't looked at it in 
depth) is email-1.0 by Joe Armstrong, in the erlang.org "User 
Contributions" section.

I would like to send emails based on mail urls e.g. :

       mailto:jones@REDACTED?subject=Lookup%20failed

i.e. urls that contain a mail address and possibly additional 
arguments, it should also be possible to attach additional data to the 
mail in some easy manner.

I'm interested in this as I'm currently implementing CPL (Call 
Processing Language - RFC 3880), for use with YXA (see 
http://www.stacken.kth.se/projekt/yxa/) and it would be nice if I could 
support the CPL mail tag, if it isn't to much work - it's currently a 
non-essential feature.



From thomas.xa.johnsson@REDACTED  Tue Dec  7 10:36:12 2004
From: thomas.xa.johnsson@REDACTED (Thomas Johnsson XA (LN/EAB))
Date: Tue, 7 Dec 2004 10:36:12 +0100 
Subject: Finding unique elements of a list
Message-ID: <A943FD84BD9ED41193460008C79180500CE0BFC0@ESEALNT419.al.sw.ericsson.se>


Luke Gorrie writes:
> I've always liked this one that Magnus Fr?berg showed me years ago:
> 
>   remove_dups([])    -> [];
>   remove_dups([H|T]) -> [H | [X || X <- T, X /= H]].

This is wrong, you probably mean

   remove_dups([])    -> [];
   remove_dups([H|T]) -> [H | [X || X <- remove_dups(T), X /= H]].

-- Thomas


From erlang@REDACTED  Tue Dec  7 11:04:52 2004
From: erlang@REDACTED (Peter L)
Date: Tue, 7 Dec 2004 11:04:52 +0100 (CET)
Subject: sending emails with erlang (for use with YXA/CPL)
In-Reply-To: <E66A629F-482F-11D9-BEC7-00039378D5B8@it.su.se>
References: <E66A629F-482F-11D9-BEC7-00039378D5B8@it.su.se>
Message-ID: <1509.193.15.249.14.1102413892.squirrel@www69.webcows.se>

There is an all erlang implementation of an SMTP client provided
by Michael Bradford in the user contrib section that I am using
myself. It works well and supports most of the things you need.

> Is there some easy way to get erlang to send emails or should I simply
> use os:cmd/1 and call some appropriate unix command ?
> The only thing I found that might do the job (I haven't looked at it in
> depth) is email-1.0 by Joe Armstrong, in the erlang.org "User
> Contributions" section.
>
> I would like to send emails based on mail urls e.g. :
>
>        mailto:jones@REDACTED?subject=Lookup%20failed
>
> i.e. urls that contain a mail address and possibly additional
> arguments, it should also be possible to attach additional data to the
> mail in some easy manner.
>
> I'm interested in this as I'm currently implementing CPL (Call
> Processing Language - RFC 3880), for use with YXA (see
> http://www.stacken.kth.se/projekt/yxa/) and it would be nice if I could
> support the CPL mail tag, if it isn't to much work - it's currently a
> non-essential feature.
>


-- 
Peter Lund
Web: http://lundata.se
Tel: +46 70 543 9416



From vlad_dumitrescu@REDACTED  Tue Dec  7 11:16:50 2004
From: vlad_dumitrescu@REDACTED (Vlad Dumitrescu)
Date: Tue, 7 Dec 2004 11:16:50 +0100
Subject: sending emails with erlang (for use with YXA/CPL)
References: <E66A629F-482F-11D9-BEC7-00039378D5B8@it.su.se>
Message-ID: <BAY15-DAV1211469DA9E455FF26A2B39CB50@phx.gbl>

You can also try
http://www.csd.uu.se/ftp/mirror/erlang/user.html#smtp_client-1.1.

regards,
Vlad

----- Original Message ----- 
From: "Administrat?r" <hakan.stenholm@REDACTED>
To: "Erlang Questions List" <erlang-questions@REDACTED>
Sent: Tuesday, December 07, 2004 10:10 AM
Subject: sending emails with erlang (for use with YXA/CPL)


> Is there some easy way to get erlang to send emails or should I simply
> use os:cmd/1 and call some appropriate unix command ?
> The only thing I found that might do the job (I haven't looked at it in
> depth) is email-1.0 by Joe Armstrong, in the erlang.org "User
> Contributions" section.
>
> I would like to send emails based on mail urls e.g. :
>
>        mailto:jones@REDACTED?subject=Lookup%20failed
>
> i.e. urls that contain a mail address and possibly additional
> arguments, it should also be possible to attach additional data to the
> mail in some easy manner.
>
> I'm interested in this as I'm currently implementing CPL (Call
> Processing Language - RFC 3880), for use with YXA (see
> http://www.stacken.kth.se/projekt/yxa/) and it would be nice if I could
> support the CPL mail tag, if it isn't to much work - it's currently a
> non-essential feature.
>
>


From Jouni.Ryno@REDACTED  Tue Dec  7 11:43:59 2004
From: Jouni.Ryno@REDACTED (Jouni =?ISO-8859-1?Q?Ryn=F6?=)
Date: Tue, 07 Dec 2004 12:43:59 +0200
Subject: OS X patch for run_erl.c
In-Reply-To: <o1653fh8ls.fsf@erix.ericsson.se>
References: <51370.193.15.249.14.1102337875.squirrel@www69.webcows.se>
	 <o1653fh8ls.fsf@erix.ericsson.se>
Message-ID: <1102416239.6090.14.camel@adic>

I still have the problem with Linux 2.6 kernel and udev etc,
run_erl [19205] Tue Dec  7 12:40:01 2004
Could not open pty master

Fixed it with my own version of it, with a call to

 #include <pty.h>
 #include <utmp.h>

 {
     static char ttyname[] = "                                   ";

     if (0 == openpty(&mfd, &sfd, ttyname, NULL, NULL)) {
       *ptyslave = ttyname;
       return mfd;
     }
   }


On Mon, 2004-12-06 at 14:14 +0100, Bjorn Gustavsson wrote:
> Thanks!
> 
> We have look at this issue. I can't promise we'll be able to do that
> for R10B-2.
> 
> /Bjorn
> 
> "Peter L" <erlang@REDACTED> writes:
> 
> > This Sunday I discovered that my FreeBSD 4.3 had exactly the same
> > problem as Sean had in his OS X, with mknod/mkfifo in the run_erl.c
> > file. When I applied the same patch as Sean did I was able to start
> > erlang as a daemon, which I had not been able to do before on FreeBSD.
> > 
> > Looking at this file run_erl.c in the latest R10B-1a version of
> > OTP/Erlang, I was still unable to see any fix for this problem,
> > despite it being well over a year since Sean reported the problem
> > the first time.
> > 
> > Is it in Erlang or in FreeBSD / OS X, where the problem needs to be
> > fixed? Because I guess that Erlang is supposed to support also OS X
> > and FreeBSD...
> > 
> > A maybe related problem I had, was that I was not able to attach
> > properly to the erlang daemon with the supplied bin/to_erl. Is there
> > perhaps another fix that has to be applied for this to work under
> > FreeBSD or OS X? Sean, any idea?
> > 
> > /Peter
> > 
> > %% ============ From the erlang list archive ===============
> > 
> >     * To: Bengt Kleberg
> >     * Subject: Re: OS X patch for run_erl.c
> >     * From: Sean Hinde
> >     * Date: Fri, 19 Sep 2003 23:46:54 +0100
> >     * Cc: Erlang Questions
> > 
> > On Monday, September 15, 2003, at 09:07  am, Bengt Kleberg wrote:
> > 
> > > Sean Hinde wrote:
> > >> Hi,
> > >> The following patch to run_erl.c is required for Erlang to run in
> > >> embedded mode under OS X. Without this patch run_erl fails to create
> > >> the
> > >
> > > ...deleted
> > >> +#ifdef HAVE_MACH_O_DYLD_H
> > >> +  if ((mkfifo(name, perm) < 0) && (errno != EEXIST))
> > >> +#else
> > >> +    if ((mknod(name, S_IFIFO | perm, 0) < 0) && (errno != EEXIST))
> > >> +#endif
> > >
> > > perhaps mkfifo() is deprecated (or something even worse :-), but would
> > > it not have been better to test like this:
> > >
> > > #ifdef HAVE_MKFIFO
> > > 	if ((mkfifo(name, perm) < 0) && (errno != EEXIST))
> > > #else
> > > 	if ((mknod(name, S_IFIFO | perm, 0) < 0) && (errno != EEXIST))
> > > #endif
> > >
> > 
> > No doubt. I leave it to the real maintainers of Erlang to take care of
> > elegance in the build process. All this difficult cross platform C
> > stuff is light years from anything I claim competence in, beyond the
> > annoying necessity to occasionally waste many hours making Erlang work
> > on my favourite laptop :)
> > 
> > Sean
> > 
> > 
> > -- 
> > Peter Lund
> > Web: http://lundata.se
> > Tel: +46 70 543 9416
> > 
> 
-- 

  Jouni Ryn?                            mailto://Jouni.Ryno@REDACTED/
                                        http://www.geo.fmi.fi/~ryno/
  Finnish Meteorological Institute      http://www.fmi.fi/
  Space Research                        http://www.geo.fmi.fi/
  P.O.BOX 503                           Tel      (+358)-9-19294656
  FIN-00101 Helsinki                    FAX      (+358)-9-19294603
  Finland                               priv-GSM (+358)-50-5302903
  
  "It's just zeros and ones, it cannot be hard"




From luna@REDACTED  Tue Dec  7 12:42:05 2004
From: luna@REDACTED (Daniel Luna)
Date: Tue, 7 Dec 2004 12:42:05 +0100 (CET)
Subject: Finding unique elements of a list
In-Reply-To: <l3653ea6qh.fsf@gant.netcamp.se>
References: <F7698C6BAED2D41189E00002A50A625F06A07CD4@v-exchange.volition.net>
 <l3653ea6qh.fsf@gant.netcamp.se>
Message-ID: <Pine.LNX.4.61.0412071237310.20904@Psilocybe.Update.UU.SE>

On Tue, 7 Dec 2004, Luke Gorrie wrote:
> James Hague <jamesh@REDACTED> writes:
>> Is there a standard Erlang library function or idiom for removing duplicate
>> items from an unsorted list?  Ideally the elements in the final list should
>> be in the order they were passed in, except without duplicates:
>
> To be faster I think Vance-style with a dictionary or set to get O(N)
> or O(N log N) is the way, and I think this is reasonably nice:
>
>  remove_dups_linear(L) -> remove_dups_linear(L, gb_sets:new()).
>
>  remove_dups_linear([], _) -> [];
>  remove_dups_linear([H|T], S) ->
>      case gb_sets:is_member(H, S) of
>          true  -> remove_dups_linear(T, S);
>          false -> [H|remove_dups_linear(T, gb_sets:insert(H, S))]
>      end.

But this one isn't linear.

I assume that gb_sets:is_member is O(log(n)) as for most balanced trees.

Then the complexity is O(Nlog(N)) for all of it.

/Luna
-- 
Daniel Luna                           | Top reasons that I have a beard:
luna@REDACTED                     |  a) Laziness.
http://www.update.uu.se/~luna/        |  b) I can.
Don't look at my homepage (it stinks).|  c) I can get away with it.


From luke@REDACTED  Tue Dec  7 12:05:01 2004
From: luke@REDACTED (Luke Gorrie)
Date: 07 Dec 2004 12:05:01 +0100
Subject: Finding unique elements of a list
In-Reply-To: <Pine.LNX.4.61.0412071237310.20904@Psilocybe.Update.UU.SE>
References: <F7698C6BAED2D41189E00002A50A625F06A07CD4@v-exchange.volition.net>
	<l3653ea6qh.fsf@gant.netcamp.se>
	<Pine.LNX.4.61.0412071237310.20904@Psilocybe.Update.UU.SE>
Message-ID: <l3acsq8j36.fsf@gant.netcamp.se>

Daniel Luna <luna@REDACTED> writes:

> But this one isn't linear.

Oops. I switched from dict to gb_set after choosing the function name :-)

Cheers,
Luke



From ernie.makris@REDACTED  Tue Dec  7 14:46:15 2004
From: ernie.makris@REDACTED (Ernie Makris)
Date: Tue, 7 Dec 2004 08:46:15 -0500
Subject: OS X patch for run_erl.c
References: <51370.193.15.249.14.1102337875.squirrel@www69.webcows.se> <o1653fh8ls.fsf@erix.ericsson.se> <1102416239.6090.14.camel@adic>
Message-ID: <000a01c4dc63$225bc890$4601a8c0@matrix>

Hi Jouni,

Can you send me the full patch for run_erl.c. I too am trying to execute
run_erl under
Fedora Core 3 and having problems.

Thanks
Ernie
----- Original Message ----- 
From: "Jouni Ryn?" <Jouni.Ryno@REDACTED>
To: "Bjorn Gustavsson" <bjorn@REDACTED>
Cc: <erlang-questions@REDACTED>
Sent: Tuesday, December 07, 2004 5:43 AM
Subject: Re: OS X patch for run_erl.c


> I still have the problem with Linux 2.6 kernel and udev etc,
> run_erl [19205] Tue Dec  7 12:40:01 2004
> Could not open pty master
>
> Fixed it with my own version of it, with a call to
>
>  #include <pty.h>
>  #include <utmp.h>
>
>  {
>      static char ttyname[] = "                                   ";
>
>      if (0 == openpty(&mfd, &sfd, ttyname, NULL, NULL)) {
>        *ptyslave = ttyname;
>        return mfd;
>      }
>    }
>
>
> On Mon, 2004-12-06 at 14:14 +0100, Bjorn Gustavsson wrote:
> > Thanks!
> >
> > We have look at this issue. I can't promise we'll be able to do that
> > for R10B-2.
> >
> > /Bjorn
> >
> > "Peter L" <erlang@REDACTED> writes:
> >
> > > This Sunday I discovered that my FreeBSD 4.3 had exactly the same
> > > problem as Sean had in his OS X, with mknod/mkfifo in the run_erl.c
> > > file. When I applied the same patch as Sean did I was able to start
> > > erlang as a daemon, which I had not been able to do before on FreeBSD.
> > >
> > > Looking at this file run_erl.c in the latest R10B-1a version of
> > > OTP/Erlang, I was still unable to see any fix for this problem,
> > > despite it being well over a year since Sean reported the problem
> > > the first time.
> > >
> > > Is it in Erlang or in FreeBSD / OS X, where the problem needs to be
> > > fixed? Because I guess that Erlang is supposed to support also OS X
> > > and FreeBSD...
> > >
> > > A maybe related problem I had, was that I was not able to attach
> > > properly to the erlang daemon with the supplied bin/to_erl. Is there
> > > perhaps another fix that has to be applied for this to work under
> > > FreeBSD or OS X? Sean, any idea?
> > >
> > > /Peter
> > >
> > > %% ============ From the erlang list archive ===============
> > >
> > >     * To: Bengt Kleberg
> > >     * Subject: Re: OS X patch for run_erl.c
> > >     * From: Sean Hinde
> > >     * Date: Fri, 19 Sep 2003 23:46:54 +0100
> > >     * Cc: Erlang Questions
> > >
> > > On Monday, September 15, 2003, at 09:07  am, Bengt Kleberg wrote:
> > >
> > > > Sean Hinde wrote:
> > > >> Hi,
> > > >> The following patch to run_erl.c is required for Erlang to run in
> > > >> embedded mode under OS X. Without this patch run_erl fails to
create
> > > >> the
> > > >
> > > > ...deleted
> > > >> +#ifdef HAVE_MACH_O_DYLD_H
> > > >> +  if ((mkfifo(name, perm) < 0) && (errno != EEXIST))
> > > >> +#else
> > > >> +    if ((mknod(name, S_IFIFO | perm, 0) < 0) && (errno != EEXIST))
> > > >> +#endif
> > > >
> > > > perhaps mkfifo() is deprecated (or something even worse :-), but
would
> > > > it not have been better to test like this:
> > > >
> > > > #ifdef HAVE_MKFIFO
> > > > if ((mkfifo(name, perm) < 0) && (errno != EEXIST))
> > > > #else
> > > > if ((mknod(name, S_IFIFO | perm, 0) < 0) && (errno != EEXIST))
> > > > #endif
> > > >
> > >
> > > No doubt. I leave it to the real maintainers of Erlang to take care of
> > > elegance in the build process. All this difficult cross platform C
> > > stuff is light years from anything I claim competence in, beyond the
> > > annoying necessity to occasionally waste many hours making Erlang work
> > > on my favourite laptop :)
> > >
> > > Sean
> > >
> > >
> > > -- 
> > > Peter Lund
> > > Web: http://lundata.se
> > > Tel: +46 70 543 9416
> > >
> >
> -- 
>
>   Jouni Ryn?                            mailto://Jouni.Ryno@REDACTED/
>                                         http://www.geo.fmi.fi/~ryno/
>   Finnish Meteorological Institute      http://www.fmi.fi/
>   Space Research                        http://www.geo.fmi.fi/
>   P.O.BOX 503                           Tel      (+358)-9-19294656
>   FIN-00101 Helsinki                    FAX      (+358)-9-19294603
>   Finland                               priv-GSM (+358)-50-5302903
>
>   "It's just zeros and ones, it cannot be hard"
>
>



From ernie.makris@REDACTED  Tue Dec  7 18:57:24 2004
From: ernie.makris@REDACTED (ernie.makris@REDACTED)
Date: Tue, 07 Dec 2004 17:57:24 +0000
Subject: OS X patch for run_erl.c
Message-ID: <120720041757.17884.41B5EF0400068FEE000045DC22058844849C079D050E03D20A07029D0A@comcast.net>

Hello Jouni,

I have one addition to the patch:

{
>     static char ttyname[] = "                                   ";
> 
>     if (0 == openpty(&mfd, &sfd, ttyname, NULL, NULL)) {
>       close(sfd);    /* ADDED THIS */
>       *ptyslave = ttyname;
>       return mfd;
>     }
>   }

If the slave fd is not closed, the parent never gets the close indication and run_erl runs forever after the erl process has ended.

Thanks
Ernie


> Here's my diff (from some version .. of run_erl.c)
> 
> 47a48,49
> > #include <pty.h>
> > #include <utmp.h>
> 153c155
> < 	      program_name, (int)(Need), FILENAME_BUFSIZ));		   \
> ---
> > 	      program_name, (Need), FILENAME_BUFSIZ));			   \
> 608c610
> < static int find_next_log_num(void) {
> ---
> > static int find_next_log_num() {
> 753a756
> >   int sfd;
> 820a824,831
> >   {
> >     static char ttyname[] = "                                   ";
> > 
> >     if (0 == openpty(&mfd, &sfd, ttyname, NULL, NULL)) {
> >       *ptyslave = ttyname;
> >       return mfd;
> >     }
> >   }
> 
> On Tue, 2004-12-07 at 08:46 -0500, Ernie Makris wrote:
> > Hi Jouni,
> > 
> > Can you send me the full patch for run_erl.c. I too am trying to execute
> > run_erl under
> > Fedora Core 3 and having problems.
> > 
> > Thanks
> > Ernie
> > ----- Original Message ----- 
> > From: "Jouni Ryn?" <Jouni.Ryno@REDACTED>
> > To: "Bjorn Gustavsson" <bjorn@REDACTED>
> > Cc: <erlang-questions@REDACTED>
> > Sent: Tuesday, December 07, 2004 5:43 AM
> > Subject: Re: OS X patch for run_erl.c
> > 
> > 
> > > I still have the problem with Linux 2.6 kernel and udev etc,
> > > run_erl [19205] Tue Dec  7 12:40:01 2004
> > > Could not open pty master
> > >
> > > Fixed it with my own version of it, with a call to
> > >
> > >  #include <pty.h>
> > >  #include <utmp.h>
> > >
> > >  {
> > >      static char ttyname[] = "                                   ";
> > >
> > >      if (0 == openpty(&mfd, &sfd, ttyname, NULL, NULL)) {
> > >        *ptyslave = ttyname;
> > >        return mfd;
> > >      }
> > >    }
> > >
> > >
> > > On Mon, 2004-12-06 at 14:14 +0100, Bjorn Gustavsson wrote:
> > > > Thanks!
> > > >
> > > > We have look at this issue. I can't promise we'll be able to do that
> > > > for R10B-2.
> > > >
> > > > /Bjorn
> > > >
> > > > "Peter L" <erlang@REDACTED> writes:
> > > >
> > > > > This Sunday I discovered that my FreeBSD 4.3 had exactly the same
> > > > > problem as Sean had in his OS X, with mknod/mkfifo in the run_erl.c
> > > > > file. When I applied the same patch as Sean did I was able to start
> > > > > erlang as a daemon, which I had not been able to do before on FreeBSD.
> > > > >
> > > > > Looking at this file run_erl.c in the latest R10B-1a version of
> > > > > OTP/Erlang, I was still unable to see any fix for this problem,
> > > > > despite it being well over a year since Sean reported the problem
> > > > > the first time.
> > > > >
> > > > > Is it in Erlang or in FreeBSD / OS X, where the problem needs to be
> > > > > fixed? Because I guess that Erlang is supposed to support also OS X
> > > > > and FreeBSD...
> > > > >
> > > > > A maybe related problem I had, was that I was not able to attach
> > > > > properly to the erlang daemon with the supplied bin/to_erl. Is there
> > > > > perhaps another fix that has to be applied for this to work under
> > > > > FreeBSD or OS X? Sean, any idea?
> > > > >
> > > > > /Peter
> > > > >
> > > > > %% ============ From the erlang list archive ===============
> > > > >
> > > > >     * To: Bengt Kleberg
> > > > >     * Subject: Re: OS X patch for run_erl.c
> > > > >     * From: Sean Hinde
> > > > >     * Date: Fri, 19 Sep 2003 23:46:54 +0100
> > > > >     * Cc: Erlang Questions
> > > > >
> > > > > On Monday, September 15, 2003, at 09:07  am, Bengt Kleberg wrote:
> > > > >
> > > > > > Sean Hinde wrote:
> > > > > >> Hi,
> > > > > >> The following patch to run_erl.c is required for Erlang to run in
> > > > > >> embedded mode under OS X. Without this patch run_erl fails to
> > create
> > > > > >> the
> > > > > >
> > > > > > ...deleted
> > > > > >> +#ifdef HAVE_MACH_O_DYLD_H
> > > > > >> +  if ((mkfifo(name, perm) < 0) && (errno != EEXIST))
> > > > > >> +#else
> > > > > >> +    if ((mknod(name, S_IFIFO | perm, 0) < 0) && (errno != EEXIST))
> > > > > >> +#endif
> > > > > >
> > > > > > perhaps mkfifo() is deprecated (or something even worse :-), but
> > would
> > > > > > it not have been better to test like this:
> > > > > >
> > > > > > #ifdef HAVE_MKFIFO
> > > > > > if ((mkfifo(name, perm) < 0) && (errno != EEXIST))
> > > > > > #else
> > > > > > if ((mknod(name, S_IFIFO | perm, 0) < 0) && (errno != EEXIST))
> > > > > > #endif
> > > > > >
> > > > >
> > > > > No doubt. I leave it to the real maintainers of Erlang to take care of
> > > > > elegance in the build process. All this difficult cross platform C
> > > > > stuff is light years from anything I claim competence in, beyond the
> > > > > annoying necessity to occasionally waste many hours making Erlang work
> > > > > on my favourite laptop :)
> > > > >
> > > > > Sean
> > > > >
> > > > >
> > > > > -- 
> > > > > Peter Lund
> > > > > Web: http://lundata.se
> > > > > Tel: +46 70 543 9416
> > > > >
> > > >
> > > -- 
> > >
> > >   Jouni Ryn?                            mailto://Jouni.Ryno@REDACTED/
> > >                                         http://www.geo.fmi.fi/~ryno/
> > >   Finnish Meteorological Institute      http://www.fmi.fi/
> > >   Space Research                        http://www.geo.fmi.fi/
> > >   P.O.BOX 503                           Tel      (+358)-9-19294656
> > >   FIN-00101 Helsinki                    FAX      (+358)-9-19294603
> > >   Finland                               priv-GSM (+358)-50-5302903
> > >
> > >   "It's just zeros and ones, it cannot be hard"
> > >
> > >
> > 
> -- 
> 
>   Jouni Ryn?                            mailto://Jouni.Ryno@REDACTED/
>                                         http://www.geo.fmi.fi/~ryno/
>   Finnish Meteorological Institute      http://www.fmi.fi/
>   Space Research                        http://www.geo.fmi.fi/
>   P.O.BOX 503                           Tel      (+358)-9-19294656
>   FIN-00101 Helsinki                    FAX      (+358)-9-19294603
>   Finland                               priv-GSM (+358)-50-5302903
>   
>   "It's just zeros and ones, it cannot be hard"
> 
> 


From joe@REDACTED  Tue Dec  7 19:26:13 2004
From: joe@REDACTED (Joe Armstrong)
Date: Tue, 7 Dec 2004 19:26:13 +0100 (CET)
Subject: sending emails with erlang (for use with YXA/CPL)
In-Reply-To: <E66A629F-482F-11D9-BEC7-00039378D5B8@it.su.se>
References: <E66A629F-482F-11D9-BEC7-00039378D5B8@it.su.se>
Message-ID: <Pine.LNX.4.61.0412071922230.21683@enfield.sics.se>


So what's wrong with the stuff in email-1.0? it's pretty easy to 
understand and tweak if it doesn't do what you want. The guts of it do
this:

send(To, Subject, Data) ->
     {TmpFile, S} = open_tmp_file("/tmp", ".mail"),
     io:format(S, "To: <~s>~n", [To]),
     io:format(S, "Subject: ~s~n~n", [Subject]),
     io:format(S, "~s~n",[Data]),
     io:format(S, ".~nquit~n", []),
     close(S),
     io:format("sending ...~n", []),
     unix:cmd("/usr/lib/sendmail -t > /dev/null < " ++ TmpFile),
     delete(TmpFile).

      ... not exactly rocket sience but it does the job

<<I use this code in my wiki to email the user their password
if they forget it :-) >>

     Using a full sntp client or pop3 client for that matter is overkill

    Cheers

/Joe




On Tue, 7 Dec 2004, Administrat?r wrote:

> Is there some easy way to get erlang to send emails or should I simply use 
> os:cmd/1 and call some appropriate unix command ?
> The only thing I found that might do the job (I haven't looked at it in 
> depth) is email-1.0 by Joe Armstrong, in the erlang.org "User Contributions" 
> section.
>
> I would like to send emails based on mail urls e.g. :
>
>      mailto:jones@REDACTED?subject=Lookup%20failed
>
> i.e. urls that contain a mail address and possibly additional arguments, it 
> should also be possible to attach additional data to the mail in some easy 
> manner.
>
> I'm interested in this as I'm currently implementing CPL (Call Processing 
> Language - RFC 3880), for use with YXA (see 
> http://www.stacken.kth.se/projekt/yxa/) and it would be nice if I could 
> support the CPL mail tag, if it isn't to much work - it's currently a 
> non-essential feature.
>

From blaguehubert@REDACTED  Wed Dec  8 01:54:30 2004
From: blaguehubert@REDACTED (blaguehubert)
Date: Wed, 08 Dec 2004 00:54:30 +0000
Subject: [Fwd: http client hell]
Message-ID: <41B650C6.6010201@free.fr>

I'm trying to use the http:request function, but I randomly get this

horrible message before it crash my program:

*=ERROR REPORT==== 7-Dec-2004::23:39:22 ===
** Generic server <0.64.0> terminating
** Last message in was {tcp_closed,#Port<0.137>}
** When Server state == {state,{request,
                                  #Ref<0.0.0.822>,
                                  <0.114.0>,
                                  0,
                                  http,
                                  {"perso.efrei.fr",80},
                                  "/~colombi//",
                                  [],
                                  get,
                                  {http_request_h,
                                      undefined,
                                      "keep-alive",
                                      undefined,
                                      undefined,
                                      undefined,
                                      undefined,

etc...



Here is my code :
*
req(Url) ->
case (catch req(Url,10)) of
       {ok, Body} ->
               {ok, Body};
       _ ->
               error
       end.

req(Url, NbTry) ->
{ok, RequestId} = http:request(get, {Url, []}, [], [{sync, false}]),
receive
{http, {RequestId, Result}} ->
       {N, Header, Body} = Result,
       {ok, binary_to_list(Body)}
after 60000 ->
       if
               NbTry) > 0 ->
                       req(Url, NbEssais-1);
               true ->
                       error
       end
end.

Like you can see, I use catch statement and timeout because I thought it 
was only an exception, but it seems more serious...





From jay@REDACTED  Wed Dec  8 07:33:46 2004
From: jay@REDACTED (Jay Nelson)
Date: Tue, 07 Dec 2004 22:33:46 -0800
Subject: Embedded telecom databases get a benchmark of their own
Message-ID: <41B6A04A.9040908@duomark.com>

Article on Linux Devices website describing a new open source telecoms 
benchmark:

http://www.linuxdevices.com/news/NS8213292690.html

jay



From raimo@REDACTED  Wed Dec  8 08:47:01 2004
From: raimo@REDACTED (Raimo Niskanen)
Date: 08 Dec 2004 08:47:01 +0100
Subject: [Fwd: http client hell]
References: <41B650C6.6010201@free.fr>
Message-ID: <lq4qixdyfe.fsf@erix.ericsson.se>

You got the message {tcp_closed,Port}, so the http server closed the
connection for some reason, and it seems that the http client did
not expect this.

Don't ask me why, I hope someone with better http client knowledge 
will enlighten you.

blaguehubert@REDACTED (blaguehubert) writes:

> I'm trying to use the http:request function, but I randomly get this
> 
> horrible message before it crash my program:
> 
> *=ERROR REPORT==== 7-Dec-2004::23:39:22 ===
> ** Generic server <0.64.0> terminating
> ** Last message in was {tcp_closed,#Port<0.137>}
> ** When Server state == {state,{request,
>                                   #Ref<0.0.0.822>,
>                                   <0.114.0>,
>                                   0,
>                                   http,
>                                   {"perso.efrei.fr",80},
>                                   "/~colombi//",
>                                   [],
>                                   get,
>                                   {http_request_h,
>                                       undefined,
>                                       "keep-alive",
>                                       undefined,
>                                       undefined,
>                                       undefined,
>                                       undefined,
> 
> etc...
> 
> 
> 
> Here is my code :
> *
> req(Url) ->
> case (catch req(Url,10)) of
>        {ok, Body} ->
>                {ok, Body};
>        _ ->
>                error
>        end.
> 
> req(Url, NbTry) ->
> {ok, RequestId} = http:request(get, {Url, []}, [], [{sync, false}]),
> receive
> {http, {RequestId, Result}} ->
>        {N, Header, Body} = Result,
>        {ok, binary_to_list(Body)}
> after 60000 ->
>        if
>                NbTry) > 0 ->
>                        req(Url, NbEssais-1);
>                true ->
>                        error
>        end
> end.
> 
> Like you can see, I use catch statement and timeout because I thought
> it was only an exception, but it seems more serious...
> 
> 
> 

-- 

/ Raimo Niskanen, Erlang/OTP, Ericsson AB


From ingela@REDACTED  Wed Dec  8 09:58:30 2004
From: ingela@REDACTED (Ingela Anderton)
Date: Wed, 8 Dec 2004 09:58:30 +0100
Subject: [Fwd: http client hell]
References: <41B650C6.6010201@free.fr>
Message-ID: <16822.49718.29049.233958@gargle.gargle.HOWL>

blaguehubert wrote:
> I'm trying to use the http:request function, but I randomly get this
> 
> horrible message before it crash my program:
> 
> *=ERROR REPORT==== 7-Dec-2004::23:39:22 ===
> ** Generic server <0.64.0> terminating
> ** Last message in was {tcp_closed,#Port<0.137>}
> ** When Server state == {state,{request,
>                                   #Ref<0.0.0.822>,
>                                   <0.114.0>,
>                                   0,
>                                   http,
>                                   {"perso.efrei.fr",80},
>                                   "/~colombi//",
>                                   [],
>                                   get,
>                                   {http_request_h,
>                                       undefined,
>                                       "keep-alive",
>                                       undefined,
>                                       undefined,
>                                       undefined,
>                                       undefined,
> 
> etc...
You cut out the most interesting part e.i. the reason for the
termination. That the tcp port was closed might be perfectly normal
as many servers close the connection to indicate that the whole answer
has been sent instead of sending a length indicator.
Which version of intes are you using?

> Here is my code :
> *
> req(Url) ->
> case (catch req(Url,10)) of
>        {ok, Body} ->
>                {ok, Body};
>        _ ->
>                error
>        end.
> 
> req(Url, NbTry) ->
> {ok, RequestId} = http:request(get, {Url, []}, [], [{sync, false}]),
> receive
> {http, {RequestId, Result}} ->
>        {N, Header, Body} = Result,
>        {ok, binary_to_list(Body)}
> after 60000 ->
>        if
>                NbTry) > 0 ->
>                        req(Url, NbEssais-1);
>                true ->
>                        error
>        end
> end.
> 
> Like you can see, I use catch statement and timeout because I thought it 
> was only an exception, but it seems more serious... 

I fixed your code above as it did not compile. And this is the result I get
every time I try it with intes-4.0.1

3> test:req("http://perso.efrei.fr:80/~colombi/").
{ok,"<?xml version=\"1.0\" encoding=\"iso-8859-1\"?>\n<!DOCTYPE html PUBLIC \"-//W3C//DTD XHTML 1.0 Strict//EN\"\n      \"http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd\">\n<html xmlns=\"http://www.w3.org/1999/xhtml\">\n<head>\n  <meta http-equiv=\"Content-Type\" content=\"text/html; charset=iso-8859-1\" />\n  <title>No title</title>\n  <meta name=\"generator\" content=\"amaya 8.5, see http://www.w3.org/Amaya/\" />\n</head>\n\n<body>\n<p>un lien de test: <a href=\"./ici.html\">ici</a></p>\n</body>\n</html>\n"}

Your code fixed:

-module(test).

-export([req/1]).

req(Url) ->
    case (catch req(Url,10)) of
	{ok, Body} ->
	    {ok, Body};
	_ ->
	    error
    end.

req(Url, NbTry) ->
    {ok, RequestId} = http:request(get, {Url, []}, [], [{sync, false}]),
    receive
	{http, {RequestId, Result}} ->
	    {_N, _Header, Body} = Result,
	    {ok, binary_to_list(Body)}
    after 60000 ->
	    if
		NbTry > 0 ->
                       req(Url, NbTry-1);
               true ->
                       error
       end
end.

-- 
/Ingela

Ericsson AB - OTP team







From hakan.stenholm@REDACTED  Wed Dec  8 14:35:13 2004
From: hakan.stenholm@REDACTED (=?ISO-8859-1?Q?Administrat=F6r?=)
Date: Wed, 8 Dec 2004 14:35:13 +0100
Subject: sending emails with erlang (for use with YXA/CPL)
In-Reply-To: <Pine.LNX.4.61.0412071922230.21683@enfield.sics.se>
References: <E66A629F-482F-11D9-BEC7-00039378D5B8@it.su.se> <Pine.LNX.4.61.0412071922230.21683@enfield.sics.se>
Message-ID: <FD19C8F8-491D-11D9-BEC7-00039378D5B8@it.su.se>


On 2004-12-07, at 19.26, Joe Armstrong wrote:

>
> So what's wrong with the stuff in email-1.0?

Probably nothing, I'm simply looking for the solution that requires the 
least amount of work, for me to use.





From blaguehubert@REDACTED  Wed Dec  8 15:23:48 2004
From: blaguehubert@REDACTED (blaguehubert)
Date: Wed, 08 Dec 2004 14:23:48 +0000
Subject: [Fwd: http client hell]
In-Reply-To: <16822.49718.29049.233958@gargle.gargle.HOWL>
References: <41B650C6.6010201@free.fr> <16822.49718.29049.233958@gargle.gargle.HOWL>
Message-ID: <41B70E74.6030104@free.fr>

I'm using inets shipped with Erlang/OTP10.

I tried your fixed version, but I have the same error with some websites.

Here is the full error message:

5> test:req("http://www.flipcode.com").

=ERROR REPORT==== 8-Dec-2004::14:15:33 ===
** Generic server <0.55.0> terminating
** Last message in was {tcp,#Port<0.134>,
                            
<<72,84,84,80,47,49,46,49,32,50,48,48,32,79,75,13,10,68,97,116,101,58,32,87,101,100,44,32,48,56,32,68,101,99,32,50,48,48,52,32,49,53,58,48,51,58,51,57,32,71,77,84,13,10,83,101,114,118,101,114,58,32,65,112,97,99,104,101,47,49,46,51,46,50,55,32,40,85,110,105,120,41,32,32,40,82,101,100,45,72,97,116,47,76,105,110,117,120,41,32,109,111,100,95,115,115,108,47,50,46,56,46,49,50,32,79,112,101,110,83,83,76,47,48,46,57,46,54,98,32,68,65,86,47,49,46,48,46,51,32,80,72,80,47,52,46,49,46,50,32,109,111,100,95,112,101,114,108,47,49,46,50,54,13,10,67,111,110,110,101,99,116,105,111,110,58,32,99,108,111,115,101,13,10,84,114,97,110,115,102,101,114,45,69,110,99,111,100,105,110,103,58,32,99,104,117,110,107,101,100,13,10,67,111,110,116,101,110,116,45,84,121,112,101,58,32,116,101,120,116,47,104,116,109,108,13,10,13,10,102,101,55,13,10,60,104,116,109,108,62,10,60,104,101,97,100,62,60,116,105,116,108,101,62,102,108,105,112,99,111,100,101,32,45,32,68,101,118,101,108,111,112,109,101,110,116,32,78,101,119,115,60,47,116,105,116,108,101,62,10,60,115,116,121,108,101,32,116,121,112,101,61,34,116,101,120,116,47,99,115,115,34,62,10,10,10,97,46,109,101,110,117,108,105,110,107,58,108,105,110,107,32,32,32,32,123,99,111,108,111,114,58,32,35,98,57,102,102,100,48,59,32,125,10,97,46,109,101,110,117,108,105,110,107,58,118,105,115,105,116,101,100,32,123,99,111,108,111,114,58,32,35,98,57,102,102,100,48,59,32,125,10,97,46,109,101,110,117,108,105,110,107,58,97,99,116,105,118,101,32,32,123,99,111,108,111,114,58,32,35,98,57,102,102,100,48,59,32,125,10,10,97,46,109,101,110,117,108,105,110,107,101,109,112,116,121,58,108,105,110,107,32,32,32,32,123,99,111,108,111,114,58,32,35,98,57,102,102,100,48,59,32,125,10,97,46,109,101,110,117,108,105,110,107,101,109,112,116,121,58,118,105,115,105,116,101,100,32,123,99,111,108,111,114,58,32,35,98,57,102,102,100,48,59,32,125,10,97,46,109,101,110,117,108,105,110,107,101,109,112,116,121,58,97,99,116,105,118,101,32,32,123,99,111,108,111,114,58,32,35,98,57,102,102,100,48,59,32,125,10,97,46,109,101,110,117,108,105,110,107,101,109,112
07,101,109,112,116,121,58,118,105,115,105,116,101,100,44,32,97,46,109,101,110,117,108,105,110,107,101,109,112,116,121,58,97,99,116,105,118,101,32,123,116,101,120,116,45,100,101,99,111,114,97,116,105,111,110,58,32,110,111,110,101,125,10,10,97,46,111,114,97,110,103,101,108,105,110,107,58,108,105,110,107,32,32,32,32,123,32,99,111,108,111,114,58,35,70,70,65,66,48,52,59,32,125,10,97,46,111,114,97,110,103,101,108,105,110,107,58,118,105,115,105,116,101,100,32,123,32,99,111,108,111,114,58,35,70,70,65,66,48,52,59,32,125,10,97,46,111,114,97,110,103,101,108,105,110,107,58,97,99,116,105,118,101,32,32,123,32,99,111,108,111,114,58,35,70,70,65,66,48,52,59,32,125,10,10,97,46,112,97,108,101,103,114,101,101,110,58,108,105,110,107,32,32,32,32,123,99,111,108,111,114,58,32,35,98,57,102,102,100,48,59,32,125,10,97,46,112,97,108,101,103,114,101,101,110,58,118,105,115,105,116,101,100,32,123,99,111,108,111,114,58,32,35,98,57,102,102,100,48,59,32,125,10,97,46,112,97,108,101,103,114,101,101,110,58,97,99,116,105,118,101,32,32,123,99,111,108,111,114,58,32,35,98,57,102,102,100,48,59,32,125,10,10,97,46,98,108,117,101,108,105,110,107,58,108,105,110,107,32,32,32,32,123,32,99,111,108,111,114,58,35,48,51,70,48,70,70,59,32,125,10,97,46,98,108,117,101,108,105,110,107,58,118,105,115,105,116,101,100,32,123,32,99,111,108,111,114,58,35,48,51,70,48,70,70,59,32,125,10,97,46,98,108,117,101,108,105,110,107,58,97,99,116,105,118,101,32,32,123,32,99,111,108,111,114,58,35,48,51,70,48,70,70,59,32,125,10,10,97,46,115,111,102>>}
** When Server state == {state,{request,
                                   #Ref<0.0.0.261>,
                                   <0.36.0>,
                                   0,
                                   http,
                                   {"www.flipcode.com",80},
                                   "/",
                                   [],
                                   get,
                                   {http_request_h,
                                       undefined,
                                       "keep-alive",
                                       undefined,
                                       undefined,
                                       undefined,
                                       undefined,
                                       undefined,
                                       undefined,
                                       undefined,
                                       undefined,
                                       undefined,
                                       undefined,
                                       undefined,
                                       undefined,
                                       undefined,
                                       undefined,
                                       "www.flipcode.com",
                                       undefined,
                                       undefined,
                                       undefined,
                                       undefined,
                                       undefined,
                                       undefined,
                                       undefined,
                                       undefined,
                                       undefined,
                                       [],
                                       undefined,
                                       undefined,
                                       undefined,
                                       undefined,
                                       "0",
                                       undefined,
                                       undefined,
                                       undefined,
                                       undefined,
                                       undefined,
                                       undefined,
                                       []},
                                   {[],[]},
                                   
{http_options,infinity,true,[],false,false}},

                               {tcp_session,
                                   {{"www.flipcode.com",80},<0.55.0>},
                                   false,
                                   http,
                                   #Port<0.134>,
                                   1},
                               undefined,
                               undefined,
                               undefined,
                               {http_response,parse,[nolimit]},
                               {[],[]},
                               true,
                               [],
                               nolimit,
                               nolimit}
** Reason for termination ==
** {{badmatch,{noreply,{state,{request,
                                  #Ref<0.0.0.261>,
                                  <0.36.0>,
                                  0,
                                  http,
                                  {"www.flipcode.com",80},
                                  "/",
                                  [],
                                  get,
                                  {http_request_h,
                                      undefined,
                                      "keep-alive",
                                      undefined,
                                      undefined,
                                      undefined,
                                      undefined,
                                      undefined,
                                      undefined,
                                      undefined,
                                      undefined,
                                      undefined,
                                      undefined,
                                      undefined,
                                      undefined,
                                      undefined,
                                      undefined,
                                      "www.flipcode.com",
                                      undefined,
                                      undefined,
                                      undefined,
                                      undefined,
                                      undefined,
                                      undefined,
                                      undefined,
                                      undefined,
                                      undefined,
                                      [],
                                      undefined,
                                      undefined,
                                      undefined,
                                      undefined,
                                      "0",
                                      undefined,
                                      undefined,
                                      undefined,
                                      undefined,
                                      undefined,
                                      undefined,
                                      []},
                                  {[],[]},
                                  
{http_options,infinity,true,[],false,false}},
                              {tcp_session,
                                  {{"www.flipcode.com",80},<0.55.0>},
                                  false,
                                  http,
                                  #Port<0.134>,
                                  1},
                              {"HTTP/1.1",200,"OK"},
                              {http_response_h,
                                  undefined,
                                  "close",
                                  undefined,
                                  undefined,
                                  undefined,
                                  "chunked",
                                  undefined,
                                  undefined,
                                  undefined,
                                  undefined,
                                  undefined,
                                  undefined,
                                  undefined,
                                  undefined,
                                  undefined,
                                  undefined,
                                  undefined,
                                  undefined,
                                  undefined,
                                  undefined,
                                  undefined,
                                  "0",
                                  undefined,
                                  undefined,
                                  undefined,
                                  "text/html",
                                  undefined,
                                  undefined,
                                  [{"date","Wed, 08 Dec 2004 15:03:39 GMT"},
                                   {"server",
                                    "Apache/1.3.27 (Unix)  
(Red-Hat/Linux) mod_ssl/2.8.12 OpenSSL/0.9.6b DAV/1.0.3 PHP/4.1.2 
mod_perl/1.26"}]},
                              undefined,
                              {http_chunk,
                                  decode_data,
                                  [4071,
                                   
<<60,104,116,109,108,62,10,60,104,101,97,100,62,60,116,105,116,108,101,62,102,108,105,112,99,111,100,101,32,45,32,68,101,118,101,108,111,112,109,101,110,116,32,78,101,119,115,60,47,116,105,116,108,101,62,10,60,115,116,121,108,101,32,116,121,112,101,61,34,116,101,120,116,47,99,115,115,34,62,10,10,10,97,46,109,101,110,117,108,105,110,107,58,108,105,110,107,32,32,32,32,123,99,111,108,111,114,58,32,35,98,57,102,102,100,48,59,32,125,10,97,46,109,101,110,117,108,105,110,107,58,118,105,115,105,116,101,100,32,123,99,111,108,111,114,58,32,35,98,57,102,102,100,48,59,32,125,10,97,46,109,101,110,117,108,105,110,107,58,97,99,116,105,118,101,32,32,123,99,111,108,111,114,58,32,35,98,57,102,102,100,48,59,32,125,10,10,97,46,109,101,110,117,108,105,110,107,101,109,112,116,121,58,108,105,110,107,32,32,32,32,123,99,111,108,111,114,58,32,35,98,57,102,102,100,48,59,32,125,10,97,46,109,101,110,117,108,105,110,107,101,109,112,116,121,58,118,105,115,105,116,101,100,32,123,99,111,108,111,114,58,32,35,98,57,102,102,100,48,59,32,125,10,97,46,109,101,110,117,108,105,110,107,101,109,112,116,121,58,97,99,116,105,118,101,32,32,123,99,111,108,111,114,58,32,35,98,57,102,102,100,48,59,32,125,10,97,46,109,101,110,117,108,105,110,107,101,109,112,116,121,58,108,105,110,107,44,32,97,46,109,101,110,117,108,105,110,107,101,109,112,116,121,58,118,105,115,105,116,101,100,44,32,97,46,109,101,110,117,108,105,110,107,101,109,112,116,121,58,97,99,116,105,118,101,32,123,116,101,120,116,45,100,101,99,111,114,97,116,105,111,110,58,32,110,111,110,101,125,10,10,97,46,111,114,97,110,103,101,108,105,110,107,58,108,105,110,107,32,32,32,32,123,32,99,111,108,111,114,58,35,70,70,65,66,48,52,59,32,125,10,97,46,111,114,97,110,103,101,108,105,110,107,58,118,105,115,105,116,101,100,32,123,32,99,111,108,111,114,58,35,70,70,65,66,48,52,59,32,125,10,97,46,111,114,97,110,103,101,108,105,110,107,58,97,99,116,105,118,101,32,32,123,32,99,111,108,111,114,58,35,70,70,65,66,48,52,59,32,125,10,10,97,46,112,97,108,101,103,114,101,101,110,58,108,105,110,107,32,32,32,32,123,99,111,108,11
,103,114,101,101,110,58,118,105,115,105,116,101,100,32,123,99,111,108,111,114,58,32,35,98,57,102,102,100,48,59,32,125,10,97,46,112,97,108,101,103,114,101,101,110,58,97,99,116,105,118,101,32,32,123,99,111,108,111,114,58,32,35,98,57,102,102,100,48,59,32,125,10,10,97,46,98,108,117,101,108,105,110,107,58,108,105,110,107,32,32,32,32,123,32,99,111,108,111,114,58,35,48,51,70,48,70,70,59,32,125,10,97,46,98,108,117,101,108,105,110,107,58,118,105,115,105,116,101,100,32,123,32,99,111,108,111,114,58,35,48,51,70,48,70,70,59,32,125,10,97,46,98,108,117,101,108,105,110,107,58,97,99,116,105,118,101,32,32,123,32,99,111,108,111,114,58,35,48,51,70,48,70,70,59,32,125,10,10,97,46,115,111,102>>,
                                   {nolimit,<<>>,4071,nolimit}]},
                              {[],[]},
                              true,
                              [],
                              nolimit,
                              nolimit}}},
    [{httpc_handler,handle_http_body,2},
     {gen_server,handle_msg,6},
     {proc_lib,init_p,5}]}

Ingela Anderton wrote:

>blaguehubert wrote:
>  
>
>>I'm trying to use the http:request function, but I randomly get this
>>
>>horrible message before it crash my program:
>>
>>*=ERROR REPORT==== 7-Dec-2004::23:39:22 ===
>>** Generic server <0.64.0> terminating
>>** Last message in was {tcp_closed,#Port<0.137>}
>>** When Server state == {state,{request,
>>                                  #Ref<0.0.0.822>,
>>                                  <0.114.0>,
>>                                  0,
>>                                  http,
>>                                  {"perso.efrei.fr",80},
>>                                  "/~colombi//",
>>                                  [],
>>                                  get,
>>                                  {http_request_h,
>>                                      undefined,
>>                                      "keep-alive",
>>                                      undefined,
>>                                      undefined,
>>                                      undefined,
>>                                      undefined,
>>
>>etc...
>>    
>>
>You cut out the most interesting part e.i. the reason for the
>termination. That the tcp port was closed might be perfectly normal
>as many servers close the connection to indicate that the whole answer
>has been sent instead of sending a length indicator.
>Which version of intes are you using?
>
>  
>
>>Here is my code :
>>*
>>req(Url) ->
>>case (catch req(Url,10)) of
>>       {ok, Body} ->
>>               {ok, Body};
>>       _ ->
>>               error
>>       end.
>>
>>req(Url, NbTry) ->
>>{ok, RequestId} = http:request(get, {Url, []}, [], [{sync, false}]),
>>receive
>>{http, {RequestId, Result}} ->
>>       {N, Header, Body} = Result,
>>       {ok, binary_to_list(Body)}
>>after 60000 ->
>>       if
>>               NbTry) > 0 ->
>>                       req(Url, NbEssais-1);
>>               true ->
>>                       error
>>       end
>>end.
>>
>>Like you can see, I use catch statement and timeout because I thought it 
>>was only an exception, but it seems more serious... 
>>    
>>
>
>I fixed your code above as it did not compile. And this is the result I get
>every time I try it with intes-4.0.1
>
>3> test:req("http://perso.efrei.fr:80/~colombi/").
>{ok,"<?xml version=\"1.0\" encoding=\"iso-8859-1\"?>\n<!DOCTYPE html PUBLIC \"-//W3C//DTD XHTML 1.0 Strict//EN\"\n      \"http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd\">\n<html xmlns=\"http://www.w3.org/1999/xhtml\">\n<head>\n  <meta http-equiv=\"Content-Type\" content=\"text/html; charset=iso-8859-1\" />\n  <title>No title</title>\n  <meta name=\"generator\" content=\"amaya 8.5, see http://www.w3.org/Amaya/\" />\n</head>\n\n<body>\n<p>un lien de test: <a href=\"./ici.html\">ici</a></p>\n</body>\n</html>\n"}
>
>Your code fixed:
>
>-module(test).
>
>-export([req/1]).
>
>req(Url) ->
>    case (catch req(Url,10)) of
>	{ok, Body} ->
>	    {ok, Body};
>	_ ->
>	    error
>    end.
>
>req(Url, NbTry) ->
>    {ok, RequestId} = http:request(get, {Url, []}, [], [{sync, false}]),
>    receive
>	{http, {RequestId, Result}} ->
>	    {_N, _Header, Body} = Result,
>	    {ok, binary_to_list(Body)}
>    after 60000 ->
>	    if
>		NbTry > 0 ->
>                       req(Url, NbTry-1);
>               true ->
>                       error
>       end
>end.
>
>  
>




From ingela@REDACTED  Wed Dec  8 16:30:13 2004
From: ingela@REDACTED (Ingela Anderton)
Date: Wed, 8 Dec 2004 16:30:13 +0100
Subject: [Fwd: http client hell]
References: <41B650C6.6010201@free.fr>
	<16822.49718.29049.233958@gargle.gargle.HOWL>
	<41B70E74.6030104@free.fr>
Message-ID: <16823.7685.261964.864904@gargle.gargle.HOWL>


blaguehubert wrote:
> I'm using inets shipped with Erlang/OTP10.
This is a known bug in inets-4.0 that appears when
the http response is chunk decoded. It has already been fixed and
patched for commercial customers and will be part of the next open
source release.

> I tried your fixed version, but I have the same error with some websites.
I only fixed the test program!  

> Here is the full error message:
> 5> test:req("http://www.flipcode.com").
> 
> =ERROR REPORT==== 8-Dec-2004::14:15:33 ===
> ** Generic server <0.55.0> terminating
> ** Last message in was {tcp,#Port<0.134>,
>                             
> <<72,84,84,80,47,49,46,49,32,50,48,48,32,79,75,13,10,68,97,116,101,58,32,87,101,100,44,32,48,56,32,68,101,99,32,50,48,48,52,32,49,53,58,48,51,58,51,57,32,71,77,84,13,10,83,101,114,118,101,114,58,32,65,112,97,99,104,101,47,49,46,51,46,50,55,32,40,85,110,105,120,41,32,32,40,82,101,100,45,72,97,116,47,76,105,110,117,120,41,32,109,111,100,95,115,115,108,47,50,46,56,46,49,50,32,79,112,101,110,83,83,76,47,48,46,57,46,54,98,32,68,65,86,47,49,46,48,46,51,32,80,72,80,47,52,46,49,46,50,32,109,111,100,95,112,101,114,108,47,49,46,50,54,13,10,67,111,110,110,101,99,116,105,111,110,58,32,99,108,111,115,101,13,10,84,114,97,110,115,102,101,114,45,69,110,99,111,100,105,110,103,58,32,99,104,117,110,107,101,100,13,10,67,111,110,116,101,110,116,45,84,121,112,101,58,32,116,101,120,116,47,104,116,109,108,13,10,13,10,102,101,55,13,10,60,104,116,109,108,62,10,60,104,101,97,100,62,60,116,105,116,108,101,62,102,108,105,112,99,111,100,101,32,45,32,68,101,118,101,108,111,112,109,101,110,116,32,78,101,!
11!
> 9,115,60,47,116,105,116,108,101,62,10,60,115,116,121,108,101,32,116,121,112,101,61,34,116,101,120,116,47,99,115,115,34,62,10,10,10,97,46,109,101,110,117,108,105,110,107,58,108,105,110,107,32,32,32,32,123,99,111,108,111,114,58,32,35,98,57,102,102,100,48,59,32,125,10,97,46,109,101,110,117,108,105,110,107,58,118,105,115,105,116,101,100,32,123,99,111,108,111,114,58,32,35,98,57,102,102,100,48,59,32,125,10,97,46,109,101,110,117,108,105,110,107,58,97,99,116,105,118,101,32,32,123,99,111,108,111,114,58,32,35,98,57,102,102,100,48,59,32,125,10,10,97,46,109,101,110,117,108,105,110,107,101,109,112,116,121,58,108,105,110,107,32,32,32,32,123,99,111,108,111,114,58,32,35,98,57,102,102,100,48,59,32,125,10,97,46,109,101,110,117,108,105,110,107,101,109,112,116,121,58,118,105,115,105,116,101,100,32,123,99,111,108,111,114,58,32,35,98,57,102,102,100,48,59,32,125,10,97,46,109,101,110,117,108,105,110,107,101,109,112,116,121,58,97,99,116,105,118,101,32,32,123,99,111,108,111,114,58,32,35,98,57,102,1!
02!
> ,100,48,59,32,125,10,97,46,109,101,110,117,108,105,110,107,101,109,112
> 07,101,109,112,116,121,58,118,105,115,105,116,101,100,44,32,97,46,109,101,110,117,108,105,110,107,101,109,112,116,121,58,97,99,116,105,118,101,32,123,116,101,120,116,45,100,101,99,111,114,97,116,105,111,110,58,32,110,111,110,101,125,10,10,97,46,111,114,97,110,103,101,108,105,110,107,58,108,105,110,107,32,32,32,32,123,32,99,111,108,111,114,58,35,70,70,65,66,48,52,59,32,125,10,97,46,111,114,97,110,103,101,108,105,110,107,58,118,105,115,105,116,101,100,32,123,32,99,111,108,111,114,58,35,70,70,65,66,48,52,59,32,125,10,97,46,111,114,97,110,103,101,108,105,110,107,58,97,99,116,105,118,101,32,32,123,32,99,111,108,111,114,58,35,70,70,65,66,48,52,59,32,125,10,10,97,46,112,97,108,101,103,114,101,101,110,58,108,105,110,107,32,32,32,32,123,99,111,108,111,114,58,32,35,98,57,102,102,100,48,59,32,125,10,97,46,112,97,108,101,103,114,101,101,110,58,118,105,115,105,116,101,100,32,123,99,111,108,111,114,58,32,35,98,57,102,102,100,48,59,32,125,10,97,46,112,97,108,101,103,114,101,101,110,58,97!
,9!
> 9,116,105,118,101,32,32,123,99,111,108,111,114,58,32,35,98,57,102,102,100,48,59,32,125,10,10,97,46,98,108,117,101,108,105,110,107,58,108,105,110,107,32,32,32,32,123,32,99,111,108,111,114,58,35,48,51,70,48,70,70,59,32,125,10,97,46,98,108,117,101,108,105,110,107,58,118,105,115,105,116,101,100,32,123,32,99,111,108,111,114,58,35,48,51,70,48,70,70,59,32,125,10,97,46,98,108,117,101,108,105,110,107,58,97,99,116,105,118,101,32,32,123,32,99,111,108,111,114,58,35,48,51,70,48,70,70,59,32,125,10,10,97,46,115,111,102>>}
> ** When Server state == {state,{request,
>                                    #Ref<0.0.0.261>,
>                                    <0.36.0>,
>                                    0,
>                                    http,
>                                    {"www.flipcode.com",80},
>                                    "/",
>                                    [],
>                                    get,
>                                    {http_request_h,
>                                        undefined,
>                                        "keep-alive",
>                                        undefined,
>                                        undefined,
>                                        undefined,
>                                        undefined,
>                                        undefined,
>                                        undefined,
>                                        undefined,
>                                        undefined,
>                                        undefined,
>                                        undefined,
>                                        undefined,
>                                        undefined,
>                                        undefined,
>                                        undefined,
>                                        "www.flipcode.com",
>                                        undefined,
>                                        undefined,
>                                        undefined,
>                                        undefined,
>                                        undefined,
>                                        undefined,
>                                        undefined,
>                                        undefined,
>                                        undefined,
>                                        [],
>                                        undefined,
>                                        undefined,
>                                        undefined,
>                                        undefined,
>                                        "0",
>                                        undefined,
>                                        undefined,
>                                        undefined,
>                                        undefined,
>                                        undefined,
>                                        undefined,
>                                        []},
>                                    {[],[]},
>                                    
> {http_options,infinity,true,[],false,false}},
> 
>                                {tcp_session,
>                                    {{"www.flipcode.com",80},<0.55.0>},
>                                    false,
>                                    http,
>                                    #Port<0.134>,
>                                    1},
>                                undefined,
>                                undefined,
>                                undefined,
>                                {http_response,parse,[nolimit]},
>                                {[],[]},
>                                true,
>                                [],
>                                nolimit,
>                                nolimit}
> ** Reason for termination ==
> ** {{badmatch,{noreply,{state,{request,
>                                   #Ref<0.0.0.261>,
>                                   <0.36.0>,
>                                   0,
>                                   http,
>                                   {"www.flipcode.com",80},
>                                   "/",
>                                   [],
>                                   get,
>                                   {http_request_h,
>                                       undefined,
>                                       "keep-alive",
>                                       undefined,
>                                       undefined,
>                                       undefined,
>                                       undefined,
>                                       undefined,
>                                       undefined,
>                                       undefined,
>                                       undefined,
>                                       undefined,
>                                       undefined,
>                                       undefined,
>                                       undefined,
>                                       undefined,
>                                       undefined,
>                                       "www.flipcode.com",
>                                       undefined,
>                                       undefined,
>                                       undefined,
>                                       undefined,
>                                       undefined,
>                                       undefined,
>                                       undefined,
>                                       undefined,
>                                       undefined,
>                                       [],
>                                       undefined,
>                                       undefined,
>                                       undefined,
>                                       undefined,
>                                       "0",
>                                       undefined,
>                                       undefined,
>                                       undefined,
>                                       undefined,
>                                       undefined,
>                                       undefined,
>                                       []},
>                                   {[],[]},
>                                   
> {http_options,infinity,true,[],false,false}},
>                               {tcp_session,
>                                   {{"www.flipcode.com",80},<0.55.0>},
>                                   false,
>                                   http,
>                                   #Port<0.134>,
>                                   1},
>                               {"HTTP/1.1",200,"OK"},
>                               {http_response_h,
>                                   undefined,
>                                   "close",
>                                   undefined,
>                                   undefined,
>                                   undefined,
>                                   "chunked",
>                                   undefined,
>                                   undefined,
>                                   undefined,
>                                   undefined,
>                                   undefined,
>                                   undefined,
>                                   undefined,
>                                   undefined,
>                                   undefined,
>                                   undefined,
>                                   undefined,
>                                   undefined,
>                                   undefined,
>                                   undefined,
>                                   undefined,
>                                   "0",
>                                   undefined,
>                                   undefined,
>                                   undefined,
>                                   "text/html",
>                                   undefined,
>                                   undefined,
>                                   [{"date","Wed, 08 Dec 2004 15:03:39 GMT"},
>                                    {"server",
>                                     "Apache/1.3.27 (Unix)  
> (Red-Hat/Linux) mod_ssl/2.8.12 OpenSSL/0.9.6b DAV/1.0.3 PHP/4.1.2 
> mod_perl/1.26"}]},
>                               undefined,
>                               {http_chunk,
>                                   decode_data,
>                                   [4071,
>                                    
> <<60,104,116,109,108,62,10,60,104,101,97,100,62,60,116,105,116,108,101,62,102,108,105,112,99,111,100,101,32,45,32,68,101,118,101,108,111,112,109,101,110,116,32,78,101,119,115,60,47,116,105,116,108,101,62,10,60,115,116,121,108,101,32,116,121,112,101,61,34,116,101,120,116,47,99,115,115,34,62,10,10,10,97,46,109,101,110,117,108,105,110,107,58,108,105,110,107,32,32,32,32,123,99,111,108,111,114,58,32,35,98,57,102,102,100,48,59,32,125,10,97,46,109,101,110,117,108,105,110,107,58,118,105,115,105,116,101,100,32,123,99,111,108,111,114,58,32,35,98,57,102,102,100,48,59,32,125,10,97,46,109,101,110,117,108,105,110,107,58,97,99,116,105,118,101,32,32,123,99,111,108,111,114,58,32,35,98,57,102,102,100,48,59,32,125,10,10,97,46,109,101,110,117,108,105,110,107,101,109,112,116,121,58,108,105,110,107,32,32,32,32,123,99,111,108,111,114,58,32,35,98,57,102,102,100,48,59,32,125,10,97,46,109,101,110,117,108,105,110,107,101,109,112,116,121,58,118,105,115,105,116,101,100,32,123,99,111,108,111,114,58,32,!
35!
> ,98,57,102,102,100,48,59,32,125,10,97,46,109,101,110,117,108,105,110,107,101,109,112,116,121,58,97,99,116,105,118,101,32,32,123,99,111,108,111,114,58,32,35,98,57,102,102,100,48,59,32,125,10,97,46,109,101,110,117,108,105,110,107,101,109,112,116,121,58,108,105,110,107,44,32,97,46,109,101,110,117,108,105,110,107,101,109,112,116,121,58,118,105,115,105,116,101,100,44,32,97,46,109,101,110,117,108,105,110,107,101,109,112,116,121,58,97,99,116,105,118,101,32,123,116,101,120,116,45,100,101,99,111,114,97,116,105,111,110,58,32,110,111,110,101,125,10,10,97,46,111,114,97,110,103,101,108,105,110,107,58,108,105,110,107,32,32,32,32,123,32,99,111,108,111,114,58,35,70,70,65,66,48,52,59,32,125,10,97,46,111,114,97,110,103,101,108,105,110,107,58,118,105,115,105,116,101,100,32,123,32,99,111,108,111,114,58,35,70,70,65,66,48,52,59,32,125,10,97,46,111,114,97,110,103,101,108,105,110,107,58,97,99,116,105,118,101,32,32,123,32,99,111,108,111,114,58,35,70,70,65,66,48,52,59,32,125,10,10,97,46,112,97,108,!
10!
> 1,103,114,101,101,110,58,108,105,110,107,32,32,32,32,123,99,111,108,11
> ,103,114,101,101,110,58,118,105,115,105,116,101,100,32,123,99,111,108,111,114,58,32,35,98,57,102,102,100,48,59,32,125,10,97,46,112,97,108,101,103,114,101,101,110,58,97,99,116,105,118,101,32,32,123,99,111,108,111,114,58,32,35,98,57,102,102,100,48,59,32,125,10,10,97,46,98,108,117,101,108,105,110,107,58,108,105,110,107,32,32,32,32,123,32,99,111,108,111,114,58,35,48,51,70,48,70,70,59,32,125,10,97,46,98,108,117,101,108,105,110,107,58,118,105,115,105,116,101,100,32,123,32,99,111,108,111,114,58,35,48,51,70,48,70,70,59,32,125,10,97,46,98,108,117,101,108,105,110,107,58,97,99,116,105,118,101,32,32,123,32,99,111,108,111,114,58,35,48,51,70,48,70,70,59,32,125,10,10,97,46,115,111,102>>,
>                                    {nolimit,<<>>,4071,nolimit}]},
>                               {[],[]},
>                               true,
>                               [],
>                               nolimit,
>                               nolimit}}},
>     [{httpc_handler,handle_http_body,2},
>      {gen_server,handle_msg,6},
>      {proc_lib,init_p,5}]}
> 
> Ingela Anderton wrote:
> 
> >blaguehubert wrote:
> >  
> >
> >>I'm trying to use the http:request function, but I randomly get this
> >>
> >>horrible message before it crash my program:
> >>
> >>*=ERROR REPORT==== 7-Dec-2004::23:39:22 ===
> >>** Generic server <0.64.0> terminating
> >>** Last message in was {tcp_closed,#Port<0.137>}
> >>** When Server state == {state,{request,
> >>                                  #Ref<0.0.0.822>,
> >>                                  <0.114.0>,
> >>                                  0,
> >>                                  http,
> >>                                  {"perso.efrei.fr",80},
> >>                                  "/~colombi//",
> >>                                  [],
> >>                                  get,
> >>                                  {http_request_h,
> >>                                      undefined,
> >>                                      "keep-alive",
> >>                                      undefined,
> >>                                      undefined,
> >>                                      undefined,
> >>                                      undefined,
> >>
> >>etc...
> >>    
> >>
> >You cut out the most interesting part e.i. the reason for the
> >termination. That the tcp port was closed might be perfectly normal
> >as many servers close the connection to indicate that the whole answer
> >has been sent instead of sending a length indicator.
> >Which version of intes are you using?
> >
> >  
> >
> >>Here is my code :
> >>*
> >>req(Url) ->
> >>case (catch req(Url,10)) of
> >>       {ok, Body} ->
> >>               {ok, Body};
> >>       _ ->
> >>               error
> >>       end.
> >>
> >>req(Url, NbTry) ->
> >>{ok, RequestId} = http:request(get, {Url, []}, [], [{sync, false}]),
> >>receive
> >>{http, {RequestId, Result}} ->
> >>       {N, Header, Body} = Result,
> >>       {ok, binary_to_list(Body)}
> >>after 60000 ->
> >>       if
> >>               NbTry) > 0 ->
> >>                       req(Url, NbEssais-1);
> >>               true ->
> >>                       error
> >>       end
> >>end.
> >>
> >>Like you can see, I use catch statement and timeout because I thought it 
> >>was only an exception, but it seems more serious... 
> >>    
> >>
> >
> >I fixed your code above as it did not compile. And this is the result I get
> >every time I try it with intes-4.0.1
> >
> >3> test:req("http://perso.efrei.fr:80/~colombi/").
> >{ok,"<?xml version=\"1.0\" encoding=\"iso-8859-1\"?>\n<!DOCTYPE html PUBLIC \"-//W3C//DTD XHTML 1.0 Strict//EN\"\n      \"http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd\">\n<html xmlns=\"http://www.w3.org/1999/xhtml\">\n<head>\n  <meta http-equiv=\"Content-Type\" content=\"text/html; charset=iso-8859-1\" />\n  <title>No title</title>\n  <meta name=\"generator\" content=\"amaya 8.5, see http://www.w3.org/Amaya/\" />\n</head>\n\n<body>\n<p>un lien de test: <a href=\"./ici.html\">ici</a></p>\n</body>\n</html>\n"}
> >
> >Your code fixed:
> >
> >-module(test).
> >
> >-export([req/1]).
> >
> >req(Url) ->
> >    case (catch req(Url,10)) of
> >	{ok, Body} ->
> >	    {ok, Body};
> >	_ ->
> >	    error
> >    end.
> >
> >req(Url, NbTry) ->
> >    {ok, RequestId} = http:request(get, {Url, []}, [], [{sync, false}]),
> >    receive
> >	{http, {RequestId, Result}} ->
> >	    {_N, _Header, Body} = Result,
> >	    {ok, binary_to_list(Body)}
> >    after 60000 ->
> >	    if
> >		NbTry > 0 ->
> >                       req(Url, NbTry-1);
> >               true ->
> >                       error
> >       end
> >end.
> >
> >  
> >
> 

-- 
/Ingela

Ericsson AB - OTP team








From klacke@REDACTED  Wed Dec  8 21:23:05 2004
From: klacke@REDACTED (klacke@REDACTED)
Date: Wed, 8 Dec 2004 21:23:05 +0100
Subject: [Fwd: http client hell]
In-Reply-To: <16823.7685.261964.864904@gargle.gargle.HOWL>
References: <41B650C6.6010201@free.fr> <16822.49718.29049.233958@gargle.gargle.HOWL> <41B70E74.6030104@free.fr> <16823.7685.261964.864904@gargle.gargle.HOWL>
Message-ID: <20041208202305.GA11881@hyber.org>


Wow, is this a shitty attitude or what.  Here is this guy who finds 
bugs in the http client and you first write:


> I fixed your code above as it did not compile. And this is the result I get
> every time I try it with intes-4.0.1                    
> ...

indicating that there is indeed no problem found, although it was
obvious from the error_report that the client was either
wrong or overly mis-forgiving to some http server behaviors.

Then blaguehubert@REDACTED sends an other error report and you write:

> This is a known bug in inets-4.0 that appears when
> the http response is chunk decoded. It has already been fixed and
> patched for commercial customers and will be part of the next open
> source release.



Yuk ....


/klacke


-- 
Claes Wikstrom                        -- Caps lock is nowhere and
http://www.hyber.org                  -- everything is under control          


From mickael.remond@REDACTED  Wed Dec  8 22:13:15 2004
From: mickael.remond@REDACTED (Mickael Remond)
Date: Wed, 08 Dec 2004 22:13:15 +0100
Subject: repos download slow
In-Reply-To: <1102331002.41b43c7aac419@horde.scannet.dk>
References: <1102331002.41b43c7aac419@horde.scannet.dk>
Message-ID: <41B76E6B.2050503@erlang-fr.org>

bry@REDACTED wrote:
> 
> Whenever I go to download Repos it is so damn slow I decide to forego the
> experience, this is over ADSL. Anybody else been complaining about this?

Hello,

You can now download it with Bittorrent:

http://download.erlang-projects.org/repos-1.0_rc1.iso.torrent

I hope this help. Bandwith will depend on the number of client 
downloading the iso image.

I hope this helps,

-- 
Micka?l R?mond
  http://www.erlang-projects.org/


From henk@REDACTED  Thu Dec  9 08:02:45 2004
From: henk@REDACTED (Henk Bijker)
Date: Thu, 09 Dec 2004 09:02:45 +0200
Subject: bad apples in mnesia db
Message-ID: <1102575765.6752.10.camel@henk.erlfinsys.net>

Hi all, 

Is there anywhere I can find more information regarding the following
message?

Any idea's on how to prevent this from happening? 

Thanks in advance!
Henk.



(acb_application@REDACTED)3>
Mnesia('acb_application@REDACTED
acard.net'): {unpaids_bin,dcd2ets} repaired, lost 5 bad bytes
(acb_application@REDACTED)3>
Mnesia('acb_application@REDACTED
acard.net'): {unpaids_bin,dcd2ets} repaired, lost 348 bad bytes
(acb_application@REDACTED)3>
Mnesia('acb_application@REDACTED
acard.net'): {unpaids_bin,dcd2ets} repaired, lost 245 bad bytes
(acb_application@REDACTED)3>





From ingela@REDACTED  Thu Dec  9 09:45:48 2004
From: ingela@REDACTED (Ingela Anderton)
Date: Thu, 9 Dec 2004 09:45:48 +0100
Subject: [Fwd: http client hell]
References: <41B650C6.6010201@free.fr>
	<16822.49718.29049.233958@gargle.gargle.HOWL>
	<41B70E74.6030104@free.fr>
	<16823.7685.261964.864904@gargle.gargle.HOWL>
	<20041208202305.GA11881@hyber.org>
Message-ID: <16824.4284.274260.952091@gargle.gargle.HOWL>


klacke@REDACTED wrote:
> 
> Wow, is this a shitty attitude or what.  Here is this guy who finds 
> bugs in the http client and you first write:
Or maybe it is you that is bit touchy ;) Of course everyone makes
mistakes, and naturally anybody would be a little suspicions of a 
trouble report where the test program does not work!

I just want to say for the record that we are greatful for all trouble
reports, and I am sorry if I made anybody feel otherwise.

> > I fixed your code above as it did not compile. And this is the result I get
> > every time I try it with intes-4.0.1                    
> > ...
Well that is perfectly true! I cut the test program provided in the
mail and put it in the module to try and repeat the problem and it
would not compile! (You can try it yourself if you do not believe me.)
I know sometimes the truth is not what you want to hear ....

> indicating that there is indeed no problem found, although it was
> obvious from the error_report that the client was either
> wrong or overly mis-forgiving to some http server behaviors.
I did ask which version he was using? Saying the problem does not
seem to appear in the latest version that is 4.0.1! If the error
message had not been cut of originally I could have seen it was
the already known bug in inets-4.0 from the start. 

> Then blaguehubert@REDACTED sends an other error report and you write:
Yes as I asked him too, since the whole error message was not part of
the first report!

> > This is a known bug in inets-4.0 that appears when
> > the http response is chunk decoded. It has already been fixed and
> > patched for commercial customers and will be part of the next open
> > source release.
> 
> 
> 
> Yuk ....
And what is yuk about that? I is a straight forward answer!
I actually was a little surprised that the patch had not made it
to the open source version yet, the patch must just have missed the
last open source release.

/Ingela

Ericsson AB - OTP team








From dgud@REDACTED  Thu Dec  9 10:21:22 2004
From: dgud@REDACTED (Dan Gudmundsson)
Date: Thu, 9 Dec 2004 10:21:22 +0100
Subject: bad apples in mnesia db
In-Reply-To: <1102575765.6752.10.camel@henk.erlfinsys.net>
References: <1102575765.6752.10.camel@henk.erlfinsys.net>
Message-ID: <16824.6418.373109.474630@rian.du.uab.ericsson.se>



Check the code :-)

Sorry, forgot to reply..

Actually its the number bad bytes found when repairing bad data files,
this indicates that the files where not closed properly and may
contain bad/old data, it's normally ok but it's a warning so that you
won't be so suprised if the data is inconsistent.

/Dan

Henk Bijker writes:
 > Hi all, 
 > 
 > Is there anywhere I can find more information regarding the following
 > message?
 > 
 > Any idea's on how to prevent this from happening? 
 > 
 > Thanks in advance!
 > Henk.
 > 
 > 
 > 
 > (acb_application@REDACTED)3>
 > Mnesia('acb_application@REDACTED
 > acard.net'): {unpaids_bin,dcd2ets} repaired, lost 5 bad bytes
 > (acb_application@REDACTED)3>
 > Mnesia('acb_application@REDACTED
 > acard.net'): {unpaids_bin,dcd2ets} repaired, lost 348 bad bytes
 > (acb_application@REDACTED)3>
 > Mnesia('acb_application@REDACTED
 > acard.net'): {unpaids_bin,dcd2ets} repaired, lost 245 bad bytes
 > (acb_application@REDACTED)3>
 > 
 > 

-- 
Dan Gudmundsson               Project:    Mnesia, Erlang/OTP
Ericsson Utvecklings AB       Phone:      +46  8 727 5762 
UAB/F/P                       Mobile:     +46 70 519 9469
S-125 25 Stockholm            Visit addr: Armborstv 1 



From thantos@REDACTED  Thu Dec  9 13:01:17 2004
From: thantos@REDACTED (Alexander Williams)
Date: Thu, 9 Dec 2004 07:01:17 -0500
Subject: Vector Normalization Functions in Erlang
Message-ID: <577610623.20041209070117@speedfactory.net>

Hello erlang-questions,

  I'm  hacking  around  with  some  distributed approaches to a vector
  search  engine  using  Erlang,  and wondered if anyone'd already put
  together  a  neat module for managing sparse vectors efficiently, as
  well    as    things    like    normalization   to   unit   lengths,
  addition/subtraction, etc.

  Why reinvent the wheel when you can reinvent the engine, right?

-- 
Best regards,
 Alexander                          mailto:thantos@REDACTED



From matthias@REDACTED  Thu Dec  9 13:56:15 2004
From: matthias@REDACTED (Matthias Lang)
Date: Thu, 9 Dec 2004 13:56:15 +0100
Subject: Broken debian package ?
In-Reply-To: <41B432DC.9030901@dolphian.com>
References: <41B432DC.9030901@dolphian.com>
Message-ID: <16824.19311.434030.993609@antilipe.corelatus.se>


olivier writes:

 > 	I've been getting a segfault for some days now, using the
 > debian-packaged erlang:
[...]
 > 	Today's package is: erlang-base    10.b.1a-2

 > 	Just wondering if this happens to someone else..

The version in 'unstable' today (10.b.1a-2) crashes for me too---it
doesn't even get as far as a shell prompt.

The version in 'testing' works fine (albeit on another machine).

The source I compiled myself also works fine. I normally use a
source-compiled version, so I didn't notice that the debian one had
gone south.

I don't know if the current debian maintainer reads this list, I've
CCed him, maybe he already knows it's a problem.

Matthias



From maruthavanan_s@REDACTED  Thu Dec  9 15:09:02 2004
From: maruthavanan_s@REDACTED (maruthavanan s)
Date: Thu, 09 Dec 2004 19:39:02 +0530
Subject: problem in Package building
Message-ID: <BAY1-F242514DC0C8A70F6059B0DE3B70@phx.gbl>

hi

i am interfacing c and erlang in windows xp and i use erl_interface nodes 
for the communication between c node and erlang nodes..
i have completed my coding successfully....
i am having all the erl's and beam's of code in the same directory...

i tried the following command with...these were executed successfully with 
no errors...

application:load(sasl).
application:load(listener_app).
application:start(sasl).
application:start(listener_app).
systools:make_script("listener_rel").
systools:script2boot("listener_rel").
systools:make_tar("listener_rel").

i able to make a tar file and i extract my tar  file in a directory
e:\A

then i create a shortcut in windows with parameter as follows...
"E:\Program Files\erl5.3\bin\werl.exe" -boot 
E:/a/listener_rel.tar/releases/1/start
Erlang is installed in the above path
this is where my problem is it shows a crash report showing after the 
progress reports of (sasl)
where i have gone wrong Please help me


Thanks
maruthavanan

_________________________________________________________________
All the news that matters. Just the way you like it. 
http://www.msn.co.in/News/ Only at MSN News!



From maruthavanan_s@REDACTED  Thu Dec  9 15:39:49 2004
From: maruthavanan_s@REDACTED (maruthavanan s)
Date: Thu, 09 Dec 2004 14:39:49 +0000
Subject: Problem in building package
Message-ID: <BAY1-F71CB8EADE5C6999E43D36E3B70@phx.gbl>

hi
i am trying to build an package with my code my code depends on  c exe build 
with erlang interface and on ini file with windows xp
i have pasted my ini and c exe file on the erl.exe directory and also in 
working directory where i have all the beam and erl files

i have created the script,boot,tar files succesfully with no error report
when i extracted my tar file in E:/A directory
and created a short cut as
"E:\Program Files\erl5.3\bin\werl.exe" -boot 
E:/a/listener_rel.tar/releases/1/start

the file  start.boot is located in e:/releases/1/start.boot

\but when i try to eecute this i getting the following error

=INFO REPORT==== 9-Dec-2004::20:05:35 ===
    application: listener_app
    exited: "invalid return value from listener_app:start(normal,[]) -> 
{'EXIT',
\n                                     {undef,\n
         [{listener_app,start,[normal,[]]},\n
       {application_master,start_it_old,4}]}}"
    type: permanent
{"Kernel pid terminated",application_controller,shutdown}


what could be wrong?
Thanks
Maruthavanan

_________________________________________________________________
FREE pop-up blocking with the new MSN Toolbar ? get it now! 
http://toolbar.msn.click-url.com/go/onm00200415ave/direct/01/



From Bjorn.Ericsson@REDACTED  Thu Dec  9 16:15:19 2004
From: Bjorn.Ericsson@REDACTED (Ericsson, Bjorn)
Date: Thu, 9 Dec 2004 16:15:19 +0100 
Subject: einval error when gen_tcp:recv after successful gen_tcp:send
Message-ID: <4CA2E76CAAE7F44186EEF0CFE4C0AB8F0B996E@nobel.logica.co.uk>

Hello!
 
This is my first week programming Erlang, so please be patient:
 
I'm trying to make an Erlang program that communicate over one open socket
by using gen_tcp. There is a server on the other end that I don't have any
control over, so I've also made a simulator for it in Erlang for the time
being. The server interface specifies that the client, my program, initiates
communication by sending a command, and then keeps the socket open and waits
for the server to reply. 
 
I have managed to get as far as my client sends a command, my server
simulator (run on same PC, but in other shell) receives that message and
sends a response, where gen_tcp:send returns ok. The problem arise when the
client tries to use gen_tcp:recv. This is where it receives {error, einval}.
 
How do you suggest I overcome this problem?
 
CLIENT CODE:
-module(interface).
-export([request/2]).
-include("interface.hrl").
 
request(Msisdn, Port) ->
%%% Build  request from input data %%%
   Request = build_request(Msisdn),
%%% Send request and wait for response %%%   
   Response = get_response(Request, Port),
   
build_request(Msisdn) ->
"0045006" ++ Msisdn ++ "46708998851FFFFFFFFF00".
 
get_response(Request, Port) ->
%%% Setup connection %%% 
  case gen_tcp:connect("localhost", Port,
                               [list, {packet, 0},{reuseaddr, true}]) of
    {ok, Sock} ->
    %%% Send request %%%               
        case gen_tcp:send(Sock, Request) of
            ok ->
                io:format("Message sent OK~n");
            {error, Reason} ->
                io:format("Send failed(~p)~n", [Reason])
        end,
    %%% Receive response %%%
        Response = receive_response(Sock),
        case gen_tcp:close(Sock) of
            ok ->
                io:format("Closing connection~n");
            {error, Reason2} ->
                io:format("Cannot close (~p)~n", [Reason2])
        end,
        Response;
    {error, Reason} ->
        io:format("Cannot connect to server (~p)~n", [Reason])
        % ToDo: handle this problem
  end.
  
receive_response(Sock) ->
    io:format("Sock (~p)~n", [Sock]),
    case gen_tcp:recv(Sock, 4) of
        {ok, B} ->
            Size = list_to_integer(B),
            io:format("Receiving message of size (~p).~n", [Size]);
        {error, Reason} ->
            Size = 0,
            io:format("Cannot receive Size (~p)~n", [Reason])
    end,
    case gen_tcp:recv(Sock, Size) of
        {ok, Message} ->
            io:format("Receiving message (~p).~n", [Message]),
            Message;
        {error, Reason2} ->
          io:format("Cannot receive (~p)~n", [Reason2])
    end.

SERVER SIMULATOR CODE:
-module(simulator).
-export([start/1]).
 
start(Port) ->
    case gen_tcp:listen(Port, [list, {packet, 0}, 
                  {active, false},{reuseaddr, true}]) of
      {ok, LSock} ->
        io:format("OK to listen~n"),
        case gen_tcp:accept(LSock) of
          {ok, Sock} ->
            io:format("OK to accept~n"),
            do_recv(Sock),
            send_response("46708797851FFFFFFFFF", Sock), 
            case gen_tcp:close(Sock) of
                ok ->
                    io:format("Closing connection~n");
                {error, Reason2} ->
                    io:format("Cannot close (~p)~n", [Reason2])
            end;
          {error, _} ->
            io:format("Cannot accept for server~n")
         end;     
      {error, Reason} ->
        io:format("Cannot listen for server (~p)~n", [Reason])
    end.
 
do_recv(Sock) ->
    io:format("Socket used (~p)~n", [Sock]),
    MessageSize = receive_size(Sock),
    Message = receive_message(Sock, MessageSize),
    io:format("Received message (~p)~n", [Message]),
    Message.
    
receive_size(Sock) ->
    case gen_tcp:recv(Sock, 4) of
        {ok, B} ->
            Size = list_to_integer(B),
            io:format("Receiving message of size (~p)~n", [Size]),
            Size;
        {error, Reason} ->
            io:format("Cannot receive (~p)~n", [Reason])
    end.
 
receive_message(Sock, MessageSize) ->
    case gen_tcp:recv(Sock, MessageSize) of
        {ok, Message} ->
            Message;
        {error, Reason} ->
            io:format("Cannot receive (~p)~n", [Reason])
    end.

send_response(Msisdn, Sock) ->
   Response = "0069009" ++ Msisdn ++ 
   "46708889951FFFFFFFFFNot Enough funds        08",
   case gen_tcp:send(Sock, Response) of
        ok ->
            io:format("Message sent (~p).~n", [Response]);
        {error, Reason} ->
            io:format("Cannot send (~p)~n", [Reason])
    end.



This e-mail and any attachment is for authorised use by the intended recipient(s) only. It may contain proprietary material, confidential information and/or be subject to legal privilege. It should not be copied, disclosed to, retained or used by, any other party. If you are not an intended recipient then please promptly delete this e-mail and any attachment and all copies and inform the sender. Thank you.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://erlang.org/pipermail/erlang-questions/attachments/20041209/1bf84db3/attachment.htm>

From klacke@REDACTED  Thu Dec  9 19:03:13 2004
From: klacke@REDACTED (klacke@REDACTED)
Date: Thu, 9 Dec 2004 19:03:13 +0100
Subject: einval error when gen_tcp:recv after successful gen_tcp:send
In-Reply-To: <4CA2E76CAAE7F44186EEF0CFE4C0AB8F0B996E@nobel.logica.co.uk>
References: <4CA2E76CAAE7F44186EEF0CFE4C0AB8F0B996E@nobel.logica.co.uk>
Message-ID: <20041209180313.GA24997@hyber.org>

On Thu, Dec 09, 2004 at 04:15:19PM +0100, Ericsson, Bjorn wrote:


>    when the client tries to use gen_tcp:recv. This is where it receives
>    {error, einval}.
> 
> 
>    get_response(Request, Port) ->
>    %%% Setup connection %%%
>      case gen_tcp:connect("localhost", Port,
>                                   [list, {packet, 0},{reuseaddr, true}])



Your socket is (as they are by default) in active mode, this means
that it's not allowed to call recv() on the socket, instead of
actively recv()'ing your process will get messages on the form of:

        {tcp, Socket, Data}


So either you have tp specify {active, false} as a socket option
when you create the socket, or rewrite your code to receive
messages instead of actively recv() 'ing.

recv() and {active , false} has the advantage that you
get flow control for free (through TCP/IP) whereas 
{active, true} has the advantage that it's more flexible.


Read erlang man pages for 'gen_tcp' and 'inet' several times.


/klacke


-- 
Claes Wikstrom                        -- Caps lock is nowhere and
http://www.hyber.org                  -- everything is under control          


From sean.hinde@REDACTED  Thu Dec  9 23:52:06 2004
From: sean.hinde@REDACTED (Sean Hinde)
Date: Thu, 9 Dec 2004 22:52:06 +0000
Subject: einval error when gen_tcp:recv after successful gen_tcp:send
In-Reply-To: <20041209180313.GA24997@hyber.org>
References: <4CA2E76CAAE7F44186EEF0CFE4C0AB8F0B996E@nobel.logica.co.uk> <20041209180313.GA24997@hyber.org>
Message-ID: <F3977888-4A34-11D9-A37E-000A95927CCE@mac.com>


On 9 Dec 2004, at 18:03, klacke@REDACTED wrote:

> On Thu, Dec 09, 2004 at 04:15:19PM +0100, Ericsson, Bjorn wrote:
>
>
>>    when the client tries to use gen_tcp:recv. This is where it 
>> receives
>>    {error, einval}.
>>
>>
>>    get_response(Request, Port) ->
>>    %%% Setup connection %%%
>>      case gen_tcp:connect("localhost", Port,
>>                                   [list, {packet, 0},{reuseaddr, 
>> true}])
>
>
>
> Your socket is (as they are by default) in active mode, this means
> that it's not allowed to call recv() on the socket, instead of
> actively recv()'ing your process will get messages on the form of:
>
>         {tcp, Socket, Data}
>
>
> So either you have tp specify {active, false} as a socket option
> when you create the socket, or rewrite your code to receive
> messages instead of actively recv() 'ing.
>
> recv() and {active , false} has the advantage that you
> get flow control for free (through TCP/IP) whereas
> {active, true} has the advantage that it's more flexible.
>
>
> Read erlang man pages for 'gen_tcp' and 'inet' several times.
>

Also..

One very nice way to use sockets is with the use of the now well 
documented {active, once} option. This allows you to get the benefit of 
flow control, but retain the nice semantics of packets being just 
another message into your process.

call inet:setopts(S, [{active, once}]) each time after receiving a 
packet when you are ready to receive the next one.

Sean



From vlad_dumitrescu@REDACTED  Fri Dec 10 11:27:50 2004
From: vlad_dumitrescu@REDACTED (Vlad Dumitrescu)
Date: Fri, 10 Dec 2004 11:27:50 +0100
Subject: HOMEDRIVE and HOMEDIR
Message-ID: <BAY15-DAV113218891537FE2EAFF4FA9CA80@phx.gbl>

Hi,

Why are these environment variables used (on Windows), and not only HOME? It's
easy to get them out of synch when moving the home directory...

regards,
Vlad


From dietmar@REDACTED  Fri Dec 10 12:21:24 2004
From: dietmar@REDACTED (Dietmar Admin Schaefer)
Date: Fri, 10 Dec 2004 12:21:24 +0100
Subject: snmp-mnesia Questions about MIB-Tables
Message-ID: <41B986B4.9040307@ast.dfs.de>

Hi !

I try to make a snmp "controller application"



I need some advices about the following problems:

My MIB contains a couple of (static) tables as well as  a few scalar 
variables.


I created mnesia tables - one for the scalar variables:




 case mnesia:create_table(fourd_mib_variables, [{snmp, [{key, integer}]},
                 {attributes, record_info(fields, fourd_mib_variables)},
                 {disc_copies, ?HR}]) of
         {atomic, ok}     -> io:format("table fourd_mib_variables could 
be created~n"),
                             ok;
         {aborted,Reason} -> io:format("creation of database 
fourd_mib_variables failed reason.~w~n",[Reason]),
                             Reason
       end,

 an one for every MIB-table:

case mnesia:create_table(criticalHrAProcessesList, [ 
                                                                        
  {attributes, record_info(fields,criticalHrAProcessesList )},
                                                                        
  {disc_copies, ?HR}]) of
        {atomic, ok}     -> io:format("table active_processes could be 
created~n"),
                            ok;
        {aborted,Reason2} -> io:format("creation of database 
criticalHrAProcessesList failed reason.~w~n",[Reason2]),
                                      Reason2
     end,


1. question :   is this a good idea ?


I can initialize the fourd_mib_variables table with no problems but it's 
not clear to me how to initialize the
 criticalHrAProcessesList table.


I tried to ask my agent without initialization and get the following 
error message:


 cmmc_4dp:start().                             starting the application


reading config file 'cmmc.config'
sendPort               = 1234
listenPort             = 1235
sysContact             = "dietmar.schaefer"
mibActive              = 1
location               = "Frankfurt"
mibVersion             = "1.2 (19.10.2004)"
activeSwVersion        = "1.4 (17.10.2004)"
inactiveSwVersion      = "1.3 (15.6.2004)"
maxCriticalProcesses   = 10
maxNonriticalProcesses = 8
maxInternalInterfaces  = 3
maxExternalInterfaces  = 3
maxApproachClients     = 10
maxCenterClients       = 30
sendport = 1234
listenport = 1235
initialisiere CMMC-Datenbank
success writing mnesia
success writing mnesia
CMMC-Datenbank initialisiert !
could not start snmp reason: {already_started,snmp}
ok
(cmmc@REDACTED)22> Src = 2

asking the agent:

snmpwalk -m +FOURD-1-2-MIB -Ob -Cc  -v1 -c public wks2:1161 .enterprises



Mnesia(cmmc@REDACTED): <0.529.0> ({registered_name,snmp_master_agent}) 
val(mnesia_gvar, {criticalHrAProcessesList,{index,snmp}}) -> {badarg,
                                                                                   
[{ets,
                                                                                    
lookup_element,
                                                                                    
[mnesia_gvar,
                                                                                   
{criticalHrAProcessesList,
                                                                                   
{index,
                                                                                    
snmp}},
                                                                                          

                                                                                                                                  
{mnesia_lib,
                                                                                                                                         
val,
                                                                                                                                         
1},
                                                                                                                                     
{mnesia_snmp_hook,
                                                                                                                                         
get_next_index,
                                                                                                                                         
2},
                                                                                                                                     
{rpc,
                                                                                                                                         
local_call,
                                                                                                                                         
3},
                                                                                                                                     
{mnesia,
                                                                                                                                         
do_dirty_rpc,
                                                                                                                                         
5},
                                                                                                                                     
{snmp_generic_mnesia,
                                                                                                                                         
table_next,
                                                                                                                                         
2},
                                                                                                                                     
{snmp_generic,
                                                                                                                                         
handle_table_next,
                                                                                                                                         
6},
                                                                                                                                     
{snmp_generic,
                                                                                                                                         
handle_table_next,
                                                                                                                                         
6},
                                                                                                                                     
{snmpa_agent,
                                                                                                                                         
get_next_values_all_rows,
                                                                                                                                         
6},
                                                                                                                                     
{snmpa_agent,
                                                                                                                                         
get_next_table,
                                                                                                                                         
4}]} {no_exists,
                                                                                                                                                  
{criticalHrAProcessesList,
                                                                                                                                                      
{index,
                                                                                                                                                          
snmp}}}

=ERROR REPORT==== 10-Dec-2004::12:15:23 ===
** User error: Invalid return value 
{'EXIT',{aborted,{badarg,[criticalHrAProcessesList,[]]}}} from 
{snmp_generic,table_func,[{criticalHrAProcessesList,mnesia}]} (get_next)







Can someone point me to the right direction ?


Many thanks in advance



Dietmar


From vlad_dumitrescu@REDACTED  Fri Dec 10 13:32:39 2004
From: vlad_dumitrescu@REDACTED (Vlad Dumitrescu)
Date: Fri, 10 Dec 2004 13:32:39 +0100
Subject: HOMEDRIVE and HOMEDIR
References: <BAY15-DAV113218891537FE2EAFF4FA9CA80@phx.gbl>
Message-ID: <BAY15-DAV14694547E9D27687D2FD8F9CA80@phx.gbl>

To answer myself, it's a standard Windows tricky thing...

I still think the HOME variable should be used if present, however. :-)

/Vlad


----- Original Message ----- 
From: "Vlad Dumitrescu" <vlad_dumitrescu@REDACTED>
To: "- Erlang Questions" <erlang-questions@REDACTED>
Sent: Friday, December 10, 2004 11:27 AM
Subject: HOMEDRIVE and HOMEDIR


> Hi,
>
> Why are these environment variables used (on Windows), and not only HOME? It's
> easy to get them out of synch when moving the home directory...
>
> regards,
> Vlad
>


From blaguehubert@REDACTED  Fri Dec 10 17:08:52 2004
From: blaguehubert@REDACTED (blaguehubert)
Date: Fri, 10 Dec 2004 16:08:52 +0000
Subject: [Fwd: http client hell]
In-Reply-To: <16824.4284.274260.952091@gargle.gargle.HOWL>
References: <41B650C6.6010201@free.fr>	<16822.49718.29049.233958@gargle.gargle.HOWL>	<41B70E74.6030104@free.fr>	<16823.7685.261964.864904@gargle.gargle.HOWL>	<20041208202305.GA11881@hyber.org> <16824.4284.274260.952091@gargle.gargle.HOWL>
Message-ID: <41B9CA14.1030105@free.fr>

Everything is ok, I did not feel it was a "shitty attitude"...

I'm a little bored, not by Ingela Anderton, but by the fact that I did 
not have time to wait for the next release and this mean that I will 
have to write my own http client (arg!!!).

Ingela Anderton wrote:

>klacke@REDACTED wrote:
>  
>
>>Wow, is this a shitty attitude or what.  Here is this guy who finds 
>>bugs in the http client and you first write:
>>    
>>
>Or maybe it is you that is bit touchy ;) Of course everyone makes
>mistakes, and naturally anybody would be a little suspicions of a 
>trouble report where the test program does not work!
>
>I just want to say for the record that we are greatful for all trouble
>reports, and I am sorry if I made anybody feel otherwise.
>
>  
>
>>>I fixed your code above as it did not compile. And this is the result I get
>>>every time I try it with intes-4.0.1                    
>>>...
>>>      
>>>
>Well that is perfectly true! I cut the test program provided in the
>mail and put it in the module to try and repeat the problem and it
>would not compile! (You can try it yourself if you do not believe me.)
>I know sometimes the truth is not what you want to hear ....
>
>  
>
>>indicating that there is indeed no problem found, although it was
>>obvious from the error_report that the client was either
>>wrong or overly mis-forgiving to some http server behaviors.
>>    
>>
>I did ask which version he was using? Saying the problem does not
>seem to appear in the latest version that is 4.0.1! If the error
>message had not been cut of originally I could have seen it was
>the already known bug in inets-4.0 from the start. 
>
>  
>
>>Then blaguehubert@REDACTED sends an other error report and you write:
>>    
>>
>Yes as I asked him too, since the whole error message was not part of
>the first report!
>
>  
>
>>>This is a known bug in inets-4.0 that appears when
>>>the http response is chunk decoded. It has already been fixed and
>>>patched for commercial customers and will be part of the next open
>>>source release.
>>>      
>>>
>>
>>Yuk ....
>>    
>>
>And what is yuk about that? I is a straight forward answer!
>I actually was a little surprised that the patch had not made it
>to the open source version yet, the patch must just have missed the
>last open source release.
>
>/Ingela
>
>Ericsson AB - OTP team
>
>
>
>
>
>
>
>
>  
>



From erlang@REDACTED  Fri Dec 10 22:35:34 2004
From: erlang@REDACTED (Michael McDaniel)
Date: Fri, 10 Dec 2004 13:35:34 -0800
Subject: choosing distributed comm ports
Message-ID: <20041210213534.GG19273@fangora.autosys.us>

Thank you for Erlang, and thank you for the educational reading on this
list.

I have some Erlang nodes (R10B-1a) that use SSL for distributed
communications (configured and working per the excellent
otp_doc_html_R10B-1a.tar.gz documentation).  The following refers to SSL
enabled node communications.

I have some remote firewalled computers providing net services.  I want to
install Erlang nodes on those computers for remote monitoring and control.

I have noticed that the ports used when my nodes communicate are not always
the same (not a surprise).  I would like to have known fixed ports so that
I can open those specific ports on the various firewalls where my Erlang
nodes will be installed.

I do not want to write custom port communications applications to do port 
control.  Rather, I want to use the common distributed functions and semantics
(auth:node_cookie(), spawn(), monitor(), receive,  A ! {msg,[]},  etc.).

QUESTION:
Is there an easy way (and what is it) to force specific ports to be used 
during distributed communications when using the common distributed
functions and semantics?

If the answer is located in the documentation, please point me to it.  I
have not (yet?) found it.


thank you,

Michael McDaniel
Portland, Oregon, USA
http://autosys.us



From daniel.neri@REDACTED  Fri Dec 10 23:17:50 2004
From: daniel.neri@REDACTED (=?ISO-8859-1?Q?Daniel_N=E9ri?=)
Date: Fri, 10 Dec 2004 23:17:50 +0100
Subject: [Fwd: http client hell]
In-Reply-To: <41B9CA14.1030105@free.fr>
References: <41B650C6.6010201@free.fr>
	 <16822.49718.29049.233958@gargle.gargle.HOWL>
	 <41B70E74.6030104@free.fr>
	 <16823.7685.261964.864904@gargle.gargle.HOWL>
	 <20041208202305.GA11881@hyber.org>
	 <16824.4284.274260.952091@gargle.gargle.HOWL>
	 <41B9CA14.1030105@free.fr>
Message-ID: <3036fece04121014173809fce0@mail.gmail.com>

On Fri, 10 Dec 2004 16:08:52 +0000, blaguehubert <blaguehubert@REDACTED> wrote:

> I'm a little bored, not by Ingela Anderton, but by the fact that I did
> not have time to wait for the next release and this mean that I will
> have to write my own http client (arg!!!).

You can find a simple HTTP client here:

  http://www.erlang.org/examples/small_examples/index.html#1


Regards
-- 
Daniel N?ri
daniel.neri@REDACTED


From vances@REDACTED  Fri Dec 10 23:42:59 2004
From: vances@REDACTED (Vance Shipley)
Date: Fri, 10 Dec 2004 17:42:59 -0500
Subject: choosing distributed comm ports
In-Reply-To: <20041210213534.GG19273@fangora.autosys.us>
References: <20041210213534.GG19273@fangora.autosys.us>
Message-ID: <20041210224259.GG55997@frogman.motivity.ca>

Michael,

What you want to do is to create a tunnel.  This is most easily
accomplished as I detailed here:

	http://www.erlang.org/ml-archive/erlang-questions/200402/msg00119.html
	http://www.erlang.org/ml-archive/erlang-questions/200402/msg00120.html

	-Vance

On Fri, Dec 10, 2004 at 01:35:34PM -0800, Michael McDaniel wrote:
}  
}  I have some remote firewalled computers providing net services.  I want to
}  install Erlang nodes on those computers for remote monitoring and control.


From jilani.khaldi1@REDACTED  Sat Dec 11 00:09:16 2004
From: jilani.khaldi1@REDACTED (Jilani Khaldi)
Date: Sat, 11 Dec 2004 00:09:16 +0100
Subject: [Fwd: http client hell]
In-Reply-To: <3036fece04121014173809fce0@mail.gmail.com>
References: <41B650C6.6010201@free.fr>	 <16822.49718.29049.233958@gargle.gargle.HOWL>	 <41B70E74.6030104@free.fr>	 <16823.7685.261964.864904@gargle.gargle.HOWL>	 <20041208202305.GA11881@hyber.org>	 <16824.4284.274260.952091@gargle.gargle.HOWL>	 <41B9CA14.1030105@free.fr> <3036fece04121014173809fce0@mail.gmail.com>
Message-ID: <41BA2C9C.9000805@virgilio.it>


>You can find a simple HTTP client here:
>
>  http://www.erlang.org/examples/small_examples/index.html#1
>
>
>Regards
>  
>
{error,{socket_error,econnrefused}}

I have tried some http site, but this is the error I get everytime. What 
does it mean?
Thanks!

jk

-- 
Jilani KHALDI
http://jkhaldi.oltrelinux.com/



From erlang@REDACTED  Sat Dec 11 00:44:04 2004
From: erlang@REDACTED (Michael McDaniel)
Date: Fri, 10 Dec 2004 15:44:04 -0800
Subject: choosing distributed comm ports
In-Reply-To: <20041210224259.GG55997@frogman.motivity.ca>
References: <20041210213534.GG19273@fangora.autosys.us> <20041210224259.GG55997@frogman.motivity.ca>
Message-ID: <20041210234404.GK19273@fangora.autosys.us>

On Fri, Dec 10, 2004 at 05:42:59PM -0500, Vance Shipley wrote:
> Michael,
> 
> What you want to do is to create a tunnel.  This is most easily
> accomplished as I detailed here:
> 
> 	http://www.erlang.org/ml-archive/erlang-questions/200402/msg00119.html
> 	http://www.erlang.org/ml-archive/erlang-questions/200402/msg00120.html
> 
> 	-Vance
> 
> On Fri, Dec 10, 2004 at 01:35:34PM -0800, Michael McDaniel wrote:
> }  
> }  I have some remote firewalled computers providing net services.  I want to
> }  install Erlang nodes on those computers for remote monitoring and control.
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

Thank you, Vance.  That would certainly work, though I would rather reserve
tunneling with SSH for installations where the firewall configuration is
less flexible.

Since I have access to the firewalls, I would rather open up some high port that
my SSL Erlang nodes are 'guarding', waiting for a connection from/to another
SSL Erlang node, and refusing all other access.  

Hence the need to pre-determine and control what ports will be used by the 
SSL Erlang nodes.

~Michael




From bfulgham@REDACTED  Sat Dec 11 08:15:27 2004
From: bfulgham@REDACTED (Brent Fulgham)
Date: Fri, 10 Dec 2004 23:15:27 -0800
Subject: Broken debian package ?
In-Reply-To: <16824.19311.434030.993609@antilipe.corelatus.se>
References: <41B432DC.9030901@dolphian.com> <16824.19311.434030.993609@antilipe.corelatus.se>
Message-ID: <200412102315.27615.bfulgham@debian.org>

On Thursday 09 December 2004 04:56 am, Matthias Lang wrote:
>  >  Today's package is: erlang-base    10.b.1a-2
>  >
>  >  Just wondering if this happens to someone else..
>
> The version in 'unstable' today (10.b.1a-2) crashes for me too---it
> doesn't even get as far as a shell prompt.

I do read the list, and I am aware of the problem, but I can't recreate the 
issue, nor identify what's unique about the hardware of the people 
encountering the problem.

> The source I compiled myself also works fine. I normally use a
> source-compiled version, so I didn't notice that the debian one had
> gone south.

I had someone compile from the Debian souce package, and they continued to 
have crashes, so it does not appear to be something unique about my build 
environment.

Hmmm....

Do you build with HIPE enabled?  I do, which might be a contributing factor...

-Brent


From klacke@REDACTED  Sat Dec 11 14:33:44 2004
From: klacke@REDACTED (klacke@REDACTED)
Date: Sat, 11 Dec 2004 14:33:44 +0100
Subject: choosing distributed comm ports
In-Reply-To: <20041210213534.GG19273@fangora.autosys.us>
References: <20041210213534.GG19273@fangora.autosys.us>
Message-ID: <20041211133344.GA28543@hyber.org>

On Fri, Dec 10, 2004 at 01:35:34PM -0800, Michael McDaniel wrote:

> 
> I have noticed that the ports used when my nodes communicate are not always
> the same (not a surprise).  I would like to have known fixed ports so that
> I can open those specific ports on the various firewalls where my Erlang
> nodes will be installed.


There are three interesting environment variable read by the code
in inet_tcp_dist.erl which are used to control the establishment
of the listen socket for distributed erlang.

 application:get_env(kernel,inet_dist_listen_min) and
 application:get_env(kernel,inet_dist_listen_max)


    are used to set a range of the incoming portnumbers for 
    distribution, use this to force one (or a range) of
    explicit port numbers. These portnumbers can then be
    fed into firewalls

 application:get_env(kernel, inet_dist_use_interface)

    Is another security related env parameter. It's used
    to bind() the distribution to an explicit interface 
    on the box. Also useful for security reasons.



/klacke


-- 
Claes Wikstrom                        -- Caps lock is nowhere and
http://www.hyber.org                  -- everything is under control          


From joe@REDACTED  Sat Dec 11 15:09:40 2004
From: joe@REDACTED (Joe Armstrong)
Date: Sat, 11 Dec 2004 15:09:40 +0100 (CET)
Subject: [Fwd: http client hell]
In-Reply-To: <41B9CA14.1030105@free.fr>
References: <41B650C6.6010201@free.fr> <16822.49718.29049.233958@gargle.gargle.HOWL>
 <41B70E74.6030104@free.fr> <16823.7685.261964.864904@gargle.gargle.HOWL>
 <20041208202305.GA11881@hyber.org> <16824.4284.274260.952091@gargle.gargle.HOWL>
 <41B9CA14.1030105@free.fr>
Message-ID: <Pine.LNX.4.61.0412111456240.31670@enfield.sics.se>



On Fri, 10 Dec 2004, blaguehubert wrote:

> Everything is ok, I did not feel it was a "shitty attitude"...
>
> I'm a little bored, not by Ingela Anderton, but by the fact that I did not 
> have time to wait for the next release and this mean that I will have to 
> write my own http client (arg!!!).

Well hopefully not :-) - I think we (ie the people on the list might be 
able to help)

Just to clarify - what do you want? - the expression "http client" could 
mean many different things:

The simplest is:

 	get_url("http://www...") which just gets a web page

This is useful for simple applications but soon breaks down for more 
complex stuff.

A better approach is to open an HTTP stream and then send and receive 
requests on the stream. Basically:

 	{ok, Stream} = http:open("http://www....")
 	... requests ...
 	http:close(Stream)

What to do in requests is more problematic. Do you want to to synchronous 
requests, or asynchronous. Do you want to create keep-alive sessions and 
handle chucked data streams or streaming media etc.

Do you want your client to preform redirects etc?

If you could just say in a sentence or so what kind of client you want I'm 
sure somebody here might be able to help, if it's just the simpler
form (get_url) I'll post my latest version (which is in daily use)
some of the versions referenced in previous posts are probably not 
up-to-date.

Cheers

/Joe



From klacke@REDACTED  Sat Dec 11 16:42:47 2004
From: klacke@REDACTED (klacke@REDACTED)
Date: Sat, 11 Dec 2004 16:42:47 +0100
Subject: yaws 1.50
Message-ID: <20041211154247.GA30818@hyber.org>



Folks, 

Yaws 1.50

I've made a new release. Release notes, code  as well as a new shiny look of
the website at http://yaws.hyber.org
I've learnt CSS and, well it rocks.



/klacke 

-- 
Claes Wikstrom                        -- Caps lock is nowhere and
http://www.hyber.org                  -- everything is under control          


From esbjorn@REDACTED  Sat Dec 11 17:24:16 2004
From: esbjorn@REDACTED (=?ISO-8859-1?Q?Esbj=F6rn_K=F6hler?=)
Date: Sat, 11 Dec 2004 17:24:16 +0100
Subject: gentoo ebuild of R10B-1
Message-ID: <20041211172416.6ceadc80@localhost>

Hi,

Does anyone know who maintains the gentoo ebuild packages for erlang?
The (still masked) erlang-10.2.0.ebuild is from 8/10, and I figure there
is a newer version (R10B-1.)

On a side note, the 10.2.0 ebuild still creates a segfaulting erlang for
me, so I thought maybe a newer version would work better.

/esbj?rn


From erlang@REDACTED  Sun Dec 12 07:32:47 2004
From: erlang@REDACTED (Michael McDaniel)
Date: Sat, 11 Dec 2004 22:32:47 -0800
Subject: choosing distributed comm ports
In-Reply-To: <20041211133344.GA28543@hyber.org>
References: <20041210213534.GG19273@fangora.autosys.us> <20041211133344.GA28543@hyber.org>
Message-ID: <20041212063247.GB1054@fangora.autosys.us>

On Sat, Dec 11, 2004 at 02:33:44PM +0100, klacke@REDACTED wrote:
> On Fri, Dec 10, 2004 at 01:35:34PM -0800, Michael McDaniel wrote:
> 
> > 
> > I have noticed that the ports used when my nodes communicate are not always
> > the same (not a surprise).  I would like to have known fixed ports so that
> > I can open those specific ports on the various firewalls where my Erlang
> > nodes will be installed.
> 
> 
> There are three interesting environment variable read by the code
> in inet_tcp_dist.erl which are used to control the establishment
> of the listen socket for distributed erlang.
> 
>  application:get_env(kernel,inet_dist_listen_min) and
>  application:get_env(kernel,inet_dist_listen_max)
> 
> 
>     are used to set a range of the incoming portnumbers for 
>     distribution, use this to force one (or a range) of
>     explicit port numbers. These portnumbers can then be
>     fed into firewalls
> 
>  application:get_env(kernel, inet_dist_use_interface)
> 
>     Is another security related env parameter. It's used
>     to bind() the distribution to an explicit interface 
>     on the box. Also useful for security reasons.
> 
> 
> 
> /klacke
> 
> 
> -- 
> Claes Wikstrom                        -- Caps lock is nowhere and
> http://www.hyber.org                  -- everything is under control          
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

Again, thank you, Claes, for the pointer.  It led to a workable solution.

Please see my original post for more details about the original question.
All my nodes use SSL communications ( -proto_dist inet_ssl ).

I found a way to force inet_ssl to listen on a fixed port, though it
requires a modification of inet_ssl_dist.erl code.  Line 84 of that module
(R10B-1a distro) is
case ssl_prim:listen(0, [{active, false}, {packet,4}] ++

Changing the 0 to my fixed port number causes the listen to occur on the
fixed port when using '-proto_dist inet_ssl'.

Since the epmd server (... /erts/epmd/src/epmd_srv.c) listens on a 
well-known port (presently 4369), I can open up 4369 and my inet_ssl
port on the firewall to allow calling nodes to reach behind-the-firewall
nodes.

Apparently what happens is that the calling node talks first to the epmd
server on the destination node.  I am guessing that the epmd server then
tells the calling node what port to use for the continuing data exchange
(my fixed port number in inet_ssl).  Hence I need to open up these two
ports on the firewall.

So, it works, though modifying distributed source is not my ideal choice.
I do not see any other way to do it if I want to use inet_ssl.

Maybe a future release of inet_ssl_dist.erl will have the 

 application:get_env(kernel,inet_dist_listen_min)
 application:get_env(kernel,inet_dist_listen_max)

capability like inet_dist.


~Michael
http://autosys.us


From robert.balogh@REDACTED  Sun Dec 12 07:43:47 2004
From: robert.balogh@REDACTED (=?iso-8859-1?Q?R=F3bert_Balogh_=28IJ/ETH=29?=)
Date: Sun, 12 Dec 2004 07:43:47 +0100
Subject: No subject
Message-ID: <F005CD411D18D3119C8F00508B0874800F286BDF@ehubunt100.eth.ericsson.se>

Hej,

I'm a beginner in the SNMP v3. Unfortunately I have got a problem. When I configure than start an SNMP erlang agent success from a MIB browser I'd like to initiate that user what I set in the erlang SNMP config file to. I set the SNMP trace to debug in the erlang and after pressing add button in the MIB browser I got this log in the erlang. 

Could anyone please help me about it, because I don't really understand what does it means :-(

Here is the log:
---------------------
(xxAG@REDACTED)3> ** SNMP A-NET-IF LOG:
   got paket from {10,10,10,4}:33017
** SNMP A-NET-IF MPD LOG:
   v3, msgID: 22, msgFlags: [5], msgSecModel: 3
** SNMP A-NET-IF MPD DEBUG: version 3 message header:
   msgID                 = 22
   msgMaxSize            = 8192
   msgFlags              = [5]
   msgSecurityModel      = 3
   msgSecurityParameters = [48,
                            36,
                            4,
                            3,
                            101,
                            73,
                            100,
                            2,
                            1,
                            0,
                            2,
                            1,
                            0,
                            4,
                            7,
                            105,
                            110,
                            105,
                            116,
                            105,
                            97,
                            108,
                            4,
                            12,
                            24,
                            124,
                            135,
                            249,
                            238,
                            183,
                            68,
                            10,
                            244,
                            226,
                            1,
                            123,
                            4,
                            0]
** SNMP A-NET-IF A-USM LOG:
   authEngineID: "eId", userName: "initial"
** SNMP A-NET-IF MPD DEBUG: security module result when reportable [7.2.6-a]:
   Reason:    usmStatsNotInTimeWindows
   ErrorInfo: {{varbind,[1,3,6,1,6,3,15,1,1,2,0],'Counter32',2,undefined},
               "initial",
               [{securityLevel,1}]}
** SNMP A-NET-IF LOG: sending report for reason:
   {securityError,usmStatsNotInTimeWindows}
  


thank You for Your help,
br,
/Robi

ps: Yesterday was everything success but today isn't and I don't change anything...:-( Just what I did before...



From matthias@REDACTED  Sun Dec 12 10:19:21 2004
From: matthias@REDACTED (Matthias Lang)
Date: Sun, 12 Dec 2004 10:19:21 +0100
Subject: Broken debian package ?
In-Reply-To: <200412102315.27615.bfulgham@debian.org>
References: <41B432DC.9030901@dolphian.com>
	<16824.19311.434030.993609@antilipe.corelatus.se>
	<200412102315.27615.bfulgham@debian.org>
Message-ID: <16828.3353.859121.104879@antilipe.corelatus.se>

Brent Fulgham writes:

 > I am aware of the problem, but I can't recreate the issue, nor
 > identify what's unique about the hardware of the people
 > encountering the problem.

I played around a bit more 

  - The verion in debian-unstable crashes during startup on my home PC,
    an old AMD Duron running 2.4.27

  - The version in debian-testing crashes with a bit of provocation
    on a server at work:

      | matthias@REDACTED:~/corelatus/api/erl/test$ /usr/bin/erl
      | Erlang (BEAM) emulator version 5.3.6.3 [source] [hipe] [threads:0]
      |
      | Eshell V5.3.6.3  (abort with ^G)
      | 1> 
      | [1]+  Stopped                 /usr/bin/erl
      | matthias@REDACTED:~/corelatus/api/erl/test$ fg
      | /usr/bin/erl
      | Segmentation fault

    That machine is a Pentium II, also running 2.4.27

That's only two machines, but it doesn't really look hardware related.

The startup from the from-source-compile I usually use:

      | Erlang (BEAM) emulator version 5.3.6.3 [source] [hipe]

So, you have threads enabled and I don't. A fresh compile from R9C-2
source with --enable-threads produces exactly the same symptoms as in
the debian package, i.e. segfault if I ^Z and the 'fg'.

The problem seems to be related to threaded IO. Anyone have some clues
for how to proceed?

Matthias


From robert.balogh@REDACTED  Sun Dec 12 10:42:13 2004
From: robert.balogh@REDACTED (=?iso-8859-1?Q?R=F3bert_Balogh_=28IJ/ETH=29?=)
Date: Sun, 12 Dec 2004 10:42:13 +0100
Subject: snmp v3 problem
Message-ID: <F005CD411D18D3119C8F00508B0874800F286BE0@ehubunt100.eth.ericsson.se>

> Hej,
> 
> I'm a beginner in the SNMP v3. Unfortunately I have got a problem. When I configure than start an SNMP erlang agent success from a MIB browser I'd like to initiate that user what I set in the erlang SNMP config file to. I set the SNMP trace to debug in the erlang and after pressing add button in the MIB browser I got this log in the erlang. 
> 
> Could anyone please help me about it, because I don't really understand what does it means :-(
> 
> Here is the log:
> ---------------------
> (xxAG@REDACTED)3> ** SNMP A-NET-IF LOG:
>    got paket from {10,10,10,4}:33017
> ** SNMP A-NET-IF MPD LOG:
>    v3, msgID: 22, msgFlags: [5], msgSecModel: 3
> ** SNMP A-NET-IF MPD DEBUG: version 3 message header:
>    msgID                 = 22
>    msgMaxSize            = 8192
>    msgFlags              = [5]
>    msgSecurityModel      = 3
>    msgSecurityParameters = [48,
>                             36,
>                             4,
>                             3,
>                             101,
>                             73,
>                             100,
>                             2,
>                             1,
>                             0,
>                             2,
>                             1,
>                             0,
>                             4,
>                             7,
>                             105,
>                             110,
>                             105,
>                             116,
>                             105,
>                             97,
>                             108,
>                             4,
>                             12,
>                             24,
>                             124,
>                             135,
>                             249,
>                             238,
>                             183,
>                             68,
>                             10,
>                             244,
>                             226,
>                             1,
>                             123,
>                             4,
>                             0]
> ** SNMP A-NET-IF A-USM LOG:
>    authEngineID: "eId", userName: "initial"
> ** SNMP A-NET-IF MPD DEBUG: security module result when reportable [7.2.6-a]:
>    Reason:    usmStatsNotInTimeWindows
>    ErrorInfo: {{varbind,[1,3,6,1,6,3,15,1,1,2,0],'Counter32',2,undefined},
>                "initial",
>                [{securityLevel,1}]}
> ** SNMP A-NET-IF LOG: sending report for reason:
>    {securityError,usmStatsNotInTimeWindows}
>   
> 
> 
> thank You for Your help,
> br,
> /Robi
> 
> ps: Yesterday was everything success but today isn't and I don't change anything...:-( Just what I did before...
> 


From maruthavanan_s@REDACTED  Mon Dec 13 15:11:49 2004
From: maruthavanan_s@REDACTED (maruthavanan s)
Date: Mon, 13 Dec 2004 14:11:49 +0000
Subject: Problem in building package
In-Reply-To: <BAY1-F71CB8EADE5C6999E43D36E3B70@phx.gbl>
Message-ID: <BAY1-F33BC85CB69B48C3B0069FDE3AB0@phx.gbl>


hi
i am trying to build an package with my code my code depends on  c exe build 
with erlang interface and on ini file with windows xp
i have pasted my ini and c exe file on the erl.exe directory and also in 
working directory where i have all the beam and erl files

i have created the script,boot,tar files succesfully with no error report
when i extracted my tar file in E:/A directory
and created a short cut as
"E:\Program Files\erl5.3\bin\werl.exe" -boot 
E:/a/listener_rel.tar/releases/1/start

the file  start.boot is located in e:/releases/1/start.boot

\but when i try to eecute this i getting the following error

=INFO REPORT==== 9-Dec-2004::20:05:35 ===
   application: listener_app
   exited: "invalid return value from listener_app:start(normal,[]) -> 
{'EXIT',
\n                                     {undef,\n
        [{listener_app,start,[normal,[]]},\n
      {application_master,start_it_old,4}]}}"
   type: permanent
{"Kernel pid terminated",application_controller,shutdown}


what could be wrong?
Thanks

_________________________________________________________________
Express yourself instantly with MSN Messenger! Download today - it's FREE! 
http://messenger.msn.click-url.com/go/onm00200471ave/direct/01/



From kruegger@REDACTED  Tue Dec 14 02:38:09 2004
From: kruegger@REDACTED (Stephen Han)
Date: Mon, 13 Dec 2004 17:38:09 -0800
Subject: newbie question on Supervisor
Message-ID: <86f1f535041213173812a8f4a3@mail.gmail.com>

Hi

I wrote the process with supervisor behaviour.

Specification :

Spec = {no1100_sup, 
	    {no1100, start_link, [StartArgs]}, 
	    transient, 
	    10, 
	    worker, 
	    [no1100]},

{{ one_for_one, 2, 1 } , [ Spec] }

I wanted to test it so I killed the process from Appmon.
I found out that even though I killed only worker, the supervisor
process also got killed.
I tested on Mnesia, same things happened.
Does this mean that exit(Pid, kill) is not avoidable?

How can I properly test the supervisor process?

regards,


From ingela@REDACTED  Tue Dec 14 09:32:08 2004
From: ingela@REDACTED (Ingela Anderton)
Date: Tue, 14 Dec 2004 09:32:08 +0100
Subject: [Fwd: http client hell]
References: <41B650C6.6010201@free.fr>
	<16822.49718.29049.233958@gargle.gargle.HOWL>
	<41B70E74.6030104@free.fr>
	<16823.7685.261964.864904@gargle.gargle.HOWL>
	<20041208202305.GA11881@hyber.org>
	<16824.4284.274260.952091@gargle.gargle.HOWL>
	<41B9CA14.1030105@free.fr>
Message-ID: <16830.42248.369228.104446@gargle.gargle.HOWL>


blaguehubert wrote:
> I'm a little bored, not by Ingela Anderton, but by the fact that I did 
> not have time to wait for the next release and this mean that I will 
> have to write my own http client (arg!!!).
You could put in the fix yourself while waiting for the new release,
as that particular bug is quite easy to fix.

In httpc_handler.erl change the following 
---------------------------------------
   "chunked" ->
            {ok, {ChunkedHeaders, NewBody}} =
                case http_chunk:decode(Body, State#state.max_body_size, 
                                       State#state.max_header_size) of
                    {Module, Function, Args} ->
                        http_transport:setopts(Session#tcp_session.scheme, 
					       Session#tcp_session.socket, 
					       [{active, once}]),
                        {noreply, State#state{mfa = 
					      {Module, Function, Args}}};
                    Decoded ->
                        Decoded
                end,
            NewHeaders = http_chunk:handle_headers(Headers, ChunkedHeaders),
            handle_response(State#state{headers = NewHeaders, 
					body = NewBody});

---------------
to:
    "chunked" ->
	    case http_chunk:decode(Body, State#state.max_body_size, 
				   State#state.max_header_size) of
		{Module, Function, Args} ->
		    http_transport:setopts(Session#tcp_session.scheme, 
					   Session#tcp_session.socket, 
					   [{active, once}]),
		    {noreply, State#state{mfa = 
					  {Module, Function, Args}}};
		{ok, {ChunkedHeaders, NewBody}} ->
		    NewHeaders = http_chunk:handle_headers(Headers, 
							   ChunkedHeaders),
		    handle_response(State#state{headers = NewHeaders, 
						body = NewBody})
	    end;

-------------------

Of course there where other things fixed for 4.0.1 so I can not
guarantee that it will work flawlessly but it might be worth a shot!

-- 
/Ingela

Ericsson AB - OTP team







From bjorn@REDACTED  Tue Dec 14 12:12:16 2004
From: bjorn@REDACTED (Bjorn Gustavsson)
Date: 14 Dec 2004 12:12:16 +0100
Subject: HOMEDRIVE and HOMEDIR
In-Reply-To: <BAY15-DAV14694547E9D27687D2FD8F9CA80@phx.gbl>
References: <BAY15-DAV113218891537FE2EAFF4FA9CA80@phx.gbl>
	<BAY15-DAV14694547E9D27687D2FD8F9CA80@phx.gbl>
Message-ID: <o1hdmp2kxb.fsf@erix.ericsson.se>

"Vlad Dumitrescu" <vlad_dumitrescu@REDACTED> writes:

> To answer myself, it's a standard Windows tricky thing...
>
> I still think the HOME variable should be used if present, however. :-)

Yes, I agree. We might fix that in a future release.

/Bjorn

-- 
Bj?rn Gustavsson, Erlang/OTP, Ericsson AB


From keymon@REDACTED  Tue Dec 14 13:29:41 2004
From: keymon@REDACTED (=?iso-8859-1?q?H=E9ctor_Rivas_G=E1ndara?=)
Date: Tue, 14 Dec 2004 13:29:41 +0100
Subject: EI libray -> how does creation number work?
Message-ID: <200412141329.41052.keymon@wanadoo.es>


The EI library defines a creation number for the c-nodes published in EPMD, in 
order to prevent the C node from receive messages sent to an earlier c-node 
with the same registered name. 

This creation number can be specified by the user in the  ei_connect_init 
functions, but the description of this function says:

"A C node acting as a server will be assigned a creation number when it calls 
ei_publish()."

but ei_publish() does not return any creation number, and if you read the 
EI-library source code (at /lib/erl_interface/src/epmd/empd_publish.c 
function ei_epmd_r4_publish) the user defined creation number is not send to 
EPMD, but EPMD sends a new creation number that is ignored.

So, How does the creation number work? Has it any use?

-- 
Saludos ;)
Keymon

... Respete la velocidad: 2400 BPS. Las imprudencias se pagan.


From Bjorn.Ericsson@REDACTED  Tue Dec 14 14:44:58 2004
From: Bjorn.Ericsson@REDACTED (Ericsson, Bjorn)
Date: Tue, 14 Dec 2004 14:44:58 +0100
Subject: Creating binary from list variables?
Message-ID: <4CA2E76CAAE7F44186EEF0CFE4C0AB8F0B997B@nobel.logica.co.uk>

Hello, here is another newbie question. I have read chapter 6.16 of the
Reference Manual, but not succeded to solve this problem.

How do I create a binary, to send with gen_tcp:send, from variables of
different types and dynamic length, when I want the final binary to be of
fixed length? (As it usually is defined in socket interfaces)

I keep getting "badarg" when I try different approaches for lists:

sandbox1() ->
    AsciText = "Hello there!",
    Number = 123,
    Binary = <<AsciText:120, Number:32>>,
    Binary.
    
sandbox2() ->
    AsciText = "Hello Sam!",
    AsciAsBinary = list_to_binary(AsciText),
    Number = 1239,
    Binary = <<AsciAsBinary:120, Number:32>>,
    Binary.

/BE

This e-mail and any attachment is for authorised use by the intended recipient(s) only. It may contain proprietary material, confidential information and/or be subject to legal privilege. It should not be copied, disclosed to, retained or used by, any other party. If you are not an intended recipient then please promptly delete this e-mail and any attachment and all copies and inform the sender. Thank you.


From thomasl_erlang@REDACTED  Tue Dec 14 15:35:45 2004
From: thomasl_erlang@REDACTED (Thomas Lindgren)
Date: Tue, 14 Dec 2004 06:35:45 -0800 (PST)
Subject: Creating binary from list variables?
In-Reply-To: <4CA2E76CAAE7F44186EEF0CFE4C0AB8F0B997B@nobel.logica.co.uk>
Message-ID: <20041214143545.66200.qmail@web41901.mail.yahoo.com>


--- "Ericsson, Bjorn" <Bjorn.Ericsson@REDACTED>
wrote:

> How do I create a binary, to send with gen_tcp:send,
> from variables of
> different types and dynamic length, when I want the
> final binary to be of
> fixed length? (As it usually is defined in socket
> interfaces)

As far as I know, you have to pad what you send
yourself:

  %% Length is desired length, F filler byte
  Rem_length = Length-length(Lst),
  list_to_binary([Lst, lists:duplicate(Rem_length,
F)])

or something like that.
Sometimes you can exploit a special case:

   <<0:Rem_length/unsigned-integer-unit:8>>

creates a binary of Rem_length zeroes. 

Neither of these truncate a too long message, of
course. Here is one approach (if you don't care what
gets lost):

  trunc_binary(<<TruncBin:124/binary,_/binary>>) ->
     TruncBin;
  trunc_binary(Bin) -> Bin.

Best,
Thomas



		
__________________________________ 
Do you Yahoo!? 
Meet the all-new My Yahoo! - Try it today! 
http://my.yahoo.com 
 



From vlad_dumitrescu@REDACTED  Tue Dec 14 15:34:47 2004
From: vlad_dumitrescu@REDACTED (Vlad Dumitrescu)
Date: Tue, 14 Dec 2004 15:34:47 +0100
Subject: Creating binary from list variables?
References: <4CA2E76CAAE7F44186EEF0CFE4C0AB8F0B997B@nobel.logica.co.uk>
Message-ID: <BAY15-DAV1257F081A6FC1D535213319CAC0@phx.gbl>

Hi,

One way to do it (not necessarily the best) is

 sandbox3() ->
     AsciText = "Hello Sam!",
     AsciAsBinary = list_to_binary(AsciText),
     PadLen = (15-length(AsciText))*8,
     Number = 1239,
     <<AsciAsBinary/binary, 0:PadLen, Number:32>>.

One observation is that you don't have to make everything a binary, you can send
io lists. Informally, an IO list is a deep list of characters and binaries which
can be sent to an Erlang port.

 sandbox4() ->
     AsciText = "Hello Sam!",
     PadLen = (15-length(AsciText))*8,
     Number = 1239,
     [AsciText, <<0:PadLen, Number:32>>].



regards,
Vlad


From jamesh@REDACTED  Tue Dec 14 20:53:38 2004
From: jamesh@REDACTED (James Hague)
Date: Tue, 14 Dec 2004 13:53:38 -0600
Subject: Atoms in I/O lists
Message-ID: <F7698C6BAED2D41189E00002A50A625F06A07D8E@v-exchange.volition.net>

An I/O list is list where each element is an integer in the range 0-255, a
binary, or another I/O list.  Several times now I've wished that atoms could
be directly included in I/O lists rather than having to call atom_to_list
first.  With atoms legal in I/O lists, you could imagine atom_to_list call
quietly wrapped around each atom.  In addition to being simpler, this could
result in less garbage being generated, assuming that I/O lists get parsed
in the C runtime.

One reason I can think of for not allowing this is if atom names can be
Unicode.  I don't know if Erlang supports this.  Are there other objections?


From raimo@REDACTED  Wed Dec 15 08:54:39 2004
From: raimo@REDACTED (Raimo Niskanen)
Date: 15 Dec 2004 08:54:39 +0100
Subject: Atoms in I/O lists
References: <F7698C6BAED2D41189E00002A50A625F06A07D8E@v-exchange.volition.net>
Message-ID: <lqr7lsro74.fsf@erix.ericsson.se>

I do not have strong and hard objections, rather soft and swampy.

No, atoms cannot be unicode. Erlang's character map is ISO-8859-1.
Strings (lists) might have unicode characters one day, though,
since it would only mean relaxing the 0..255 range reqirement.

1) If we would allow atoms in I/O lists, why not then also floats and all
other basic types that have TYPE_to_list/1 BIFs. Internal rewrites might
be bigger than expected, e.g the flattening to a buffer has got two
passes - one to determine the size and one to write the data so
all TYPE_to_list/1 BIFSs would need internal associates a'la
TYPE_list_size/1. Plus the fact that all TYPE_to_list/1 functions 
create erlang lists, not byte buffers.

2) It is not only sending to a port that uses I/O lists. The I/O
functions in io.erl, io_lib.erl, etc uses a subset :-) of I/O lists,
without binaries, as allowed string for the format character ~s.
Would we have to allow atoms in such strings too?

3) A lot of application code out there probably mangles I/O lists
as they are known today. Such code would have to be rewritten.

So, I feel I did not present a particulary strange case against
allowing atoms in I/O lists. Then it depends on how strong
the cusomer demands are...



jamesh@REDACTED (James Hague) writes:

> An I/O list is list where each element is an integer in the range 0-255, a
> binary, or another I/O list.  Several times now I've wished that atoms could
> be directly included in I/O lists rather than having to call atom_to_list
> first.  With atoms legal in I/O lists, you could imagine atom_to_list call
> quietly wrapped around each atom.  In addition to being simpler, this could
> result in less garbage being generated, assuming that I/O lists get parsed
> in the C runtime.
> 
> One reason I can think of for not allowing this is if atom names can be
> Unicode.  I don't know if Erlang supports this.  Are there other objections?

-- 

/ Raimo Niskanen, Erlang/OTP, Ericsson AB


From klacke@REDACTED  Wed Dec 15 10:39:29 2004
From: klacke@REDACTED (klacke@REDACTED)
Date: Wed, 15 Dec 2004 10:39:29 +0100
Subject: Atoms in I/O lists
In-Reply-To: <lqr7lsro74.fsf@erix.ericsson.se>
References: <F7698C6BAED2D41189E00002A50A625F06A07D8E@v-exchange.volition.net> <lqr7lsro74.fsf@erix.ericsson.se>
Message-ID: <20041215093929.GA19780@hyber.org>

On Wed, Dec 15, 2004 at 08:54:39AM +0100, Raimo Niskanen wrote:

> 
> So, I feel I did not present a particulary strange case against
> allowing atoms in I/O lists. 


I think you did.

Yet another argument would be that: Not allowing atoms
in io lists does catch runtime bugs. Sometimes the occasional
'ok' or 'true' atom happens to find its way into a list later
destined for io.  These bugs are caught by the runtime system.

/klacke

-- 
Claes Wikstrom                        -- Caps lock is nowhere and
http://www.hyber.org                  -- everything is under control          


From bjorn@REDACTED  Wed Dec 15 10:49:03 2004
From: bjorn@REDACTED (Bjorn Gustavsson)
Date: 15 Dec 2004 10:49:03 +0100
Subject: Atoms in I/O lists
In-Reply-To: <20041215093929.GA19780@hyber.org>
References: <F7698C6BAED2D41189E00002A50A625F06A07D8E@v-exchange.volition.net>
	<lqr7lsro74.fsf@erix.ericsson.se> <20041215093929.GA19780@hyber.org>
Message-ID: <o1is73ki28.fsf@erix.ericsson.se>

klacke@REDACTED writes:

> On Wed, Dec 15, 2004 at 08:54:39AM +0100, Raimo Niskanen wrote:
> 
> > 
> > So, I feel I did not present a particulary strange case against
> > allowing atoms in I/O lists. 
> 
> 
> I think you did.
> 
> Yet another argument would be that: Not allowing atoms
> in io lists does catch runtime bugs. Sometimes the occasional
> 'ok' or 'true' atom happens to find its way into a list later
> destined for io.  These bugs are caught by the runtime system.

I agree. I think this is the best argument for not allowing any more
types in I/O lists.

/Bjorn
-- 
Bj?rn Gustavsson, Erlang/OTP, Ericsson AB


From hakan.stenholm@REDACTED  Wed Dec 15 13:51:00 2004
From: hakan.stenholm@REDACTED (=?ISO-8859-1?Q?Administrat=F6r?=)
Date: Wed, 15 Dec 2004 13:51:00 +0100
Subject: Crash in calendar:local_time_to_universal_time_dst/1
Message-ID: <F8ED7C0C-4E97-11D9-BEC7-00039378D5B8@it.su.se>

I'm currently working with the 'calendar' OTP module, to handle dates  
and time values of the iCal (RFC 2445) / CPL standards (RFC 3880).

A call  
"calendar:local_time_to_universal_time_dst({{2004,3,28},{2,0,0}})",  
crashes on my Mac but not on a RedHat Linux machine I tested on.
The return value expected is = [] - assuming that both use DST  
(Daylight Saving Time) according to the current EU standard.
This works on the Linux machine, but not the Mac, other legal dates  
like 2004-03-28 01:59:59 and 2004-03-28 03:00:00 work without problems  
and return the expected results.

The crash:

MacOSX 10.3.6

95> calendar:local_time_to_universal_time_dst({{2004,3,28},{2,0,0}}).

=ERROR REPORT==== 15-Dec-2004::13:22:41 ===
Error in process <0.85.0> with exit value:  
{badarg,[{erlang,universaltime_to_localtime,[{{1969,12,31},{23,59,59}}]} 
,{calendar,local_time_to_universal_time_dst,1},{erl_eval,do_apply,5},{sh 
ell,exprs,6},{shell,eval_loop,3}]}

** exited: {badarg,[{erlang,universaltime_to_localtime,
                             [{{1969,12,31},{23,59,59}}]},
                     {calendar,local_time_to_universal_time_dst,1},
                     {erl_eval,do_apply,5},
                     {shell,exprs,6},
                     {shell,eval_loop,3}]} **


----------------
Red Hat Linux release 7.3

1> calendar:local_time_to_universal_time_dst({{2004,3,28},{2,0,0}}).
[]



From bjorn@REDACTED  Wed Dec 15 15:36:39 2004
From: bjorn@REDACTED (Bjorn Gustavsson)
Date: 15 Dec 2004 15:36:39 +0100
Subject: Erlang/OTP R10B-2 is out
Message-ID: <o13by7d3wo.fsf@erix.ericsson.se>

Bug fix release     : otp_src_R10B-2
Build from snapshot : 2004-12-15

This is a bug fix release 2 for the R10B release.
You can download the full source distribution from
 
  http://www.erlang.org/download/otp_src_R10B-2.tar.gz
  http://www.erlang.org/download/otp_src_R10B-2.readme

Note: To unpack the TAR archive you need a GNU TAR compatible
program. For instance, on MacOS X you need to use the 'gnutar' command;
you can't use the 'tar' command or StuffIt to unpack the sources.


For installation instructions please read the README that is part of
the distribution.

The Windows binary distribution can be downloaded from

  http://www.erlang.org/download/otp_win32_R10B-2.exe

The documentation at http://www.erlang.org will be updated. You can
also download the complete HTML documentation or the Unix manual files

  http://www.erlang.org/download/otp_doc_html_R10B-2.tar.gz
  http://www.erlang.org/download/otp_doc_man_R10B-2.tar.gz

For some OTP applications there are more detailed release notes in the
HTML documentation.

We also want to thank those that sent us patches, suggestions and bug
reports,

The OTP Team
-- 
Bj?rn Gustavsson, Erlang/OTP, Ericsson AB


From rprice@REDACTED  Wed Dec 15 15:59:42 2004
From: rprice@REDACTED (Roger Price)
Date: Wed, 15 Dec 2004 15:59:42 +0100 (CET)
Subject: Guard test pid/1 allows invalid Pid into spawned program
Message-ID: <Pine.LNX.4.44.0412071410440.8015-100000@olive.rogerprice.org>

I wrote function test/0 to spawn another process and run function
hello/1 in that process.  hello/1 does nothing but print a debugging
message which displays the current processes.

I tested this by entering test:test() on the command line - all went well
and the expected debugging messages were written.  I also ran the test
using apply(test,test,[]) and all went well again.

Finally I ran spawn(test,test,[]) and argument passed to function hello/1
disappeared.

The Pid which I pass to hello/1 as an argument, is invalid even though I
have a guard test pid/1, and the Pid satisfies it.
lists:member(Pid,processes()) gives true before and after the spawn but
false within the spawned program.

If I print hello/1's debug message twice, I see that the list of current
processes changes, with the argument absent from the second message.

I use R9C-1 on a 1GHz PC under SuSE Linux 8.1.  I attach, with my
apologies for a long post, the functions test/0 and hello/1.

Here is a slightly edited trace of the test:

rprice@REDACTED:/tmp> erlc -W test.erl
rprice@REDACTED:/tmp> erlc -W hello.erl
rprice@REDACTED:/tmp> erl
Erlang (BEAM) emulator version 5.3.6.2 [source] [hipe]

Eshell V5.3.6.2  (abort with ^G)
1> % apply/3 produces the correct output
1> apply(test,test,[]) .
test/0 opened LogStream.  Processes:
[...,<0.23.0>,<0.24.0>,<0.25.0>,<0.26.0>,<0.31.0>]
checkPid 'LogStream before spawn' succeeds for pid <0.31.0>.
spawn(hello,hello,[<0.31.0>]) -> <0.32.0>.
checkPid 'LogStream after spawn' succeeds for pid <0.31.0>.
ok
hello/1 starts with processes
[...,<0.23.0>,<0.24.0>,<0.25.0>,<0.26.0>,<0.31.0>,<0.32.0>]
hello/1 continues with processes
[...,<0.23.0>,<0.24.0>,<0.25.0>,<0.26.0>,<0.31.0>,<0.32.0>,<0.33.0>]
checkPid 'Spawned LogStream' succeeds for pid <0.31.0>.

2> % spawn/3 fails
2> spawn(test,test,[]) .
<0.34.0>
test/0 opened LogStream.  Processes:
[...,<0.23.0>,<0.24.0>,<0.25.0>,<0.26.0>,<0.31.0>,<0.34.0>, <0.35.0>,<0.36.0>]
checkPid 'LogStream before spawn' succeeds for pid <0.35.0>.
spawn(hello,hello,[<0.35.0>]) -> <0.37.0>.
checkPid 'LogStream after spawn' succeeds for pid <0.35.0>.
hello/1 starts with processes
[...,<0.23.0>,<0.24.0>,<0.25.0>,<0.26.0>,<0.31.0>,<0.34.0>,<0.35.0>,<0.36.0>,
 <0.37.0>]
hello/1 continues with processes
[...,<0.23.0>,<0.24.0>,<0.25.0>,<0.26.0>,<0.31.0>,<0.36.0>,<0.37.0>]
checkPid 'Spawned LogStream' fails for pid <0.35.0>.

Note in the "spawn/3" test how <0.35.0> vanishes from the list of
processes, while hello/1 is running.

Roger

%% test.erl

-module(test).
-export([checkPid/2,pp/1,test/0]).

test() ->
   LogFile = '/tmp/hello.log',
   {ok,LogStream} = file:open(LogFile,[append]),

   pp("test/0 opened LogStream.  Processes:"),
   checkPid('LogStream before spawn',LogStream),

   ProgPid = spawn(hello, hello, [LogStream]),

   io:format("spawn(hello,hello,~p) -> ~p.~n", [[LogStream],ProgPid]),
   checkPid('LogStream after spawn',LogStream),
   ok .

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%% Check that a given pid is for a known process.
checkPid(What,Pid) when atom(What), pid(Pid) ->
  case lists:member(Pid,processes()) of
     true -> io:format("checkPid ~p succeeds for pid ~p.~n",
                       [What,Pid]),
             true;
     _    -> io:format("checkPid ~p fails for pid ~p.~n",
                       [What,Pid]),
             exit(What)
  end;
checkPid(Bad1,Bad2) ->
   io:format("Error in pint:checkPid(~p,~p)~n", [Bad1,Bad2]),
   exit('checkPid') .

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%% Not very pretty printer for processes.
pp(Prefix) -> io:format("~s [~s]~n", [Prefix,ppPr(processes())]) .
ppPr([])     -> "";
ppPr([P])    -> io_lib:format("~p", [P]);
ppPr([P|Ps]) -> io_lib:format("~p,~s", [P,ppPr(Ps)]) .

%% End of module test

%% hello.erl

-module(hello).
-export([hello/1]).

hello(LogStream) when pid(LogStream) ->
   test:pp("hello/1 starts with processes"),
   test:pp("hello/1 continues with processes"),
   test:checkPid('Spawned LogStream',LogStream),
   yessir .

%% End of module hello




From jamesh@REDACTED  Wed Dec 15 16:10:11 2004
From: jamesh@REDACTED (James Hague)
Date: Wed, 15 Dec 2004 09:10:11 -0600
Subject: Atoms in I/O lists
Message-ID: <F7698C6BAED2D41189E00002A50A625F06A07D9C@v-exchange.volition.net>

>1) If we would allow atoms in I/O lists,
>why not then also floats and all other basic
>types that have TYPE_to_list/1 BIFs.

Because there are format issues with floats, such as endianness and single
vs. double precision.  With bytes there are no such issues.

>3) A lot of application code out there
>probably mangles I/O lists as they are known
>today. Such code would have to be rewritten.

True, sadly.

I don't feel strongly about this issue.  I've simply noticed that I tend to
use atoms for a lot of things internally, because they're so simple, and end
up converting them to lists for display or writing results to a file.
Technically there's no reason to convert them to lists, because the
characters making up the atom name are already stored somewhere.


From vlad_dumitrescu@REDACTED  Wed Dec 15 16:45:35 2004
From: vlad_dumitrescu@REDACTED (Vlad Dumitrescu)
Date: Wed, 15 Dec 2004 16:45:35 +0100
Subject: Guard test pid/1 allows invalid Pid into spawned program
References: <Pine.LNX.4.44.0412071410440.8015-100000@olive.rogerprice.org>
Message-ID: <BAY15-DAV124123ABDF4917BD62BE729CAD0@phx.gbl>

Hi,

I think there is no problem. LogStream is a process linked to the process that
runs test:test() - when test terminates, then LogStream gets killed too. This is
a result of having test:test() spawned: it no longer is there when hello runs.
In the successful cases, the process running test:test() is the shell, and it
remains alive.

regards,
Vlad


From sl45sms@REDACTED  Wed Dec 15 17:04:25 2004
From: sl45sms@REDACTED (Panagiotis Skarvelis)
Date: Wed, 15 Dec 2004 08:04:25 -0800 (PST)
Subject: highlight for KDE (kate kwrite kdevelop etc...) 
Message-ID: <20041215160426.44387.qmail@web52907.mail.yahoo.com>

I just finish an xml file for Kate the editor of KDE.
You can download this file from
http://www.linuxer.gr/download/highlightErlang.1.0.tar.gz

I hope someone found it useful :)

Panagiotis Skarvelis


__________________________________________________
Do You Yahoo!?
Tired of spam?  Yahoo! Mail has the best spam protection around 
http://mail.yahoo.com 


From mickael.remond@REDACTED  Wed Dec 15 22:13:37 2004
From: mickael.remond@REDACTED (Mickael Remond)
Date: Wed, 15 Dec 2004 22:13:37 +0100
Subject: highlight for KDE (kate kwrite kdevelop etc...)
In-Reply-To: <20041215160426.44387.qmail@web52907.mail.yahoo.com>
References: <20041215160426.44387.qmail@web52907.mail.yahoo.com>
Message-ID: <41C0A901.3020602@erlang-fr.org>

Panagiotis Skarvelis wrote:
> I just finish an xml file for Kate the editor of KDE.
> You can download this file from
> http://www.linuxer.gr/download/highlightErlang.1.0.tar.gz
> 
> I hope someone found it useful :)

Yes :-)

Thank you !

-- 
Micka?l R?mond
  http://www.erlang-projects.org/


From klacke@REDACTED  Wed Dec 15 23:39:09 2004
From: klacke@REDACTED (klacke@REDACTED)
Date: Wed, 15 Dec 2004 23:39:09 +0100
Subject: yaws-1.51
Message-ID: <20041215223909.GA24685@hyber.org>

Activity level is high. New release again. 

Added support for config changes without stopping 
the running system, virt servers can be added, removed 
and changed without affecting traffic.       

As usual, source, docs etc at http://yaws.hyber.org

/klacke


-- 
Claes Wikstrom                        -- Caps lock is nowhere and
http://www.hyber.org                  -- everything is under control




From rprice@REDACTED  Wed Dec 15 22:49:30 2004
From: rprice@REDACTED (Roger Price)
Date: Wed, 15 Dec 2004 22:49:30 +0100 (CET)
Subject: Guard test pid/1 allows invalid Pid into spawned program
In-Reply-To: <BAY15-DAV124123ABDF4917BD62BE729CAD0@phx.gbl>
Message-ID: <Pine.LNX.4.44.0412152232070.1371-100000@olive.rogerprice.org>

On Wed, 15 Dec 2004, Vlad Dumitrescu wrote:

> I think there is no problem. LogStream is a process linked to the
> process that runs test:test() - when test terminates, then LogStream
> gets killed too. This is a result of having test:test() spawned: it no
> longer is there when hello runs. In the successful cases, the process
> running test:test() is the shell, and it remains alive.

Hello Vlad,

Thanks for the insight.  I was under the mistaken impression that somehow
processes, like memory, were garbage collected only when the "last
pointer" disappeared.

Does the link-based garbage collection work between nodes?  I.e, if I pass
a Pid to my hello/1 in another node must I keep test/0 alive in the
originating node to keep the "LogStream" Pid alive on some distant node?

Best Regards,
Roger



From jabba@REDACTED  Thu Dec 16 04:34:12 2004
From: jabba@REDACTED (Jani Launonen)
Date: Thu, 16 Dec 2004 05:34:12 +0200 (EET)
Subject: highlight for KDE (kate kwrite kdevelop etc...) 
In-Reply-To: <20041215160426.44387.qmail@web52907.mail.yahoo.com>
References: <20041215160426.44387.qmail@web52907.mail.yahoo.com>
Message-ID: <Pine.GSO.4.61.0412160529380.10981@stekt1>

On Wed, 15 Dec 2004, Panagiotis Skarvelis wrote:

> I just finish an xml file for Kate the editor of KDE.
> You can download this file from
> http://www.linuxer.gr/download/highlightErlang.1.0.tar.gz
>
> I hope someone found it useful :)

I did, thanks! It makes it easier to read examples and makes even my own 
code to look nicer (which is just an illusion ;)

> Panagiotis Skarvelis
>
>
> __________________________________________________
> Do You Yahoo!?
> Tired of spam?  Yahoo! Mail has the best spam protection around
> http://mail.yahoo.com

-+-+-+-
Jani Launonen


From casper2000a@REDACTED  Thu Dec 16 04:54:52 2004
From: casper2000a@REDACTED (Casper)
Date: Thu, 16 Dec 2004 09:54:52 +0600
Subject: Erl Ports and Threads
Message-ID: <!~!UENERkVCMDkAAQACAAAAAAAAAAAAAAAAABgAAAAAAAAA2eo0SPN1R0OAqojNqhlfEsKAAAAQAAAA+A1BJeev7EqkGuAbgLADoAEAAAAA@omnibis.com>

Hi all,

 

I am wondering, is Erlang Ports are thread safe? 

 

I'm planning to use Erl Interface based Erlang Port. If communicates with
the Erlang node over Standard Input and Output. The communication message
has 2 bytes header indicating the length of the message followed by the
encoded message. If the C program has multiple threads and if they write to
the stdout at the same time, will there be any communication problem?

 

Thanks in advance!

Eranga

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://erlang.org/pipermail/erlang-questions/attachments/20041216/441ed9d5/attachment.htm>

From vlad_dumitrescu@REDACTED  Thu Dec 16 09:20:15 2004
From: vlad_dumitrescu@REDACTED (Vlad Dumitrescu)
Date: Thu, 16 Dec 2004 09:20:15 +0100
Subject: Guard test pid/1 allows invalid Pid into spawned program
References: <Pine.LNX.4.44.0412152232070.1371-100000@olive.rogerprice.org>
Message-ID: <BAY15-DAV160E48F381D7CAAB47541E9CAE0@phx.gbl>

Hi,

Yes, links work even for remote processes. And no, processes are not garbage
collected, they live until the function they were spawned with ends.

Usually, if one uses spawn (as opposed to spawn_link) there is no link created
between processes, and thus the spawned process doesn't depend on the parent
remaining alive. The link can be severed, but it's bad programming practice, as
the place (process) that opened the file should also close it (otherwise the
file process could keep running for all eternity)

I'd suggest another approach: the process that opens LogStream should be a
server that can hand out the pid to clients and also receives 'close' commands
from them.

Better, the owner process can be made a logging server and send it's own pid ti
the children: the children just say 'write this data to the log', without caring
whether if the log medium is a file or a database or something else.

I hope this helps. The documentation and the book contain more detailed
information, but please ask again if it's not clear enough.

regards,
Vlad

----- Original Message ----- 
From: "Roger Price" <rprice@REDACTED>
To: "Erlang mailing list" <erlang-questions@REDACTED>
Sent: Wednesday, December 15, 2004 10:49 PM
Subject: Re: Guard test pid/1 allows invalid Pid into spawned program


> Thanks for the insight.  I was under the mistaken impression that somehow
> processes, like memory, were garbage collected only when the "last
> pointer" disappeared.
>
> Does the link-based garbage collection work between nodes?  I.e, if I pass
> a Pid to my hello/1 in another node must I keep test/0 alive in the
> originating node to keep the "LogStream" Pid alive on some distant node?
>
> Best Regards,
> Roger
>
>


From klacke@REDACTED  Thu Dec 16 10:03:53 2004
From: klacke@REDACTED (klacke@REDACTED)
Date: Thu, 16 Dec 2004 10:03:53 +0100
Subject: Erl Ports and Threads
In-Reply-To: <!~!UENERkVCMDkAAQACAAAAAAAAAAAAAAAAABgAAAAAAAAA2eo0SPN1R0OAqojNqhlfEsKAAAAQAAAA+A1BJeev7EqkGuAbgLADoAEAAAAA@omnibis.com>
References: <!~!UENERkVCMDkAAQACAAAAAAAAAAAAAAAAABgAAAAAAAAA2eo0SPN1R0OAqojNqhlfEsKAAAAQAAAA+A1BJeev7EqkGuAbgLADoAEAAAAA@omnibis.com>
Message-ID: <20041216090353.GA2204@hyber.org>

On Thu, Dec 16, 2004 at 09:54:52AM +0600, Casper wrote:
>    Hi all,
> 
> 
>    I am wondering, is Erlang Ports are thread safe?
> 
> 
>    I'm planning to use Erl Interface based Erlang Port. If communicates
>    with the Erlang node over Standard Input and Output. The communication
>    message has 2 bytes header indicating the length of the message
>    followed by the encoded message. If the C program has multiple threads
>    and if they write to the stdout at the same time, will there be any
>    communication problem?
> 

This will ofcourse wreak major havoc. You have to synchronize
the output from your threads.

It's just a pipe between erlang and the external C-program and
erlang will just sit there reading data.

write() operations on a socket/pipe from the external program
are not atomic in unix/linux. 


/klacke


-- 
Claes Wikstrom                        -- Caps lock is nowhere and
http://www.hyber.org                  -- everything is under control          


From casper2000a@REDACTED  Thu Dec 16 10:37:34 2004
From: casper2000a@REDACTED (Casper)
Date: Thu, 16 Dec 2004 15:37:34 +0600
Subject: Erl Ports and Threads
In-Reply-To: <20041216090353.GA2204@hyber.org>
Message-ID: <!~!UENERkVCMDkAAQACAAAAAAAAAAAAAAAAABgAAAAAAAAA2eo0SPN1R0OAqojNqhlfEsKAAAAQAAAAND2jD1gc9UWP5CYTb7mfbgEAAAAA@omnibis.com>

Hi Klacke,

Thanks for the response.

Here I am not talking about simultaneous write() operations on a socket/pipe
from the multiple external program. I am talking about multiple threads in
one program.

Is single write() call, atomic or not?

If that's not thread safe, what's the best way to go for inter-operability?

Cheers,
Eranga




-----Original Message-----
From: klacke@REDACTED [mailto:klacke@REDACTED] 
Sent: Thursday, December 16, 2004 3:04 PM
To: Casper
Cc: erlang-questions@REDACTED
Subject: Re: Erl Ports and Threads

On Thu, Dec 16, 2004 at 09:54:52AM +0600, Casper wrote:
>    Hi all,
> 
> 
>    I am wondering, is Erlang Ports are thread safe?
> 
> 
>    I'm planning to use Erl Interface based Erlang Port. If communicates
>    with the Erlang node over Standard Input and Output. The communication
>    message has 2 bytes header indicating the length of the message
>    followed by the encoded message. If the C program has multiple threads
>    and if they write to the stdout at the same time, will there be any
>    communication problem?
> 

This will ofcourse wreak major havoc. You have to synchronize
the output from your threads.

It's just a pipe between erlang and the external C-program and
erlang will just sit there reading data.

write() operations on a socket/pipe from the external program
are not atomic in unix/linux. 


/klacke


-- 
Claes Wikstrom                        -- Caps lock is nowhere and
http://www.hyber.org                  -- everything is under control




From klacke@REDACTED  Thu Dec 16 10:38:38 2004
From: klacke@REDACTED (klacke@REDACTED)
Date: Thu, 16 Dec 2004 10:38:38 +0100
Subject: Erl Ports and Threads
In-Reply-To: <!~!UENERkVCMDkAAQACAAAAAAAAAAAAAAAAABgAAAAAAAAA2eo0SPN1R0OAqojNqhlfEsKAAAAQAAAAND2jD1gc9UWP5CYTb7mfbgEAAAAA@omnibis.com>
References: <20041216090353.GA2204@hyber.org> <!~!UENERkVCMDkAAQACAAAAAAAAAAAAAAAAABgAAAAAAAAA2eo0SPN1R0OAqojNqhlfEsKAAAAQAAAAND2jD1gc9UWP5CYTb7mfbgEAAAAA@omnibis.com>
Message-ID: <20041216093838.GB2204@hyber.org>

On Thu, Dec 16, 2004 at 03:37:34PM +0600, Casper wrote:
> Hi Klacke,
> 
> Thanks for the response.
> 
> Here I am not talking about simultaneous write() operations on a socket/pipe
> from the multiple external program. I am talking about multiple threads in
> one program.

I know, that's what you wrote.

> 
> Is single write() call, atomic or not?
> 

Not.


> If that's not thread safe, what's the best way to go for inter-operability?

Let the write routine take a lock on a mutex before
it writes and release the lock when the write (which must
be performed in a loop until everything is written) is
finished.

Read man pages for read/write and pthread_mutex_init, lock
and friends.


/klacke


-- 
Claes Wikstrom                        -- Caps lock is nowhere and
http://www.hyber.org                  -- everything is under control          


From ulf.wiger@REDACTED  Thu Dec 16 10:50:47 2004
From: ulf.wiger@REDACTED (Ulf Wiger (AL/EAB))
Date: Thu, 16 Dec 2004 10:50:47 +0100
Subject: record deep dive
Message-ID: <37FB7AA6F5F9814FB634A7BF4C35A6F54029F0@ESEALNT442.al.sw.ericsson.se>


I took R10B-2 for a test drive using my ring benchmark
on my new Blade1500 Sun workstation (1 GHz, 1 GB RAM)

I was able to push the ring benchmark to 630,000 processes
before I got into lots of iowait. Beam used about 800 MB
of RAM.

Timings in microseconds:

N procs  spawn   send
630000   6.47106 1.96776

I also tried a ring benchmark with increasing payload,
using the hybrid heap, and it performed excellently,
with message passing cost of abot 1 us regardless of 
message size. Apparently, the compiler is capable of
realizing that the output of lists:duplicate/2 should
be put on the shared heap, since it is later passed 
as an argument in spawn_link()? Impressive.

/Uffe


From tobias.lindahl@REDACTED  Thu Dec 16 10:53:28 2004
From: tobias.lindahl@REDACTED (Tobias Lindahl)
Date: Thu, 16 Dec 2004 10:53:28 +0100 (MET)
Subject: Dialyzer v1.2.0
Message-ID: <Pine.SOL.4.53.0412161032520.6636@hamberg.it.uu.se>

Once again we are proud to announce a new release of Dialyzer.

Dialyzer v1.2.0 is available for download at:

http://www.it.uu.se/research/group/hipe/dialyzer/

Version 1.2.0 requires Erlang/OTP R10B-2 to work properly. The main new
feature is that Dialyzer now works on all platforms (no HiPE support is
required).

At the end of this message are additional release notes. We are greatful
for the feedback we have received, including new contributions and
suggestions for improvement.

Have fun!

Tobias Lindahl and Kostis Sagonas



Version 1.2.0
-------------
 - Dialyzer works with the open source and commercial versions of
Erlang/OTP
   R10B-2 on all platforms (i.e., HiPE support is not a prerequisite
anymore).
 - Whenever a .beam file contains abstract code (i.e., has been compiled
   with the +debug_info option), the analysis starts from this code rather
   than from BEAM bytecode -- this makes the results identical to those
   obtained by analysis starting from source code.
   (This is a contribution from Bjorn Gustavsson - Thanks!)
 - Added -pa command line option for easier handling of -include_lib()
   directives in source code.
 - Includes all changes added in v1.1.1; see below.  The "Add Recursively"
   functionality is also available in the command-line mode (-r option).

Version 1.1.1 (internal, not released to the masses :-)
-------------
 - The command-line mode of Dialyzer spits discrepancies in stdout
   by default and returns a numeric value indicating its exit status.
 - Fixed problems using the shell script on Solaris machines.
 - Fixed small inconsistencies in Dialyzer's documentation and help.
 - Added "Add Recursively" button contributed by Martin Bjorklund
(thanks!).



From kostis@REDACTED  Thu Dec 16 11:05:07 2004
From: kostis@REDACTED (Kostis Sagonas)
Date: Thu, 16 Dec 2004 11:05:07 +0100 (MET)
Subject: record deep dive
In-Reply-To: Mail from '"Ulf Wiger (AL/EAB)" <ulf.wiger@ericsson.com>'
      dated: Thu, 16 Dec 2004 10:50:47 +0100
Message-ID: <200412161005.iBGA57XO000342@spikklubban.it.uu.se>

Ulf Wiger wrote:
 >
 > I also tried a ring benchmark with increasing payload,
 > using the hybrid heap, and it performed excellently,
 > with message passing cost of abot 1 us regardless of 
 > message size.

Yes.  The hybrid heap architecture has a lot of potential,
and we have spent some effort in making it more robust than
in previous versions, but let me stress that the hybrid heap
is still experimental and requires more tuning.  I strongly
recommend playing with it, but it is still not mature to be
used in production code.

 > Apparently, the compiler is capable of
 > realizing that the output of lists:duplicate/2 should
 > be put on the shared heap, since it is later passed 
 > as an argument in spawn_link()? Impressive.

Thanks for the "Impressive", but unfortunately this is not true.
We still have not hooked the changes to the compiler for the
hybrid heap.  This ability to recognize shared data is currently
only part of the hybrid runtime system.  As such it does indeed
help but it has a runtime cost.  Wait till you see the changes
in the compiler being hooked in... and then you'll really be
impressed !

Cheers,
Kostis.



From hedeland@REDACTED  Thu Dec 16 11:10:30 2004
From: hedeland@REDACTED (Per Hedeland)
Date: Thu, 16 Dec 2004 11:10:30 +0100 (CET)
Subject: Erl Ports and Threads
In-Reply-To: <20041216093838.GB2204@hyber.org>
Message-ID: <200412161010.iBGAAUPl080157@tordmule.bluetail.com>

klacke@REDACTED wrote:
>
>On Thu, Dec 16, 2004 at 03:37:34PM +0600, Casper wrote:
>> Hi Klacke,
>> 
>> Thanks for the response.
>> 
>> Here I am not talking about simultaneous write() operations on a socket/pipe
>> from the multiple external program. I am talking about multiple threads in
>> one program.
>
>I know, that's what you wrote.
>
>> 
>> Is single write() call, atomic or not?
>> 
>
>Not.

<nitpick>
Actually this isn't entirely true - when different processes write to
the same pipe, their writes are guaranteed (on a standards-compliant
Unix:-) to be atomic if they are "small enough" - from a decent
pathconf(2) man page:

     _PC_PIPE_BUF
             The maximum number of bytes which will be written atomically to a
             pipe.

(The number is typically 4096 or so.) Whether this is still true for
different threads in a process writing to the same pipe I don't know,
but I would expect so (might be implementation-dependant though). And of
course relying on the undocumented fact that the port communication
channel *is* a pipe, and not e.g. a local socket which doesn't have such
guarantees, isn't a very good idea.
</nitpick>

For all practical purposes, I agree with Klacke's answer.:-)

--Per


From dhanasekaran.c@REDACTED  Thu Dec 16 11:58:26 2004
From: dhanasekaran.c@REDACTED (Chenniappan,Dhanasekaran [DBA])
Date: Thu, 16 Dec 2004 16:28:26 +0530
Subject: Erl Ports and Threads
Message-ID: <124AD2BD2C1FB245B2E4FF0E90AE0E324A5790@ntserver.dbachn>

Hi,
> If that's not thread safe, what's the best way to go for
inter-operability?

 I think, You may use the erl_interface for interoperability instead of
port.

With Regards
Dhanas.

-----Original Message-----
From: klacke@REDACTED [mailto:klacke@REDACTED] 
Sent: Thursday, December 16, 2004 3:09 PM
To: Casper
Cc: klacke@REDACTED; erlang-questions@REDACTED
Subject: Re: Erl Ports and Threads

On Thu, Dec 16, 2004 at 03:37:34PM +0600, Casper wrote:
> Hi Klacke,
> 
> Thanks for the response.
> 
> Here I am not talking about simultaneous write() operations on a
socket/pipe
> from the multiple external program. I am talking about multiple threads in
> one program.

I know, that's what you wrote.

> 
> Is single write() call, atomic or not?
> 

Not.


> If that's not thread safe, what's the best way to go for
inter-operability?

Let the write routine take a lock on a mutex before
it writes and release the lock when the write (which must
be performed in a loop until everything is written) is
finished.

Read man pages for read/write and pthread_mutex_init, lock
and friends.


/klacke


-- 
Claes Wikstrom                        -- Caps lock is nowhere and
http://www.hyber.org                  -- everything is under control



From mickael.remond@REDACTED  Thu Dec 16 17:00:05 2004
From: mickael.remond@REDACTED (Mickael Remond)
Date: Thu, 16 Dec 2004 17:00:05 +0100
Subject: Promotion in Objectweb
Message-ID: <41C1B105.3050601@erlang-fr.org>

Hello,

Objectweb is a well-known consortium focusing on Open Source middleware. 
It has been mainly exclusively Java until now. The consoritum is mainly 
targeting IT managers and architects.

The Erlang-projects association is joining the consortium to push and 
promote Erlang-based middleware solution from the inside. We think that 
software like J-EAI for example can help the consortium solve some of 
there needs.

I hope that this will help a little bit more increasing the promotion 
and the interest in Erlang.

For more details, see:

http://www.erlang-projects.org/Public/news/erlang-projects_join/view

Cheers,

-- 
Micka?l R?mond


From vlad_dumitrescu@REDACTED  Thu Dec 16 20:00:53 2004
From: vlad_dumitrescu@REDACTED (Vlad Dumitrescu)
Date: Thu, 16 Dec 2004 20:00:53 +0100
Subject: Dialyzer v1.2.0
References: <Pine.SOL.4.53.0412161032520.6636@hamberg.it.uu.se>
Message-ID: <BAY15-DAV3B58CCDF090EC98B6D41F9CAE0@phx.gbl>

> Version 1.2.0 requires Erlang/OTP R10B-2 to work properly. The main new
> feature is that Dialyzer now works on all platforms (no HiPE support is
> required).

Super cool!

However, in practice it still won't compile and start on Windows... The 
reason is that all paths in the scripts are Unix-like. Cygwin understands 
them properly, but Erlang isn't compiled as a Cygwin application and expects 
regular Windows paths.

The first problem is with the link to hipe_icode_type.hrl: erl chokes on it. 
I copied the file instead, and commented out the lines that create the link, 
because I don't know how to handle this.

The nex problem is when launching dialyzer. I enclosed the definitions 
DIALYZER_OTP and DIALYZER_DIR inside a
    `cygpath -m #`
where # is the value created at setup. I don't have a patch, as it's trivial 
and you might want to handle the OS dependencies in your own way.

After that, I got it up and running!
regards,
Vlad




From sl45sms@REDACTED  Thu Dec 16 21:07:46 2004
From: sl45sms@REDACTED (Panagiotis Skarvelis)
Date: Thu, 16 Dec 2004 12:07:46 -0800 (PST)
Subject: highlight for KDE (kate kwrite kdevelop etc...) 
Message-ID: <20041216200746.63430.qmail@web52903.mail.yahoo.com>

Thank's to both :-)

Panagiotis Skarvelis



		
__________________________________ 
Do you Yahoo!? 
Jazz up your holiday email with celebrity designs. Learn more. 
http://celebrity.mail.yahoo.com


From ulf.wiger@REDACTED  Thu Dec 16 21:51:57 2004
From: ulf.wiger@REDACTED (Ulf Wiger)
Date: Thu, 16 Dec 2004 21:51:57 +0100
Subject: record deep dive
In-Reply-To: <200412161005.iBGA57XO000342@spikklubban.it.uu.se>
References: <200412161005.iBGA57XO000342@spikklubban.it.uu.se>
Message-ID: <opsi4b8vstmk9l4x@mail1.telia.com>

On Thu, 16 Dec 2004 11:05:07 +0100 (MET), Kostis Sagonas  
<kostis@REDACTED> wrote:

>  > Apparently, the compiler is capable of
>  > realizing that the output of lists:duplicate/2 should
>  > be put on the shared heap, since it is later passed
>  > as an argument in spawn_link()? Impressive.
>
> Thanks for the "Impressive", but unfortunately this is not true.
> We still have not hooked the changes to the compiler for the
> hybrid heap.  This ability to recognize shared data is currently
> only part of the hybrid runtime system.  As such it does indeed
> help but it has a runtime cost.  Wait till you see the changes
> in the compiler being hooked in... and then you'll really be
> impressed !

Ok, then, John Hughes and I don't have to feel too stupid
for staring at the ring.S listing and not being able to
spot the instructions that directed the data to the shared
heap in the first place! I feel much better now.  (:

Uffe
-- 
Ulf Wiger



From jabba@REDACTED  Fri Dec 17 06:45:05 2004
From: jabba@REDACTED (Jani Launonen)
Date: Fri, 17 Dec 2004 07:45:05 +0200 (EET)
Subject: Good taste design question about nested FSMs
Message-ID: <Pine.GSO.4.61.0412170715490.29575@stekt1>

Hello everybody,

I'd be very grateful if erlangers of good taste would give me an advice 
about how to make a nested FSMs in an OTP compatible manner. I'm in a 
process of trying to improve my skills concerning OTP and my judgement for 
good erlang code.

Little background: I met couple of friends that I hadn't seen long time and 
was surprised that they still used to play MUD (www.icesus.org). Not that 
I'm that interested in playing but the game itself ;) Or that at least for 
now...

As these fellows described the workings of MUD a thought started to creep 
along my spine and I realised how well such a multiplayer game would suit 
erlang. Of course I was naive to think that I've done a revelation --- just 
yesterday I noticed that Micka?l R?mond and Thierry Mallard had alredy 
shown a whole lot more sophisticated ideas about such multiuser games in 
http://www.erlang.se/euc/03/proceedings/0950Mikael.pdf. Haven't read it 
fully yet.

But the MUD still is a very interesting proving ground for me to improve my 
skills. So far I've managed to make a "portmapper" along the Joe's 
tcp_server, but implemented in gen_server -behaviour and included in 
supervision tree. The portmapper asks player supervisor to start an 
"acceptor" child that listens to a tcp port the mapper has opened for 
listening and keeps track of number of active childs. Pretty straight 
forward thing.

I implemented initially the child as a gen_fsm, but I don't think its that 
elegant solution. Firstly one have to ask supervisor to create it and after 
that send a message to start listening to the port (gen_tcp:accept) so that 
the fsm-process doesn't hang on the init function and make supervisor 
nervous. Other than that the fsm should communicate with the tcp port (with 
handle_info) and with other player/room processes that is reacted in 
"normal" state functions. More over the player could be in different states 
(fighting, in a group and such) that should be handled in nested FSMs.

So --- finally --- the question is; apparently I should ditch the gen_fsm 
and make my custom fsm that can be attached to the supervisor tree (by 
using sys and proc_lib), so that I can make consistent receive loop, but 
the inherent "nestedness" is still a mess. Should I make a process for each 
substate machine (which would make communicating with other players 
difficult as the "active" process is changing) or should I make nested 
receive loops that handle the substate machines?

If my explanation of problem was a mess by itself, as for clarification.

Cheers,

-+-+-+-
Jani Launonen 
Student. . . . . . . . . .University of Oulu, Dept. of El. & Inf. Eng.

"Life is what happens to you while you're busy making future plans."
  - Alfred E. Neuman

From ulf.wiger@REDACTED  Fri Dec 17 08:18:13 2004
From: ulf.wiger@REDACTED (Ulf Wiger)
Date: Fri, 17 Dec 2004 08:18:13 +0100
Subject: Good taste design question about nested FSMs
In-Reply-To: <Pine.GSO.4.61.0412170715490.29575@stekt1>
References: <Pine.GSO.4.61.0412170715490.29575@stekt1>
Message-ID: <opsi448nbvmk9l4x@mail1.telia.com>

On Fri, 17 Dec 2004 07:45:05 +0200 (EET), Jani Launonen <jabba@REDACTED>  
wrote:

> So --- finally --- the question is; apparently I should ditch the
> gen_fsm and make my custom fsm that can be attached to the supervisor
> tree (by using sys and proc_lib), so that I can make consistent
> receive loop, but the inherent "nestedness" is still a mess. Should
> I make a process for each substate machine (which would make
> communicating with other players difficult as the "active" process
> is changing) or should I make nested receive loops that handle the
> substate machines?

This sort of thing is a bit difficult to do well in gen_fsm, as you
point out. One one hand, it provides built-in support for the OTP
system messages, which is good, but on the other hand, it provides
no good support for nested state machines.

Whether you should break up your state machine in multiple processes
is difficult to answer in general. Sometimes, it's a good idea, but
other times it makes things more complex. You have to go through the
different models in your mind (or on paper) and try to get a feeling
for it.

Reasonably simple sub states are nicely done with nested receives.
The gen_fsm model has a problem if you may want to handle system
messages (suspend, resume, change_code, get_status) in sub states -
gen_fsm only does that on the top level. If this is a real problem,
you can either use multiple gen_fsms (which may buy you other
problems, like synchronization bugs), or you might try out plain_fsm,
on jungerl. It tries to help you out with system messages without
forcing you to leave the textbook Erlang fsm programming model.

Discussions about plain_fsm on erlang-questions include:

http://www.erlang.org/ml-archive/erlang-questions/200402/msg00043.html

http://www.erlang.org/ml-archive/erlang-questions/200407/msg00032.html

/Uffe
-- 
Ulf Wiger



From klacke@REDACTED  Fri Dec 17 10:37:41 2004
From: klacke@REDACTED (klacke@REDACTED)
Date: Fri, 17 Dec 2004 10:37:41 +0100
Subject: Good taste design question about nested FSMs
In-Reply-To: <opsi448nbvmk9l4x@mail1.telia.com>
References: <Pine.GSO.4.61.0412170715490.29575@stekt1> <opsi448nbvmk9l4x@mail1.telia.com>
Message-ID: <20041217093741.GA17950@hyber.org>

On Fri, Dec 17, 2004 at 08:18:13AM +0100, Ulf Wiger wrote:

> 
> Discussions about plain_fsm on erlang-questions include:
> 
> http://www.erlang.org/ml-archive/erlang-questions/200402/msg00043.html
> 
> http://www.erlang.org/ml-archive/erlang-questions/200407/msg00032.html


Nice stuff, I never saw these posts on the list. Best of both
worlds !!

/klacke



-- 
Claes Wikstrom                        -- Caps lock is nowhere and
http://www.hyber.org                  -- everything is under control          


From dietmar@REDACTED  Fri Dec 17 13:11:26 2004
From: dietmar@REDACTED (Dietmar Admin Schaefer)
Date: Fri, 17 Dec 2004 13:11:26 +0100
Subject: snmp-mnesia
Message-ID: <41C2CCEE.7050906@ast.dfs.de>

HI !


Is there anybody out there who works with snmp/mnesia ?

Is the question that I asked last week too complicated to be answered or 
too stupid to ask ?


My question was about how to represent MIB-Tables to mnesia tables.


regard


Dietmar


From jabba@REDACTED  Fri Dec 17 20:10:57 2004
From: jabba@REDACTED (Jani Launonen)
Date: Fri, 17 Dec 2004 21:10:57 +0200 (EET)
Subject: Good taste design question about nested FSMs
In-Reply-To: <opsi448nbvmk9l4x@mail1.telia.com>
References: <Pine.GSO.4.61.0412170715490.29575@stekt1> <opsi448nbvmk9l4x@mail1.telia.com>
Message-ID: <Pine.GSO.4.61.0412172101450.6817@stekt1>

> On Fri, 17 Dec 2004 07:45:05 +0200 (EET), Jani Launonen <jabba@REDACTED> 
> wrote:
>
>> So --- finally --- the question is; apparently I should ditch the
>> gen_fsm and make my custom fsm that can be attached to the supervisor
>> tree (by using sys and proc_lib), so that I can make consistent
>> receive loop, but the inherent "nestedness" is still a mess. Should
>> I make a process for each substate machine (which would make
>> communicating with other players difficult as the "active" process
>> is changing) or should I make nested receive loops that handle the
>> substate machines?
>
> This sort of thing is a bit difficult to do well in gen_fsm, as you
> point out. One one hand, it provides built-in support for the OTP
> system messages, which is good, but on the other hand, it provides
> no good support for nested state machines.
>
> Whether you should break up your state machine in multiple processes
> is difficult to answer in general. Sometimes, it's a good idea, but
> other times it makes things more complex. You have to go through the
> different models in your mind (or on paper) and try to get a feeling
> for it.
>
> Reasonably simple sub states are nicely done with nested receives.
> The gen_fsm model has a problem if you may want to handle system
> messages (suspend, resume, change_code, get_status) in sub states -
> gen_fsm only does that on the top level. If this is a real problem,
> you can either use multiple gen_fsms (which may buy you other
> problems, like synchronization bugs), or you might try out plain_fsm,
> on jungerl. It tries to help you out with system messages without
> forcing you to leave the textbook Erlang fsm programming model.
>
> Discussions about plain_fsm on erlang-questions include:
>
> http://www.erlang.org/ml-archive/erlang-questions/200402/msg00043.html
>
> http://www.erlang.org/ml-archive/erlang-questions/200407/msg00032.html

Hey, the plain_fsm looks great! How come it didn't register in my mind as I 
was scanning the mailinglist... perhaps because it didn't straight out 
spell nes-ted FSM. But it really looks something that I've been looking 
for. Thanks!

> /Uffe
> -- 
> Ulf Wiger
>

-+-+-+-
Jani Launonen 
Student. . . . . . . . . .University of Oulu, Dept. of El. & Inf. Eng.

"Life is what happens to you while you're busy making future plans."
  - Alfred E. Neuman


From jamesh@REDACTED  Fri Dec 17 20:35:58 2004
From: jamesh@REDACTED (James Hague)
Date: Fri, 17 Dec 2004 13:35:58 -0600
Subject: record deep dive
Message-ID: <F7698C6BAED2D41189E00002A50A625F06A07DD2@v-exchange.volition.net>

Kostis Sagonas wrote:

>Wait till you see the changes in the compiler
>being hooked in... and then you'll really be
>impressed !

Aw, you're not just going to leave us wondering, are you?


From matthias@REDACTED  Fri Dec 17 22:05:34 2004
From: matthias@REDACTED (Matthias Lang)
Date: Fri, 17 Dec 2004 22:05:34 +0100
Subject: snmp-mnesia
In-Reply-To: <41C2CCEE.7050906@ast.dfs.de>
References: <41C2CCEE.7050906@ast.dfs.de>
Message-ID: <16835.18974.884916.644545@antilipe.corelatus.se>

Dietmar Admin Schaefer writes:

 > Is there anybody out there who works with snmp/mnesia ?

 > Is the question that I asked last week too complicated to be answered or 
 > too stupid to ask ?

It's hard to judge because I don't use SNMP or mnesia, but you could
make it easier for people to answer your question by cutting the
problem down to concise and complete example.

This might help:

  http://www.drxyzzy.org/mnesia-snmp/

Matt


From casper2000a@REDACTED  Mon Dec 20 09:26:24 2004
From: casper2000a@REDACTED (Casper)
Date: Mon, 20 Dec 2004 14:26:24 +0600
Subject: Mnesia adding 1,000,000 records, core dump
In-Reply-To: <16835.18974.884916.644545@antilipe.corelatus.se>
Message-ID: <!~!UENERkVCMDkAAQACAAAAAAAAAAAAAAAAABgAAAAAAAAA2eo0SPN1R0OAqojNqhlfEsKAAAAQAAAAyQ/1a/VsbEKrGhn2okYHzwEAAAAA@omnibis.com>

Hi all,

I'm trying to load test a Mnesia database. Attached are header and erlang
files I used to do the test. Below are the commands I ran (start erlang with
mnesia directory).

mnesia:start().
loadtest:init().
loadtest:init_counters().
loadtest:add_dummy_data(1000000).

Once I do that, the shell goes adding new entries and after some time prints
"Abnormal error" and hangs for ever. If I only tries to add about 100,000
records, it works without any problems.

If you need the erl_crash.dump file, I can send. It's about 19MB (8MB
compressd).

What could be there error?

Thanks in advance!
Eranga
-------------- next part --------------
A non-text attachment was scrubbed...
Name: loadtest.hrl
Type: application/octet-stream
Size: 317 bytes
Desc: not available
URL: <http://erlang.org/pipermail/erlang-questions/attachments/20041220/3e915898/attachment.obj>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: loadtest.erl
Type: application/octet-stream
Size: 3134 bytes
Desc: not available
URL: <http://erlang.org/pipermail/erlang-questions/attachments/20041220/3e915898/attachment-0001.obj>

From casper2000a@REDACTED  Mon Dec 20 09:39:19 2004
From: casper2000a@REDACTED (Casper)
Date: Mon, 20 Dec 2004 14:39:19 +0600
Subject: Mnesia adding 1,000,000 records, core dump
Message-ID: <!~!UENERkVCMDkAAQACAAAAAAAAAAAAAAAAABgAAAAAAAAA2eo0SPN1R0OAqojNqhlfEsKAAAAQAAAArf8+JIvjrkqkhhCV6hGINwEAAAAA@omnibis.com>

Hi all,

After checking the Erl_crash.dump file, I found the below line.

Slogan: no more index entries in atom_tab (max=1048576)
System version: Erlang (BEAM) emulator version 5.3.6.3 [source] [threads:0]

According to that the system has gone space out of atom table entries. How
can I increase the atom_tab (max=1048576) value?

Thanks!
Eranga




-----Original Message-----
From: Casper [mailto:casper2000a@REDACTED] 
Sent: Monday, December 20, 2004 2:26 PM
To: 'erlang-questions@REDACTED'
Subject: Mnesia adding 1,000,000 records, core dump

Hi all,

I'm trying to load test a Mnesia database. Attached are header and erlang
files I used to do the test. Below are the commands I ran (start erlang with
mnesia directory).

mnesia:start().
loadtest:init().
loadtest:init_counters().
loadtest:add_dummy_data(1000000).

Once I do that, the shell goes adding new entries and after some time prints
"Abnormal error" and hangs for ever. If I only tries to add about 100,000
records, it works without any problems.

If you need the erl_crash.dump file, I can send. It's about 19MB (8MB
compressd).

What could be there error?

Thanks in advance!
Eranga



From matthias@REDACTED  Mon Dec 20 09:50:36 2004
From: matthias@REDACTED (Matthias Lang)
Date: Mon, 20 Dec 2004 09:50:36 +0100
Subject: Mnesia adding 1,000,000 records, core dump
In-Reply-To: <!~!UENERkVCMDkAAQACAAAAAAAAAAAAAAAAABgAAAAAAAAA2eo0SPN1R0OAqojNqhlfEsKAAAAQAAAArf8+JIvjrkqkhhCV6hGINwEAAAAA@omnibis.com>
References: <!~!UENERkVCMDkAAQACAAAAAAAAAAAAAAAAABgAAAAAAAAA2eo0SPN1R0OAqojNqhlfEsKAAAAQAAAArf8+JIvjrkqkhhCV6hGINwEAAAAA@omnibis.com>
Message-ID: <16838.37468.230775.737472@antilipe.corelatus.se>

Casper writes:

 > Slogan: no more index entries in atom_tab (max=1048576)
 > System version: Erlang (BEAM) emulator version 5.3.6.3 [source] [threads:0]
 > 
 > According to that the system has gone space out of atom table entries. How
 > can I increase the atom_tab (max=1048576) value?

Increasing the size of the atom table is probably not the solution you
really want. 

What you really want to do is to avoid dynamically creating huge
numbers of atoms, don't call list_to_atom.

See also:

  http://www.erlang.org/ml-archive/erlang-questions/200309/msg00329.html

Matthias


From casper2000a@REDACTED  Mon Dec 20 10:12:50 2004
From: casper2000a@REDACTED (Casper)
Date: Mon, 20 Dec 2004 15:12:50 +0600
Subject: Mnesia adding 1,000,000 records, core dump
In-Reply-To: <16838.37468.230775.737472@antilipe.corelatus.se>
Message-ID: <!~!UENERkVCMDkAAQACAAAAAAAAAAAAAAAAABgAAAAAAAAA2eo0SPN1R0OAqojNqhlfEsKAAAAQAAAA7JgL4s2O502Sb2Qd157UOQEAAAAA@omnibis.com>

Hi Matthias,

Thanks a lot for the advice. What discussed in the email you pointed is like
Greek to me. Will try to understand.

Cheers,
Eranga




-----Original Message-----
From: Matthias [mailto:matthias@REDACTED] On Behalf Of Matthias Lang
Sent: Monday, December 20, 2004 2:51 PM
To: Casper
Cc: erlang-questions@REDACTED
Subject: RE: Mnesia adding 1,000,000 records, core dump

Casper writes:

 > Slogan: no more index entries in atom_tab (max=1048576)
 > System version: Erlang (BEAM) emulator version 5.3.6.3 [source]
[threads:0]
 > 
 > According to that the system has gone space out of atom table entries.
How
 > can I increase the atom_tab (max=1048576) value?

Increasing the size of the atom table is probably not the solution you
really want. 

What you really want to do is to avoid dynamically creating huge
numbers of atoms, don't call list_to_atom.

See also:

  http://www.erlang.org/ml-archive/erlang-questions/200309/msg00329.html

Matthias



From vlad_dumitrescu@REDACTED  Mon Dec 20 10:04:07 2004
From: vlad_dumitrescu@REDACTED (Vlad Dumitrescu)
Date: Mon, 20 Dec 2004 10:04:07 +0100
Subject: Mnesia adding 1,000,000 records, core dump
References: <!~!UENERkVCMDkAAQACAAAAAAAAAAAAAAAAABgAAAAAAAAA2eo0SPN1R0OAqojNqhlfEsKAAAAQAAAArf8+JIvjrkqkhhCV6hGINwEAAAAA@omnibis.com>
Message-ID: <BAY15-DAV1483D6BB4DE07ED16E3F719CA20@phx.gbl>

From: "Casper" <casper2000a@REDACTED>
> According to that the system has gone space out of atom table entries. How
> can I increase the atom_tab (max=1048576) value?

Hi,

You can't (with reservation for recent changes). But even if you could, you
should not use atoms like this anyway. Since atoms aren't garbage collected,
they are a precious resource.

I suppose you use list_to_atom to build atoms. Instead of for example
list_to_atom("root"+Number), consider using {root, Number} as a value. It works
just as fine as Mnesia index.

regards,
Vlad


From matthias@REDACTED  Mon Dec 20 10:28:05 2004
From: matthias@REDACTED (Matthias Lang)
Date: Mon, 20 Dec 2004 10:28:05 +0100
Subject: Mnesia adding 1,000,000 records, core dump
In-Reply-To: <!~!UENERkVCMDkAAQACAAAAAAAAAAAAAAAAABgAAAAAAAAA2eo0SPN1R0OAqojNqhlfEsKAAAAQAAAA7JgL4s2O502Sb2Qd157UOQEAAAAA@omnibis.com>
References: <16838.37468.230775.737472@antilipe.corelatus.se>
	<!~!UENERkVCMDkAAQACAAAAAAAAAAAAAAAAABgAAAAAAAAA2eo0SPN1R0OAqojNqhlfEsKAAAAQAAAA7JgL4s2O502Sb2Qd157UOQEAAAAA@omnibis.com>
Message-ID: <16838.39717.901503.798868@antilipe.corelatus.se>

Casper writes:

 > Thanks a lot for the advice. What discussed in the email you pointed is like
 > Greek to me. Will try to understand.

The mail I provided a link to is only of interest if you're wondering
"is it really necessary for an Erlang implementation to have a finite
atom table, and is it really so hard to GC that table?"

If you're just interested in writing programs that work, it suffices
to know that

  - in general, don't use 'list_to_atom'. (There are few legitimate 
    uses outside of debugging, it probably wouldn't hurt for the
    documentation to say something along those lines.)

  - it general, if something is a string (list), you should store it 
    as a string (list)

  - the current Erlang implementation has a limit to the number of
    atoms you can have. That limit is high enough so that this will
    only cause problems if you're doing strange things, such as
    creating atoms dynamically with 'list_to_atom'.

Matt


From ulf.wiger@REDACTED  Mon Dec 20 11:43:42 2004
From: ulf.wiger@REDACTED (Ulf Wiger (AL/EAB))
Date: Mon, 20 Dec 2004 11:43:42 +0100
Subject: bug in register/2 BIF
Message-ID: <37FB7AA6F5F9814FB634A7BF4C35A6F54029F2@ESEALNT442.al.sw.ericsson.se>


This bug was introduced in OTP R8B, I believe.

According to the manual for register(), you should get
exit (badarg) if the given process is not active. This
was true in OTP R7B, but isn't anymore (I've tried it 
on R7B, R8B, R9B, R9C and R10B.) If the P is not 
alive (but a local pid), register(Name, P) always 
returns true.

One really interesting characteristic of this bug is 
this:

1> Dead = spawn(fun() -> true end).
<0.39.0>
2> register(foo, self()).
true
3> register(foo,Dead).
true

This is what the manual says:

"register(Name, P)
Associates the name Name with the port or pid P. Name, which 
must be an atom, can be used instead of a port or pid in the 
send operator (Name ! Message). 
Returns true. 
Failure: badarg if P is not an active port or process, if P 
is on another node, if Name is already in use, if the port 
or process is already registered (already has a name), if 
Name is not an atom, or if Name is the atom undefined. "


/Uffe


From casper2000a@REDACTED  Mon Dec 20 11:59:44 2004
From: casper2000a@REDACTED (Casper)
Date: Mon, 20 Dec 2004 16:59:44 +0600
Subject: Mnesia adding 1,000,000 records, core dump
In-Reply-To: <16838.39717.901503.798868@antilipe.corelatus.se>
Message-ID: <!~!UENERkVCMDkAAQACAAAAAAAAAAAAAAAAABgAAAAAAAAA2eo0SPN1R0OAqojNqhlfEsKAAAAQAAAAMJ4dxXDv1EeH58GqrdSJtwEAAAAA@omnibis.com>


Thanks to all who enlighten me! It's clear now.



-----Original Message-----
From: Matthias [mailto:matthias@REDACTED] On Behalf Of Matthias Lang
Sent: Monday, December 20, 2004 3:28 PM
To: Casper
Cc: erlang-questions@REDACTED
Subject: RE: Mnesia adding 1,000,000 records, core dump

Casper writes:

 > Thanks a lot for the advice. What discussed in the email you pointed is
like
 > Greek to me. Will try to understand.

The mail I provided a link to is only of interest if you're wondering
"is it really necessary for an Erlang implementation to have a finite
atom table, and is it really so hard to GC that table?"

If you're just interested in writing programs that work, it suffices
to know that

  - in general, don't use 'list_to_atom'. (There are few legitimate 
    uses outside of debugging, it probably wouldn't hurt for the
    documentation to say something along those lines.)

  - it general, if something is a string (list), you should store it 
    as a string (list)

  - the current Erlang implementation has a limit to the number of
    atoms you can have. That limit is high enough so that this will
    only cause problems if you're doing strange things, such as
    creating atoms dynamically with 'list_to_atom'.

Matt



From joe.armstrong@REDACTED  Mon Dec 20 13:10:38 2004
From: joe.armstrong@REDACTED (Joe Armstrong (AL/EAB))
Date: Mon, 20 Dec 2004 13:10:38 +0100
Subject: Bug? in error_handler:unregistered_name/3
Message-ID: <D6A41B94D27EA643BAE9319F5348603FB8CEFC@ESEALNT895.al.sw.ericsson.se>


What's happened to error_handler:unregistered_name/3

According to page 101 of the Erlang book if I make a module (say my_error_handler) 
which exports unregistered_name(Name, Pid, Msg)

If then in some process (Pid) I call

	process_flag(error_handler, my_error_handler)
	abc ! foo

Then the standard error handler will be replaced by my_error handler
and if abc is not a registered process then

	my_error_handler:unregistered_name(abc, Pid, foo)

will be called.

This does not seem to work at all.

What's happened here?

/Joe




From bry@REDACTED  Mon Dec 20 15:42:32 2004
From: bry@REDACTED (bry@REDACTED)
Date: Mon, 20 Dec 2004 15:42:32 +0100
Subject: http://downloads.erlang-projects.org down?
Message-ID: <1103553752.41c6e4d84f252@horde.scannet.dk>


When i try to download something from http://downloads.erlang-projects.org it
says the file does not exist, if i go directly to
http://downloads.erlang-projects.org it says that it is not configured
correctly. 



From erlang@REDACTED  Mon Dec 20 15:39:03 2004
From: erlang@REDACTED (Inswitch Solutions - Erlang Evaluation)
Date: Mon, 20 Dec 2004 12:39:03 -0200
Subject: Binary matching
Message-ID: <06aa01c4e6a1$b6c678c0$1e00a8c0@Inswitch251>


Does someone know why the following binary matching fails?
 <<X:7/binary, Rest/binary>> = <<0>>

Is there an elegant way to group an octect binary in seven bits values?
For example eight characters (1 character = 1 byte) in seven octects.


thanks,
Eduardo Figoli
INSwitch Solutions





-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://erlang.org/pipermail/erlang-questions/attachments/20041220/d5c3a4b6/attachment.htm>

From sean.hinde@REDACTED  Mon Dec 20 16:05:22 2004
From: sean.hinde@REDACTED (Sean Hinde)
Date: Mon, 20 Dec 2004 15:05:22 +0000
Subject: [ANN] md5_crypt
Message-ID: <92574EC7-5298-11D9-8B4C-000A95927CCE@mac.com>

Hi,

Just finished fiddling with an Erlang port of the apache/BSD style md5 
crypt routine used for logins and htpasswd files etc

Someone might find it useful.

Sean

-------------- next part --------------
A non-text attachment was scrubbed...
Name: md5_crypt.erl
Type: application/octet-stream
Size: 6335 bytes
Desc: not available
URL: <http://erlang.org/pipermail/erlang-questions/attachments/20041220/123c8e17/attachment.obj>

From carsten@REDACTED  Mon Dec 20 16:24:29 2004
From: carsten@REDACTED (Carsten Schultz)
Date: Mon, 20 Dec 2004 16:24:29 +0100
Subject: Binary matching
In-Reply-To: <06aa01c4e6a1$b6c678c0$1e00a8c0@Inswitch251>
References: <06aa01c4e6a1$b6c678c0$1e00a8c0@Inswitch251>
Message-ID: <20041220152428.GB10362@penne.localnet>

Hi!

On Mon, Dec 20, 2004 at 12:39:03PM -0200, 
Inswitch Solutions - Erlang Evaluation wrote:
>  
> Does someone know why the following binary matching fails?
>  <<X:7/binary, Rest/binary>> = <<0>>
>  
> Is there an elegant way to group an octect binary in seven bits values?
> For example eight characters (1 character = 1 byte) in seven octects.

I think you have to align the Rest on a byte boundary yourself.  For
example the following code is working:

base64encode(B) ->
    base64encode(B, []).
base64encode(<<A:6,B:6,C:6,D:6,T/binary>>, Acc) -> 
						% 3+ bytes
    base64encode(T, [b64e(D),b64e(C),b64e(B),b64e(A)|Acc]);
base64encode(<<A:6,B:6,C:4>>, Acc) -> 
						% 2 bytes
    lists:reverse([$=,b64e(C bsl 2),b64e(B),b64e(A)|Acc]);
base64encode(<<A:6,B:2>>, Acc) ->
						% 1 byte
    lists:reverse([$=,$=,b64e(B bsl 4), b64e(A)|Acc]);
base64encode(<<>>, Acc) ->
						% 0 bytes
    lists:reverse(Acc).


b64e(X) when X >= 0, X < 26 -> 
    X + $A;
b64e(X) when X >= 26, X < 52 -> 
    X - 26 + $a;
b64e(X) when X >= 52, X < 62 ->
    X - 52 + $0;
b64e(62) ->
    $+;
b64e(63) ->
    $/.


HTH

Carsten

-- 
Carsten Schultz (2:38, 33:47), FB Mathematik, FU Berlin
http://carsten.codimi.de/
PGP/GPG key on the pgp.net key servers, 
fingerprint on my home page.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: not available
URL: <http://erlang.org/pipermail/erlang-questions/attachments/20041220/6c051065/attachment.bin>

From mickael.remond@REDACTED  Mon Dec 20 16:24:02 2004
From: mickael.remond@REDACTED (Mickael Remond)
Date: Mon, 20 Dec 2004 16:24:02 +0100
Subject: http://downloads.erlang-projects.org down?
In-Reply-To: <1103553752.41c6e4d84f252@horde.scannet.dk>
References: <1103553752.41c6e4d84f252@horde.scannet.dk>
Message-ID: <41C6EE92.8030905@erlang-fr.org>

bry@REDACTED wrote:
> When i try to download something from http://downloads.erlang-projects.org it
> says the file does not exist, if i go directly to
> http://downloads.erlang-projects.org it says that it is not configured
> correctly. 

Yes. The server had a crash disk. The machine under repair. I hope the 
hosting provider will reinstall the Operating system soon.

When ready I will reinstall the application from our backup and send a 
notice to this mailing list.

For the record, this is the second hard drive crash we had on 
erlang-projects subsite. I will really start to like hardware RAID 
solution a lot :-)

Sorry for the inconvenience. We are doing our best to put it online as 
soon as possible.

-- 
Micka?l R?mond


From mickael.remond@REDACTED  Mon Dec 20 16:56:55 2004
From: mickael.remond@REDACTED (Mickael Remond)
Date: Mon, 20 Dec 2004 16:56:55 +0100
Subject: http://downloads.erlang-projects.org down?
In-Reply-To: <41C6EE92.8030905@erlang-fr.org>
References: <1103553752.41c6e4d84f252@horde.scannet.dk> <41C6EE92.8030905@erlang-fr.org>
Message-ID: <41C6F647.8020406@erlang-fr.org>

Hello again,

Fredrik Thulin did set up a mirror of the Erlang REPOS CDROM before our 
crash disk.

So if you are especially interested in the REPOS image, you can get it from:

ftp://ftp.su.se/pub/mirrors/download.erlang-projects.org/

I hope this helps :-)

-- 
Micka?l R?mond


From bjorn@REDACTED  Thu Dec  9 11:05:48 2004
From: bjorn@REDACTED (Bjorn Gustavsson)
Date: 09 Dec 2004 11:05:48 +0100
Subject: About the open-source and commercial releases
In-Reply-To: <16824.4284.274260.952091@gargle.gargle.HOWL>
References: <41B650C6.6010201@free.fr>
	<16822.49718.29049.233958@gargle.gargle.HOWL>
	<41B70E74.6030104@free.fr>
	<16823.7685.261964.864904@gargle.gargle.HOWL>
	<20041208202305.GA11881@hyber.org>
	<16824.4284.274260.952091@gargle.gargle.HOWL>
Message-ID: <o1oeh36b2b.fsf_-_@erix.ericsson.se>

Ingela Anderton <ingela@REDACTED> writes:

> I actually was a little surprised that the patch had not made it
> to the open source version yet, the patch must just have missed the
> last open source release.

It is not directly related to the original question, but I will
take this opportunity to inform about open-source and commercial
releases, and how we now try to break the "old tradition" of releasing
the R*B-1 release more than six months after the original R*B release.

We never intended to delay the initial bug fix release by six months.
It was just that there was another necessary correction to some appliation
or the other that could be included if we just waited a few days more. When
that correction was finished, we could have another great correction in
another application if we just waited four days more... and so on.

Therefore, our current policy is to release the open-source releases regularily,
at least once every two months. Generally, we set our internal deadline
for when the next release will be out, and patches that are not ready will
have to wait for the next release.

We've have always had the policy that paying customers should have access
to patches before we release them as open-source.


We usually don't announce beforehand when an open-source release is to be released,
but this time I can tell that it is very likely that R10B-2 will be released
next week.

/Bjorn
-- 
Bj?rn Gustavsson, Erlang/OTP, Ericsson AB


From frej@REDACTED  Tue Dec 21 07:04:05 2004
From: frej@REDACTED (Frej Drejhammar)
Date: 21 Dec 2004 07:04:05 +0100
Subject: Win XP SP2 and gen_tcp:connect/3
Message-ID: <tz8fz20jiga.fsf@pingvin.sics.se>

Hi,

On my friend's Wintendo-box (XP SP2, English, corporate pro) running
R10B-2, gen_tcp:connect/3 fails with enotsock. It does so
irrespectively of port number and options to connect. It is not a
firewall issue, telnet connects just fine. Has anyone else seen this?

Best regards,

Frej


From matthias@REDACTED  Tue Dec 21 11:02:32 2004
From: matthias@REDACTED (Matthias Lang)
Date: Tue, 21 Dec 2004 11:02:32 +0100
Subject: Win XP SP2 and gen_tcp:connect/3
In-Reply-To: <tz8fz20jiga.fsf@pingvin.sics.se>
References: <tz8fz20jiga.fsf@pingvin.sics.se>
Message-ID: <16839.62648.591702.639438@antilipe.corelatus.se>

Frej Drejhammar writes:

 > On my friend's Wintendo-box (XP SP2, English, corporate pro) running
 > R10B-2, gen_tcp:connect/3 fails with enotsock. It does so
 > irrespectively of port number and options to connect. It is not a
 > firewall issue, telnet connects just fine. Has anyone else seen this?

Other people have seen this:

  http://www.erlang.org/ml-archive/erlang-questions/200411/msg00120.html

As far as I know, nobody actually knows why it happens and nobody has a
fix.

Matthias


From danie@REDACTED  Tue Dec 21 23:20:03 2004
From: danie@REDACTED (Danie Schutte)
Date: Wed, 22 Dec 2004 00:20:03 +0200
Subject: Binary matching
In-Reply-To: <06aa01c4e6a1$b6c678c0$1e00a8c0@Inswitch251>
References: <06aa01c4e6a1$b6c678c0$1e00a8c0@Inswitch251>
Message-ID: <1103667603.6547.7.camel@blackbox.erlfinsys.net>

On Mon, 2004-12-20 at 12:39 -0200, Inswitch Solutions - Erlang
Evaluation wrote:

>  
> Does someone know why the following binary matching fails?
>  <<X:7/binary, Rest/binary>> = <<0>>

I think the problem lies with the fact that X:7/binary will look for 7
binary BYTES of data and not 7 BITS.
<<X:7, Rest/binary>> = <<0>> might be more what you are looking for.


>  
> Is there an elegant way to group an octect binary in seven bits
> values?
> For example eight characters (1 character = 1 byte) in seven octects.
>  
>  
> thanks,
> Eduardo Figoli
> INSwitch Solutions
>  
>  
>  
>  
>  

-- 
Danie Schutte <danie@REDACTED>
Erlang Financial Systems
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://erlang.org/pipermail/erlang-questions/attachments/20041222/3a9aed5e/attachment.htm>

From vances@REDACTED  Tue Dec 21 23:41:54 2004
From: vances@REDACTED (Vance Shipley)
Date: Tue, 21 Dec 2004 17:41:54 -0500
Subject: Binary matching
In-Reply-To: <1103667603.6547.7.camel@blackbox.erlfinsys.net>
References: <06aa01c4e6a1$b6c678c0$1e00a8c0@Inswitch251> <1103667603.6547.7.camel@blackbox.erlfinsys.net>
Message-ID: <20041221224154.GK22325@frogman.motivity.ca>

Danie,

That would make Rest a binary of 1 bit which is illegal.

	-Vance


On Wed, Dec 22, 2004 at 12:20:03AM +0200, Danie Schutte wrote:
}  On Mon, 2004-12-20 at 12:39 -0200, Inswitch Solutions - Erlang
}  Evaluation wrote:
}  
}  >  
}  > Does someone know why the following binary matching fails?
}  >  <<X:7/binary, Rest/binary>> = <<0>>
}  
}  I think the problem lies with the fact that X:7/binary will look for 7
}  binary BYTES of data and not 7 BITS.
}  <<X:7, Rest/binary>> = <<0>> might be more what you are looking for.
}  
}  
}  >  
}  > Is there an elegant way to group an octect binary in seven bits
}  > values?
}  > For example eight characters (1 character = 1 byte) in seven octects.
}  >  
}  >  
}  > thanks,
}  > Eduardo Figoli
}  > INSwitch Solutions
}  >  
}  >  
}  >  
}  >  
}  >  
}  
}  -- 
}  Danie Schutte <danie@REDACTED>
}  Erlang Financial Systems


From dhanasekaran.c@REDACTED  Wed Dec 22 06:38:42 2004
From: dhanasekaran.c@REDACTED (Chenniappan,Dhanasekaran [DBA])
Date: Wed, 22 Dec 2004 11:08:42 +0530
Subject: Erl Ports and Threads
Message-ID: <124AD2BD2C1FB245B2E4FF0E90AE0E324A57C4@ntserver.dbachn>

Hi, 
 I am ODBC module for Microsoft SQL server. I am struggling with call a
stored procedure, If anybody have some working example (With MS SQL stored
procedure) please help me.

With Regards
Dhanas

-----Original Message-----
From: Per Hedeland [mailto:hedeland@REDACTED] 
Sent: Thursday, December 16, 2004 3:41 PM
To: casper2000a@REDACTED
Cc: erlang-questions@REDACTED; klacke@REDACTED
Subject: Re: Erl Ports and Threads

klacke@REDACTED wrote:
>
>On Thu, Dec 16, 2004 at 03:37:34PM +0600, Casper wrote:
>> Hi Klacke,
>> 
>> Thanks for the response.
>> 
>> Here I am not talking about simultaneous write() operations on a
socket/pipe
>> from the multiple external program. I am talking about multiple threads
in
>> one program.
>
>I know, that's what you wrote.
>
>> 
>> Is single write() call, atomic or not?
>> 
>
>Not.

<nitpick>
Actually this isn't entirely true - when different processes write to
the same pipe, their writes are guaranteed (on a standards-compliant
Unix:-) to be atomic if they are "small enough" - from a decent
pathconf(2) man page:

     _PC_PIPE_BUF
             The maximum number of bytes which will be written atomically to
a
             pipe.

(The number is typically 4096 or so.) Whether this is still true for
different threads in a process writing to the same pipe I don't know,
but I would expect so (might be implementation-dependant though). And of
course relying on the undocumented fact that the port communication
channel *is* a pipe, and not e.g. a local socket which doesn't have such
guarantees, isn't a very good idea.
</nitpick>

For all practical purposes, I agree with Klacke's answer.:-)

--Per


From casper2000a@REDACTED  Wed Dec 22 07:13:47 2004
From: casper2000a@REDACTED (Casper)
Date: Wed, 22 Dec 2004 12:13:47 +0600
Subject: Erlang Port Drivers and Erl functions
In-Reply-To: <1103667603.6547.7.camel@blackbox.erlfinsys.net>
Message-ID: <!~!UENERkVCMDkAAQACAAAAAAAAAAAAAAAAABgAAAAAAAAA2eo0SPN1R0OAqojNqhlfEsKAAAAQAAAAeZq7z+xtWkyuNcEurj+sTAEAAAAA@omnibis.com>

Hi All,

 

I'm developing a Erlang <-> C port using Port Drivers. As per the
communication message, I try to use Erl_Interface functions for easy message
encoding/decoding. But the program generates a segmentation fault.

 

Below are the details I do.

 

Start Port Driver:

            Port = open_port({spawn, './lib.so'}, []),                %%
inside a spawned process

 

Send a message to the C Port:

            Port ! {self(), {command, term_to_binary(Msg)}},  %% binary
encode the message

 

Receive the message in the C Port:

            static void nmslib_drv_output(ErlDrvData handle, char *buff, int
bufflen) {

                        ETERM *tuple;

                        tuple = erl_decode(buff);

 

As soon as I do the erl_decode, the program gets segmentation fault. As I
suspect, the buff doesn't have the correct message format that erl_decode
expect. Could anybody kindly advice me what's that difference.

 

If I don't use erl_decode, inside the C Port, the Port Driver works without
any problems.

 

Thanks in advance!

Eranga

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://erlang.org/pipermail/erlang-questions/attachments/20041222/736d4aea/attachment.htm>

From techieyang@REDACTED  Wed Dec 22 11:12:52 2004
From: techieyang@REDACTED (techie)
Date: Wed, 22 Dec 2004 18:12:52 +0800
Subject: Newbie question: cross-compile PPC860/LINUX
Message-ID: <41C948A4.2000004@baniantech.com>

I am a newbie to the Erlang/Opt. Looking forward to your help. Thanks!

Several question:

First Question: Failed when Cross-Compile R10B-2, target --
PPC860/Linux, Host -- RedHat-ES3-X86

Error Information got:

ppc_8xx-gcc -g -O2
-I/opt/erlang/test/otp_src_R10B-2/erts/i686-pc-linux-gnu -DHYBRID
-DHAVE_CONFIG_H -Wall -Wstrict-prototypes -Wmissing-prototypes -Ibeam
-Isys/unix -Isys/common -Ii686-pc-linux-gnu -Ii686-pc-linux-gnu/hybrid
-Izlib -Ihipe -o
/opt/erlang/test/otp_src_R10B-2/bin/i686-pc-linux-gnu/hipe_mkliterals.hybrid
/opt/erlang/test/otp_src_R10B-2/erts/obj.hybrid.beam/i686-pc-linux-gnu/hipe_mkliterals.o
/opt/erlang/test/otp_src_R10B-2/bin/i686-pc-linux-gnu/hipe_mkliterals.hybrid
-c > i686-pc-linux-gnu/hybrid/hipe_literals.h
/bin/sh: line 1:
/opt/erlang/test/otp_src_R10B-2/bin/i686-pc-linux-gnu/hipe_mkliterals.hybrid:
cannot execute binary file
make[4]: *** [i686-pc-linux-gnu/hybrid/hipe_literals.h] Error 126
make[4]: Leaving directory `/opt/erlang/test/otp_src_R10B-2/erts/emulator'
make[3]: *** [hybrid] Error 2
make[3]: Leaving directory `/opt/erlang/test/otp_src_R10B-2/erts/emulator'
make[2]: *** [hybrid] Error 2
make[2]: Leaving directory `/opt/erlang/test/otp_src_R10B-2/erts'
make[1]: *** [emulator] Error 2
make[1]: Leaving directory `/opt/erlang/test/otp_src_R10B-2'
make: *** [noboot] Error 2

I just use the configuration learning from the step-by-step-cross-compile

(1)./configure --prefix=/opt/erlang/otp/erlang --without-ssl --disable-hipe
(2) find . -name Makefile -exec ../mung \{\} \;
(3) ../mung make/i686-pc-linux-gnu/otp.mk

Question Two: How many files needed to run on the PPC860-target
(Embedded-ppc860-board)

I can cross-compile the PPC860-Linux with R7B4. Then i will copy the
built-files onto the target RAMDISK.
I tried to use the way in the 'system principles', but failed.
Can anybody give me a hints how to make this step success to run my
target-otp on the target board.

Thanks a lot!

Best Regards
Techie



From mikpe@REDACTED  Wed Dec 22 12:48:06 2004
From: mikpe@REDACTED (Mikael Pettersson)
Date: Wed, 22 Dec 2004 12:48:06 +0100 (MET)
Subject: Newbie question: cross-compile PPC860/LINUX
Message-ID: <200412221148.iBMBm6M4028669@harpo.it.uu.se>

On Wed, 22 Dec 2004 18:12:52 +0800, techieyang@REDACTED wrote:
>First Question: Failed when Cross-Compile R10B-2, target --
>PPC860/Linux, Host -- RedHat-ES3-X86
>
>Error Information got:
>
>ppc_8xx-gcc -g -O2
>-I/opt/erlang/test/otp_src_R10B-2/erts/i686-pc-linux-gnu -DHYBRID
>-DHAVE_CONFIG_H -Wall -Wstrict-prototypes -Wmissing-prototypes -Ibeam
>-Isys/unix -Isys/common -Ii686-pc-linux-gnu -Ii686-pc-linux-gnu/hybrid
>-Izlib -Ihipe -o
>/opt/erlang/test/otp_src_R10B-2/bin/i686-pc-linux-gnu/hipe_mkliterals.hybrid
>/opt/erlang/test/otp_src_R10B-2/erts/obj.hybrid.beam/i686-pc-linux-gnu/hipe_mkliterals.o
>/opt/erlang/test/otp_src_R10B-2/bin/i686-pc-linux-gnu/hipe_mkliterals.hybrid
>-c > i686-pc-linux-gnu/hybrid/hipe_literals.h
>/bin/sh: line 1:
>/opt/erlang/test/otp_src_R10B-2/bin/i686-pc-linux-gnu/hipe_mkliterals.hybrid:
>cannot execute binary file
>make[4]: *** [i686-pc-linux-gnu/hybrid/hipe_literals.h] Error 126

The problem is that $HOST can't execute a file compiled for $TARGET.
Maybe I haven't looked sufficiently closely, but I have yet to find
any sensible cross-compilation infrastructure in Erlang/OTP.

In this case, rebuilding hipe_mkliterals with a non-cross-compiler
should allow you to make some progress.

(FWIW, I regularly cross-compile Linux kernels on x86 for x86-64
or ppc32, but that's easy since there's a built-in cross-compilation
infrastructure in the Linux kernel source code.)

>Question Two: How many files needed to run on the PPC860-target
>(Embedded-ppc860-board)
>
>I can cross-compile the PPC860-Linux with R7B4. Then i will copy the
>built-files onto the target RAMDISK.

Given that you have compile a pile of .erl and .java files,
I'd suggest that you bootstrap on the host and install somewhere,
then remove the host-specific executables (not .beam or .class
files), rebuild them with a cross-compiler, and reinstall them.

/Mikael


From dietmar@REDACTED  Wed Dec 22 13:26:52 2004
From: dietmar@REDACTED (Dietmar Admin Schaefer)
Date: Wed, 22 Dec 2004 13:26:52 +0100
Subject: Question about records
Message-ID: <41C9680C.4080802@ast.dfs.de>

Hi !


I have:

-record(cmmc_config,{...}).


after reading the docs (more or less carefully) it's not yet clear to me


what   lists:foldl(F, #cmmc_config{}, Lines) means.
(#cmmc_config{} is the point)


regards



Dietmar


From techieyang@REDACTED  Wed Dec 22 13:31:58 2004
From: techieyang@REDACTED (techie)
Date: Wed, 22 Dec 2004 20:31:58 +0800
Subject: Newbie question: cross-compile PPC860/LINUX
In-Reply-To: <200412221148.iBMBm6M4028669@harpo.it.uu.se>
References: <200412221148.iBMBm6M4028669@harpo.it.uu.se>
Message-ID: <41C9693E.4080902@baniantech.com>


If it doesnt disturb you too much, could you please give more hints? Thanks!
Suppose i have already cross-compile the Erlang-PPC32-target on my
X86-linux-host
How to make a "bootstrap on the host and install somewhere"? Thanks!


>
>Mikael Pettersson>
>Given that you have compile a pile of .erl and .java files,
>I'd suggest that you bootstrap on the host and install somewhere,
>then remove the host-specific executables (not .beam or .class
>files), rebuild them with a cross-compiler, and reinstall them.
>
>/Mikael
>
>
>  
>



From fredrik.linder@REDACTED  Wed Dec 22 13:55:48 2004
From: fredrik.linder@REDACTED (Fredrik Linder)
Date: Wed, 22 Dec 2004 13:55:48 +0100
Subject: Question about records
Message-ID: <AFEA9020CFE63C448D6506ECE64F2F9874BA70@kangoo.sto.satsafe.se>

Hi

I hope this makes it clearer:

With

-record(a, {b, c}).

1> A1 = #a{}.
{a, undefined, undefined}.
2> A1#a.b.
undefined.
3> A1#A.c.
undefined.
4> A2 = #a{b = 1}.
{a, 1, undefined}.
5> A2#a.b.
1
6> lists:foldl(fun(A, Acc) -> Acc end, #a{}, [1]).
[{a, undefined, undefined}].

Merry Christmas
/Fredrik

> -----Original Message-----
> From: owner-erlang-questions@REDACTED
> [mailto:owner-erlang-questions@REDACTED]On Behalf Of Dietmar Admin
> Schaefer
> Sent: den 22 december 2004 13:27
> To: 'Erlang Questions List'
> Subject: Question about records
> 
> 
> Hi !
> 
> 
> I have:
> 
> -record(cmmc_config,{...}).
> 
> 
> after reading the docs (more or less carefully) it's not yet 
> clear to me
> 
> 
> what   lists:foldl(F, #cmmc_config{}, Lines) means.
> (#cmmc_config{} is the point)
> 
> 
> regards
> 
> 
> 
> Dietmar
> 


From carsten@REDACTED  Wed Dec 22 13:58:28 2004
From: carsten@REDACTED (Carsten Schultz)
Date: Wed, 22 Dec 2004 13:58:28 +0100
Subject: Question about records
In-Reply-To: <41C9680C.4080802@ast.dfs.de>
References: <41C9680C.4080802@ast.dfs.de>
Message-ID: <20041222125827.GD16468@penne.localnet>

Hi!

On Wed, Dec 22, 2004 at 01:26:52PM +0100, Dietmar Admin Schaefer wrote:
> -record(cmmc_config,{...}).
> 
> 
> after reading the docs (more or less carefully) it's not yet clear to me
> 
> 
> what   lists:foldl(F, #cmmc_config{}, Lines) means.
> (#cmmc_config{} is the point)

#cmmc_config{} is just a record with all entries set to their default
values.  

Take care,

Carsten

-- 
Carsten Schultz (2:38, 33:47), FB Mathematik, FU Berlin
http://carsten.codimi.de/
PGP/GPG key on the pgp.net key servers, 
fingerprint on my home page.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: not available
URL: <http://erlang.org/pipermail/erlang-questions/attachments/20041222/0c347259/attachment.bin>

From matthias@REDACTED  Wed Dec 22 15:14:51 2004
From: matthias@REDACTED (Matthias Lang)
Date: Wed, 22 Dec 2004 15:14:51 +0100
Subject: Newbie question: cross-compile PPC860/LINUX
In-Reply-To: <200412221148.iBMBm6M4028669@harpo.it.uu.se>
References: <200412221148.iBMBm6M4028669@harpo.it.uu.se>
Message-ID: <16841.33115.988701.454939@antilipe.corelatus.se>

Mikael Pettersson writes:

 > The problem is that $HOST can't execute a file compiled for $TARGET.
 > Maybe I haven't looked sufficiently closely, but I have yet to find
 > any sensible cross-compilation infrastructure in Erlang/OTP.

There isn't any. It'd be a fair chunk of work to get that working.

The cross-compile for VxWorks was driven by a bunch of sed scripts,
that's what inspired my hack for PPC860-linux. 

Matt


From jamesh@REDACTED  Thu Dec 23 22:59:05 2004
From: jamesh@REDACTED (James Hague)
Date: Thu, 23 Dec 2004 15:59:05 -0600
Subject: Windows: What's the proper way to run without a console?
Message-ID: <F7698C6BAED2D41189E00002A50A625F06A07DFB@v-exchange.volition.net>

There are three command-line switches:

-noshell: means boot up, but don't start an interactive top-level
-noinput: not sure what this means
-detached: means run without standard input or output (this is just like
running werl normally, except that there's no visible indication that it's
running)

There's also an environment variable, ERL_CONSOLE_MODE, which can be set to
"window" or "tty:something". (What does the latter of these *do* under
Windows?)

Now there's also the issue of when the Erlang GUI application shows up.
"-noshell" still causes it to be visible, which doesn't seem like the
correct behavior (or maybe it is; maybe it's where standard output goes by
default?).  I tracked through the code and, yes, the GUI console is still
initialized when -noshell is specified.  Bug?  If so, I haven't had any luck
fixing it!

Could someone enlighten me about how all these switches and ERL_CONSOLE_MODE
are supposed to work?

James


From bry@REDACTED  Fri Dec 24 18:07:42 2004
From: bry@REDACTED (bry@REDACTED)
Date: Fri, 24 Dec 2004 18:07:42 +0100
Subject: commandline start erlang, next commandline go to first instance?
Message-ID: <1103908062.41cc4cde7eb7c@horde.scannet.dk>



Hey, i was just wondering how one would do the following in a windows
environment.

start erlang and pass it a commandline for analysis, every subsequent call to
erlang via the commandline does not start a new instance but instead passes to
the originally started instance?



From olgeni@REDACTED  Sat Dec 25 16:45:44 2004
From: olgeni@REDACTED (Jimmy Olgeni)
Date: Sat, 25 Dec 2004 16:45:44 +0100 (CET)
Subject: rb:show/1 problem
Message-ID: <20041225164349.K16425@olgeni.olgeni>


[reposting - looks like it got lost somewhere]

Hi,

I still get the rb:show/1 problems with R10B-2: when called twice, it 
displays "rb: ERROR! A report on bad form was encountered. It can not 
be printed to the log.", then the details field, then "rb: Can not 
reopen standard_io. Logging aborted.".

Does anybody have a patch to fix this? :)

-- 
jimmy


From dietmar@REDACTED  Mon Dec 27 12:01:26 2004
From: dietmar@REDACTED (Dietmar Admin Schaefer)
Date: Mon, 27 Dec 2004 12:01:26 +0100
Subject: Problem with snmpa
Message-ID: <41CFEB86.8090108@ast.dfs.de>

Hi !


Using otp_R10B-2 starting snmpa (application:start(snmpa))

gives me:


 application:start(snmpa).
{error,{"unknown POSIX error","snmpa.app"}}.


Any ideas ?



Regards


Dietmar




From vlad_dumitrescu@REDACTED  Mon Dec 27 13:25:45 2004
From: vlad_dumitrescu@REDACTED (Vlad Dumitrescu)
Date: Mon, 27 Dec 2004 13:25:45 +0100
Subject: Problem with snmpa
References: <41CFEB86.8090108@ast.dfs.de>
Message-ID: <BAY15-DAV9C0780CDDFB53C4DCBC089C990@phx.gbl>

Hi,

> Using otp_R10B-2 starting snmpa (application:start(snmpa))
> gives me:
>  application:start(snmpa).
> {error,{"unknown POSIX error","snmpa.app"}}.

I get the same error message on Windows, but the problem might not be OS
specific (although I guess it is).

In my case, I don't have any snmpa.app file that the runtime can find. This
makes it return an error from application_controller:

    case code:where_is_file(FName) of
 non_existing ->
     {error, {file:format_error({error,enoent}), FName}};

For some strange reason,
    file:format_error({error,enoent})
evaluates to
    "unknown POSIX error".

I don't really know how you should proceed, but at least you know where to start
:-)

regards,
Vlad


From vlad_dumitrescu@REDACTED  Mon Dec 27 13:28:29 2004
From: vlad_dumitrescu@REDACTED (Vlad Dumitrescu)
Date: Mon, 27 Dec 2004 13:28:29 +0100
Subject: Problem with snmpa
References: <41CFEB86.8090108@ast.dfs.de>
Message-ID: <BAY15-DAV141D10C5BC5EF16B13F3659C990@phx.gbl>

> Using otp_R10B-2 starting snmpa (application:start(snmpa))
> gives me:
>  application:start(snmpa).
> {error,{"unknown POSIX error","snmpa.app"}}.

Hi again,

After reading the docs, I'd venture to guess that snmpa is not an application.
The application is called snmp, and there are several interface modules, among
others snmp and snmpa.

regards,
Vlad


From vlad_dumitrescu@REDACTED  Mon Dec 27 13:44:38 2004
From: vlad_dumitrescu@REDACTED (Vlad Dumitrescu)
Date: Mon, 27 Dec 2004 13:44:38 +0100
Subject: Problem with snmpa (bug found)
References: <41CFEB86.8090108@ast.dfs.de>
Message-ID: <BAY15-DAV1613A97CE8FF9D32D03EDC9C990@phx.gbl>

To round up this issue, I think the bug is simply that in application_controller
there is a simple error:

1401c1401
<      {error, {file:format_error({error,enoent}), FName}};
---
>      {error, {file:format_error(enoent), FName}};

regards,
Vlad


From dietmar@REDACTED  Mon Dec 27 14:37:20 2004
From: dietmar@REDACTED (Dietmar Admin Schaefer)
Date: Mon, 27 Dec 2004 14:37:20 +0100
Subject: Problem with snmpa
In-Reply-To: <41CFEB86.8090108@ast.dfs.de>
References: <41CFEB86.8090108@ast.dfs.de>
Message-ID: <41D01010.8090006@ast.dfs.de>

> Using otp_R10B-2 starting snmpa (application:start(snmpa))
> gives me:
>  application:start(snmpa).
> {error,{"unknown POSIX error","snmpa.app"}}.

Hi again,

After reading the docs, I'd venture to guess that snmpa is not an 
application.
The application is called snmp, and there are several interface modules, 
among
others snmp and snmpa.

regards,
Vlad

O.K.

I edited the application_controler AND I set TOP_ERL environment 
variable to the propper location -> ok.


Which one actually did it -> unknown.


Regards


Dietmar




From bjorn@REDACTED  Mon Dec 27 21:48:21 2004
From: bjorn@REDACTED (=?ISO-8859-1?Q?Bj=F6rn_Bylander?=)
Date: Mon, 27 Dec 2004 21:48:21 +0100
Subject: unknown_catch_try_state under R10B-2
Message-ID: <41D07515.8000306@loxysoft.se>


I get the following output from make:all([load]) with R10B-2 under 
Windows. The module (admctrl_srv.erl) compiles fine under R9C-2, R9C-1 
and R9C-0. I haven't tried other R10 releases.

Recompile: admctrl_srv
admctrl_srv: function handle_info/2+2032:
   Internal consistency check failed - please report this bug.
   Instruction: {'%live',1}
   Error:       unknown_catch_try_state:

./Admctrl_srv.erl:103: Warning: variable 'From' is unused

Apart from the obvious bug I find the first warning message odd in that 
the 'a' in the module file name is an upper case 'a' and not a lower 
case 'a' as it actually is and also is written in the warnings following 
it...

Is there any info in the error message about the row on which this 
occurs or do I have to try locate the row some other way?

/Bj?rn Bylander

-------------- next part --------------
A non-text attachment was scrubbed...
Name: bjorn.vcf
Type: text/x-vcard
Size: 276 bytes
Desc: not available
URL: <http://erlang.org/pipermail/erlang-questions/attachments/20041227/b1a10bee/attachment.vcf>

From bjorn@REDACTED  Mon Dec 27 23:05:03 2004
From: bjorn@REDACTED (=?ISO-8859-1?Q?Bj=F6rn_Bylander?=)
Date: Mon, 27 Dec 2004 23:05:03 +0100
Subject: unknown_catch_try_state under R10B-2
In-Reply-To: <41D07515.8000306@loxysoft.se>
References: <41D07515.8000306@loxysoft.se>
Message-ID: <41D0870F.9040600@loxysoft.se>


I've reduced the incorrect code to this:

-module(foomod).
-export([handle_info/2]).
-record(state, { foo = undefined }).

handle_info({_}, State) ->
   foo(),
   State#state{foo = bar},
   case ok of
   _ ->
     case catch foo() of
     ok ->
       {stop, State}
     end
   end;

handle_info(_, State) ->
   (catch begin
     foo(),
     State#state{foo = bar}
   end),
   case ok of
   _ ->
     case catch foo() of
     ok ->
       {stop, State}
     end
   end.

foo() -> ok.

On 2004-12-27 21:48, Bj?rn Bylander wrote:
> 
> I get the following output from make:all([load]) with R10B-2 under 
> Windows. The module (admctrl_srv.erl) compiles fine under R9C-2, R9C-1 
> and R9C-0. I haven't tried other R10 releases.
> 
> Recompile: admctrl_srv
> admctrl_srv: function handle_info/2+2032:
>   Internal consistency check failed - please report this bug.
>   Instruction: {'%live',1}
>   Error:       unknown_catch_try_state:
> 
> ./Admctrl_srv.erl:103: Warning: variable 'From' is unused
> 
> Apart from the obvious bug I find the first warning message odd in that 
> the 'a' in the module file name is an upper case 'a' and not a lower 
> case 'a' as it actually is and also is written in the warnings following 
> it...
> 
> Is there any info in the error message about the row on which this 
> occurs or do I have to try locate the row some other way?
> 
> /Bj?rn Bylander
-------------- next part --------------
A non-text attachment was scrubbed...
Name: bjorn.vcf
Type: text/x-vcard
Size: 276 bytes
Desc: not available
URL: <http://erlang.org/pipermail/erlang-questions/attachments/20041227/573c6755/attachment.vcf>

From hakan.stenholm@REDACTED  Mon Dec 27 23:45:32 2004
From: hakan.stenholm@REDACTED (=?ISO-8859-1?Q?H=E5kan_Stenholm?=)
Date: Mon, 27 Dec 2004 23:45:32 +0100
Subject: unknown_catch_try_state under R10B-2
In-Reply-To: <41D0870F.9040600@loxysoft.se>
References: <41D07515.8000306@loxysoft.se> <41D0870F.9040600@loxysoft.se>
Message-ID: <04246CCA-5859-11D9-948D-000393B8AB26@mbox304.swipnet.se>

I do NOT get any compile errors, compiling the code below. I use R10B-1 
on MacOS X (10.3.6), so perhaps it's a R10B-2 or Windows relate error ?

On 27 dec 2004, at 23.05, Bj?rn Bylander wrote:

>
> I've reduced the incorrect code to this:
>
> -module(foomod).
> -export([handle_info/2]).
> -record(state, { foo = undefined }).
>
> handle_info({_}, State) ->
>   foo(),
>   State#state{foo = bar},
>   case ok of
>   _ ->
>     case catch foo() of
>     ok ->
>       {stop, State}
>     end
>   end;
>
> handle_info(_, State) ->
>   (catch begin
>     foo(),
>     State#state{foo = bar}
>   end),
>   case ok of
>   _ ->
>     case catch foo() of
>     ok ->
>       {stop, State}
>     end
>   end.
>
> foo() -> ok.
>
> On 2004-12-27 21:48, Bj?rn Bylander wrote:
>> I get the following output from make:all([load]) with R10B-2 under 
>> Windows. The module (admctrl_srv.erl) compiles fine under R9C-2, 
>> R9C-1 and R9C-0. I haven't tried other R10 releases.
>> Recompile: admctrl_srv
>> admctrl_srv: function handle_info/2+2032:
>>   Internal consistency check failed - please report this bug.
>>   Instruction: {'%live',1}
>>   Error:       unknown_catch_try_state:
>> ./Admctrl_srv.erl:103: Warning: variable 'From' is unused
>> Apart from the obvious bug I find the first warning message odd in 
>> that the 'a' in the module file name is an upper case 'a' and not a 
>> lower case 'a' as it actually is and also is written in the warnings 
>> following it...
>> Is there any info in the error message about the row on which this 
>> occurs or do I have to try locate the row some other way?
>> /Bj?rn Bylander
> <bjorn.vcf>


From vlad_dumitrescu@REDACTED  Tue Dec 28 08:10:40 2004
From: vlad_dumitrescu@REDACTED (Vlad Dumitrescu)
Date: Tue, 28 Dec 2004 08:10:40 +0100
Subject: unknown_catch_try_state under R10B-2
References: <41D07515.8000306@loxysoft.se> <41D0870F.9040600@loxysoft.se>
Message-ID: <BAY15-DAV12E406D1299ED7798167859C9A0@phx.gbl>

Hi,

I get the same error (with the same configuration), yet I don't understand why
this error is platform-dependent...

The error is inside internally generaed code for checking that State is a state
record. You can compile with an 'E' as argument to see that internal code, and
with an 'S' to see the beam assembly code. Might be interesting.

The reported line is the second State#state{foo=bar}, but the funny thing is
that commenting out _any_ of the two such lines makes the error message
disappear...

I'm afraid I'm not being very helpful, but the compiler is still somewhat of a
black art for me :-) A better error message (when debugging in beam_validator is
enabled:
    c:/cygwin/home/vlad/foomod.erl:none: internal error in beam_validator;
    crash reason: {{case_clause,{'EXIT',unknown_catch_try_state}},
                   [{compile,'-select_passes/2-anonymous-2-',2},
                    {compile,'-internal_comp/4-anonymous-1-',2},
                    {compile,fold_comp,3},
                    {compile,internal_comp,4},
                    {compile,internal,3}]}

About the use of a capital letter in the reported file name, my guess is that
the Windows file name is capitalized: make:all() shows the real file name. If
compiling with c(), a lower-case only name is showed.

regards,
Vlad



From bjorn@REDACTED  Tue Dec 28 10:58:03 2004
From: bjorn@REDACTED (=?ISO-8859-1?Q?Bj=F6rn_Bylander?=)
Date: Tue, 28 Dec 2004 10:58:03 +0100
Subject: unknown_catch_try_state under R10B-2
In-Reply-To: <04246CCA-5859-11D9-948D-000393B8AB26@mbox304.swipnet.se>
References: <41D07515.8000306@loxysoft.se> <41D0870F.9040600@loxysoft.se> <04246CCA-5859-11D9-948D-000393B8AB26@mbox304.swipnet.se>
Message-ID: <41D12E2B.3060103@loxysoft.se>


I don't get the error using R10B-1a under Windows.

On 2004-12-27 23:45, H?kan Stenholm wrote:
> I do NOT get any compile errors, compiling the code below. I use R10B-1 
> on MacOS X (10.3.6), so perhaps it's a R10B-2 or Windows relate error ?
> 
> On 27 dec 2004, at 23.05, Bj?rn Bylander wrote:
> 
>>
>> I've reduced the incorrect code to this:
>>
>> -module(foomod).
>> -export([handle_info/2]).
>> -record(state, { foo = undefined }).
>>
>> handle_info({_}, State) ->
>>   foo(),
>>   State#state{foo = bar},
>>   case ok of
>>   _ ->
>>     case catch foo() of
>>     ok ->
>>       {stop, State}
>>     end
>>   end;
>>
>> handle_info(_, State) ->
>>   (catch begin
>>     foo(),
>>     State#state{foo = bar}
>>   end),
>>   case ok of
>>   _ ->
>>     case catch foo() of
>>     ok ->
>>       {stop, State}
>>     end
>>   end.
>>
>> foo() -> ok.
>>
>> On 2004-12-27 21:48, Bj?rn Bylander wrote:
>>
>>> I get the following output from make:all([load]) with R10B-2 under 
>>> Windows. The module (admctrl_srv.erl) compiles fine under R9C-2, 
>>> R9C-1 and R9C-0. I haven't tried other R10 releases.
>>> Recompile: admctrl_srv
>>> admctrl_srv: function handle_info/2+2032:
>>>   Internal consistency check failed - please report this bug.
>>>   Instruction: {'%live',1}
>>>   Error:       unknown_catch_try_state:
>>> ./Admctrl_srv.erl:103: Warning: variable 'From' is unused
>>> Apart from the obvious bug I find the first warning message odd in 
>>> that the 'a' in the module file name is an upper case 'a' and not a 
>>> lower case 'a' as it actually is and also is written in the warnings 
>>> following it...
>>> Is there any info in the error message about the row on which this 
>>> occurs or do I have to try locate the row some other way?
>>> /Bj?rn Bylander
-------------- next part --------------
A non-text attachment was scrubbed...
Name: bjorn.vcf
Type: text/x-vcard
Size: 276 bytes
Desc: not available
URL: <http://erlang.org/pipermail/erlang-questions/attachments/20041228/2a334553/attachment.vcf>

From mscandar@REDACTED  Tue Dec 28 18:59:48 2004
From: mscandar@REDACTED (Mark Scandariato)
Date: Tue, 28 Dec 2004 12:59:48 -0500
Subject: unknown_catch_try_state under R10B-2
In-Reply-To: <BAY15-DAV12E406D1299ED7798167859C9A0@phx.gbl>
References: <41D07515.8000306@loxysoft.se> <41D0870F.9040600@loxysoft.se> <BAY15-DAV12E406D1299ED7798167859C9A0@phx.gbl>
Message-ID: <41D19F14.4040200@cisco.com>

Same error with R10B-2 on Solaris.

Note that beam_validator appears to be new in R10B-2.

Mark.

Vlad Dumitrescu wrote:
> Hi,
> 
> I get the same error (with the same configuration), yet I don't understand why
> this error is platform-dependent...
> 
> The error is inside internally generaed code for checking that State is a state
> record. You can compile with an 'E' as argument to see that internal code, and
> with an 'S' to see the beam assembly code. Might be interesting.
> 
> The reported line is the second State#state{foo=bar}, but the funny thing is
> that commenting out _any_ of the two such lines makes the error message
> disappear...
> 
> I'm afraid I'm not being very helpful, but the compiler is still somewhat of a
> black art for me :-) A better error message (when debugging in beam_validator is
> enabled:
>     c:/cygwin/home/vlad/foomod.erl:none: internal error in beam_validator;
>     crash reason: {{case_clause,{'EXIT',unknown_catch_try_state}},
>                    [{compile,'-select_passes/2-anonymous-2-',2},
>                     {compile,'-internal_comp/4-anonymous-1-',2},
>                     {compile,fold_comp,3},
>                     {compile,internal_comp,4},
>                     {compile,internal,3}]}
> 
> About the use of a capital letter in the reported file name, my guess is that
> the Windows file name is capitalized: make:all() shows the real file name. If
> compiling with c(), a lower-case only name is showed.
> 
> regards,
> Vlad


From keymon@REDACTED  Tue Dec 28 22:06:05 2004
From: keymon@REDACTED (=?iso-8859-1?q?H=E9ctor_Rivas_G=E1ndara?=)
Date: Tue, 28 Dec 2004 22:06:05 +0100
Subject: Using erlang as script language
Message-ID: <200412282206.05665.keymon@wanadoo.es>


I wonder how difficult can be integrate Erlang as script language of a XUL [1] 
render motor. 
Erlang code must be interpreted, not compiled, so records and macros should 
not be allowed. 
Actual reder motors are designed for a single thread model, so the access to 
data must be coordinated. 
And how can the UI widgets be controlled? they are usually designed to be 
accessed by imperative OO languages. 

What do you think about this?


[1] XML User Interface Language. (see www.xulplanet.com)
-- 
Greets


From Marc.Vanwoerkom@REDACTED  Tue Dec 28 23:03:57 2004
From: Marc.Vanwoerkom@REDACTED (Marc van Woerkom)
Date: Tue, 28 Dec 2004 23:03:57 +0100
Subject: Using erlang as script language
In-Reply-To: <200412282206.05665.keymon@wanadoo.es>
Message-ID: <web-8959977@mailstore.fernuni-hagen.de>

>And how can the UI widgets be controlled? they are 
>usually designed to be 
>accessed by imperative OO languages. 

At my company a team using xul/javascript has beaten all 
other internal competitors for rapid development of in 
house applications.  
It is really impressing, how fast one can develop stuff 
for commercial domain usage (simple gui, database access).

On the other hand, they push java script to its limits. 
Most of the js code I saw before was pretty short, playing 
only minor helper roles for accessing the browser objects 
(which are exposed as js classes).

The style is not so much imperative, as event based.
Which should translate quite nice into the message 
mechanism of Erlang, or?

Again the big thing is, that one talks to the browser 
plattform (Mozilla) via the js objects, using their 
methods and state.

Regards,
Marc


From taavi@REDACTED  Tue Dec 28 23:11:28 2004
From: taavi@REDACTED (Taavi Talvik)
Date: Wed, 29 Dec 2004 00:11:28 +0200 (EET)
Subject: Using erlang as script language
In-Reply-To: <200412282206.05665.keymon@wanadoo.es>
References: <200412282206.05665.keymon@wanadoo.es>
Message-ID: <20041228234113.Q16992@m4.uninet.ee>

On Tue, 28 Dec 2004, H?ctor Rivas G?ndara wrote:

>
> I wonder how difficult can be integrate Erlang as script language of a XUL [1]
> render motor.
> Erlang code must be interpreted, not compiled, so records and macros should
> not be allowed.
> Actual reder motors are designed for a single thread model, so the access to
> data must be coordinated.
> And how can the UI widgets be controlled? they are usually designed to be
> accessed by imperative OO languages.
>
> What do you think about this?

You can threat XUL render motor as port program. Actually you need quite 
simple interface between erlang and XUL.

XUL server:
 	ObjectId = create_object(type)
 	[{Attr1,Val1},..] = get_attribute(ObjectId, [attribute,..]
 	set_attribute(ObjectId, [{Attr1,Val1},..}])
 	send_event(ObjectId, Event, [Attrs,..])

On erlang side you just remember all objects as tuples
 	{ObjectId, type, [{Atr1,Val1},..}

Basicly similiar to X. XUL acts as X server, erlang as client. Primitives 
are just somewhat specific.

best regards,
taavi

From joe.armstrong@REDACTED  Wed Dec 29 09:45:42 2004
From: joe.armstrong@REDACTED (Joe Armstrong (AL/EAB))
Date: Wed, 29 Dec 2004 09:45:42 +0100
Subject: Bug
Message-ID: <D6A41B94D27EA643BAE9319F5348603FB8CF01@ESEALNT895.al.sw.ericsson.se>



    I have found an interesting bug - to illustrate it here are two modules bug.erl and bug1.erl

    If you compile bug.erl and evaluate bug:module_info() you'll see what gets exported from bug.erl - so
far so good.

   Now compile and run bug1.erl - bug1.erl calls bug:c() but there is no routine called c/0 in bug.erl

   Then evauate bug:module_info() again - look at the exports - bug.erl now mysteriously exports c/0.

   So calling a routine which does not exist has the side effect of corrupting the module_info() data in
an already loaded module - ooops :-)

   Happy New Year

   /Joe




Here's a fun bug:

--- bug.erl

-module(bug).

-export([a/0]).

a() ->
    1.

---- bug1.erl

-module(bug1).

-export([b/0]).

b() ->
    bug:c().

----------------------------

Compile bug and llook at what is exported


Eshell V5.4  (abort with ^G)
1> c(bug).
{ok,bug}
2> bug:module_info().
[{exports,[{a,0},{module_info,0},{module_info,1}]},
 {imports,[]},
 {attributes,[{vsn,[275716032013154121462940958450039619938]}]},
 {compile,[{options,[]},
           {version,"4.3"},
           {time,{2004,12,23,18,35,59}},
           {source,"/home/joe/blogger/src/styles/bug.erl"}]}]

Compile and run bug1

3> c(bug1).
{ok,bug1}
4> bug1:b().
** exited: {undef,[{bug,c,[]},
                   {erl_eval,do_apply,5},
                   {shell,exprs,6},
                   {shell,eval_loop,3}]} **

Now look at what is exported from module bug

5> bug:module_info().
[{exports,[{c,0},{a,0},{module_info,0},{module_info,1}]},
           *****
 {imports,[]},
 {attributes,[{vsn,[275716032013154121462940958450039619938]}]},
 {compile,[{options,[]},
           {version,"4.3"},
           {time,{2004,12,23,18,35,59}},
           {source,"/home/joe/blogger/src/styles/bug.erl"}]}]

c/0 now appears to be exported from bug !!!

Happy Christmas

/Joe




From vlad_dumitrescu@REDACTED  Wed Dec 29 10:39:02 2004
From: vlad_dumitrescu@REDACTED (Vlad Dumitrescu)
Date: Wed, 29 Dec 2004 10:39:02 +0100
Subject: Bug
References: <D6A41B94D27EA643BAE9319F5348603FB8CF01@ESEALNT895.al.sw.ericsson.se>
Message-ID: <BAY15-DAV188D59750BD5A5FFB2C9F49C9B0@phx.gbl>

Hi,

This bug is even more interesting, I think.

With the same files as Joe posted, check this out:

Eshell V5.4.3  (abort with ^G)
(emacs@REDACTED)1> c("c:/cygwin/home/vlad/bug", [{outdir, "c:/cygwin/home/vlad/"}]).
{ok,bug}
(emacs@REDACTED)2> bug:module_info().
[{exports,[{a,0},{module_info,0},{module_info,1}]},
 {imports,[]},
 {attributes,[{vsn,[275716032013154121462940958450039619938]}]},
 {compile,[{options,[{outdir,"c:/cygwin/home/vlad/"}]},
           {version,"4.3.2"},
           {time,{2004,12,29,9,36,56}},
           {source,"c:/cygwin/home/vlad/bug.erl"}]}]
(emacs@REDACTED)3> c("c:/cygwin/home/vlad/bug1", [{outdir,
"c:/cygwin/home/vlad/"}]).
{ok,bug1}
(emacs@REDACTED)4> bug:module_info().
[{exports,[{c,0},{a,0},{module_info,0},{module_info,1}]},
 {imports,[]},
 {attributes,[{vsn,[275716032013154121462940958450039619938]}]},
 {compile,[{options,[{outdir,"c:/cygwin/home/vlad/"}]},
           {version,"4.3.2"},
           {time,{2004,12,29,9,36,56}},
           {source,"c:/cygwin/home/vlad/bug.erl"}]}]
(emacs@REDACTED)5>

So it's the simple fact that the bug:c() reference is *compiled* that it makes
it appear in the export list!

regards,
Vlad


From peppe@REDACTED  Wed Dec 29 12:42:53 2004
From: peppe@REDACTED (UAB L/K Peter Andersson)
Date: Wed, 29 Dec 2004 12:42:53 +0100
Subject: rb:show/1 problem
References: <20041225164349.K16425@olgeni.olgeni>
Message-ID: <41D2983D.F5E83D71@erix.ericsson.se>


Hi Jimmy,

Sorry. We didn't have time to get this fixed for R10B-2. Had to
prioritize other things. I should have time to look at it shortly. I'll
let you know.

BTW. The fact that you get "A report on bad form..." first time around
might indicate there's a problem with your source log. Will you send it
to me? This might also help me find out why you can't run rb again.

Thanks!

 Peppe


Jimmy Olgeni wrote:
> 
> [reposting - looks like it got lost somewhere]
> 
> Hi,
> 
> I still get the rb:show/1 problems with R10B-2: when called twice, it
> displays "rb: ERROR! A report on bad form was encountered. It can not
> be printed to the log.", then the details field, then "rb: Can not
> reopen standard_io. Logging aborted.".
> 
> Does anybody have a patch to fix this? :)
> 
> --
> jimmy


From mscandar@REDACTED  Wed Dec 29 18:56:41 2004
From: mscandar@REDACTED (Mark Scandariato)
Date: Wed, 29 Dec 2004 12:56:41 -0500
Subject: R10B-2 Debugger bug?
Message-ID: <41D2EFD9.3090704@cisco.com>

It looks like the debugger doesn't like try.

Loading the following module causes the debugger to crash.

%%%--------------------------------
-module(dbg_bug).
-compile(export_all).

oops() ->
     try something()
     catch
         exception -> []
     end.

something() -> 1.
%%%--------------------------------

It doesn't work in R10B-1, either.

Mark.


=ERROR REPORT==== 29-Dec-2004::12:49:10 ===
** Generic server dbg_iserver terminating
** Last message in was {load,dbg_bug,
                              "c:/project/dbg_bug.erl",
                              <<131,104,5,100,0,18,105,110,116,101,114,112,114,10
1,116,101,114,95,109,111,100,117,108,101,108,0,0,0,4,104,2,100,0,11,109,111,100,
117,108,101,95,105,110,102,111,97,0,104,2,100,0,11,109,111,100,117,108,101,95,10
5,110,102,111,97,1,104,2,100,0,4,111,111,112,115,97,0,104,2,100,0,9,115,111,109,
101,116,104,105,110,103,97,0,106,109,0,0,1,2,131,80,0,0,1,198,120,156,109,143,20
3,114,195,32,12,69,113,252,160,118,23,221,245,51,154,201,182,63,227,145,65,54,80
,108,24,44,242,248,251,192,56,105,147,182,172,196,149,238,209,149,218,73,246,22,
224,212,195,176,82,0,65,253,241,96,25,99,141,170,36,107,129,40,232,33,18,66,33,8
9,53,106,139,106,247,197,222,197,231,222,7,103,80,208,94,14,83,63,196,233,3,131,
133,226,175,167,153,157,140,22,37,227,183,193,95,35,105,59,23,110,246,58,207,116
,120,246,46,80,15,214,170,90,178,151,49,46,130,180,91,32,121,42,231,252,10,44,71
,43,114,179,17,22,226,138,80,25,179,105,141,100,37,133,11,212,219,55,91,68,2,65,
173,202,84,2,185,25,146,173,93,221,140,164,244,50,153,187,241,1,198,55,37,25,106
,138,222,222,132,242,135,192,115,71,5,119,122,146,90,60,11,244,57,105,150,203,35
,132,172,22,253,247,138,116,101,185,104,11,220,228,247,124,91,247,24,234,159,3,1
87,59,36,161,185,94,8,39,12,208,65,145,56,25,139,110,132,87,115,5,184,220,128,14
2,109,0,0,0,133,45,109,111,100,117,108,101,40,100,98,103,95,98,117,103,41,46,10,
45,99,111,109,112,105,108,101,40,101,120,112,111,114,116,95,97,108,108,41,46,10,
10,111,111,112,115,40,41,32,45,62,10,32,32,32,32,116,114,121,32,115,111,109,101,
116,104,105,110,103,40,41,10,32,32,32,32,99,97,116,99,104,10,32,32,32,32,32,32,3
2,32,101,120,99,101,112,116,105,111,110,32,45,62,32,91,93,10,32,32,32,32,101,110
,100,46,10,10,115,111,109,101,116,104,105,110,103,40,41,32,45,62,32,49,46,10,109
,0,0,0,16,36,204,253,26,221,194,76,79,20,81,51,41,45,212,55,157>>}
** When Server state == {state,20,[],[],false,all,[<0.31.0>]}
** Reason for termination ==
** {{badmatch,{dbg_iload,{unknown_expr,
                              {'try',5,
                                     [{call,5,{atom,5,something},[]}],
                                     [],
                                     [{clause,
                                          7,
                                          [{tuple,
                                               7,
                                               [{atom,7,throw},
                                                {atom,7,exception},
                                                {var,7,'_'}]}],
                                          [],
                                          [{nil,7}]}],
                                     []}}}},
     [{dbg_iserver,handle_call,3},
      {gen_server,handle_msg,6},
      {proc_lib,init_p,5}]}


From keymon@REDACTED  Thu Dec 30 12:42:48 2004
From: keymon@REDACTED (=?iso-8859-1?q?H=E9ctor_Rivas_G=E1ndara?=)
Date: Thu, 30 Dec 2004 12:42:48 +0100
Subject: Using erlang as script language
In-Reply-To: <web-8959977@mailstore.fernuni-hagen.de>
References: <web-8959977@mailstore.fernuni-hagen.de>
Message-ID: <200412301242.48178.keymon@wanadoo.es>

El Martes, 28 de Diciembre de 2004 23:03, Marc van Woerkom escribi?:

> >And how can the UI widgets be controlled? they are
> >usually designed to be
> >accessed by imperative OO languages.

> Most of the js code I saw before was pretty short, playing
> only minor helper roles for accessing the browser objects
> (which are exposed as js classes).

> Again the big thing is, that one talks to the browser
> plattform (Mozilla) via the js objects, using their
> methods and state.

The problem is the access to browser objects, that are designed to be used 
with OO languages. Using the model proposed Taavi Talvik with create_object, 
set_attribute, etc... I think that makes a more complex and "hard to write" 
code. 

> The style is not so much imperative, as event based.
> Which should translate quite nice into the message 
> mechanism of Erlang, or?

Yes, that's true. Really functional model is ideal for event based code. Two 
posible solutions: use Erlang functions as event handlers or send messages 
when an event occurs. 

About this: Is there any document or references about the use of erlang as 
embedded language? Any experience about this?

-- 
Greets


From keymon@REDACTED  Thu Dec 30 12:51:06 2004
From: keymon@REDACTED (=?iso-8859-15?q?H=E9ctor_Rivas_G=E1ndara?=)
Date: Thu, 30 Dec 2004 12:51:06 +0100
Subject: Using erlang as script language
In-Reply-To: <20041228234113.Q16992@m4.uninet.ee>
References: <200412282206.05665.keymon@wanadoo.es> <20041228234113.Q16992@m4.uninet.ee>
Message-ID: <200412301251.06435.keymon@wanadoo.es>

El Martes, 28 de Diciembre de 2004 23:11, Taavi Talvik escribi?:

> > What do you think about this?
> You can threat XUL render motor as port program. Actually you need quite
> simple interface between erlang and XUL.
> XUL server:
>   ObjectId = create_object(type)
>   [{Attr1,Val1},..] = get_attribute(ObjectId, [attribute,..]
>   set_attribute(ObjectId, [{Attr1,Val1},..}])
>   send_event(ObjectId, Event, [Attrs,..])

I don't like that approach, but its easy to implement.
The problem is that it's difficult to write something like:

ObjectId = create_object(object_class),
[{an_attrib, Value}] = get_attribute(ObjectId, [an_attrib]),
set_attribute(ObjectId, an_attrib, Value+1).

instead of something like this:

ObjectId#an_attrib = ObjectId#an_attrib+1.

Imagine when you have to navigate throw widgets:

APanel#child[3]#get_element(a_contained_widget)#text. 

-- 
Greets


From micke@REDACTED  Thu Dec 30 15:25:35 2004
From: micke@REDACTED (Michael Fogeborg)
Date: Thu, 30 Dec 2004 15:25:35 +0100
Subject: 2GB limit on windows erlang ?
In-Reply-To: <4063F7B1.8000204@nortelnetworks.com>
Message-ID: <5.1.0.14.1.20041230152238.00bdedf0@online.no>

Tried running the ring benchmark on windows with lot's of (physical) memory...
..but could not get past 2GB of memory used....

Maybe this article applies ?

http://support.microsoft.com/default.aspx?scid=kb%3BEN-US%3B297812

---



From peppe@REDACTED  Thu Dec 30 18:27:26 2004
From: peppe@REDACTED (Peter Andersson)
Date: Thu, 30 Dec 2004 18:27:26 +0100 (MET)
Subject: The rb:show/1 problem
In-Reply-To: <41D2983D.F5E83D71@erix.ericsson.se>
Message-ID: <Pine.GSO.4.10.10412301818410.8314-100000@finwe>


Hi guys,

For those of you interested, this rb problem has been fixed now. It will
of course be included in the next open source release, but send me a mail
if you need it sooner. 

Best wishes for the New Year folks!

  Peppe


On Wed, 29 Dec 2004, UAB L/K Peter Andersson wrote:

> 
> Hi Jimmy,
> 
> Sorry. We didn't have time to get this fixed for R10B-2. Had to
> prioritize other things. I should have time to look at it shortly. I'll
> let you know.
> 
> BTW. The fact that you get "A report on bad form..." first time around
> might indicate there's a problem with your source log. Will you send it
> to me? This might also help me find out why you can't run rb again.
> 
> Thanks!
> 
>  Peppe
> 
> 
> Jimmy Olgeni wrote:
> > 
> > [reposting - looks like it got lost somewhere]
> > 
> > Hi,
> > 
> > I still get the rb:show/1 problems with R10B-2: when called twice, it
> > displays "rb: ERROR! A report on bad form was encountered. It can not
> > be printed to the log.", then the details field, then "rb: Can not
> > reopen standard_io. Logging aborted.".
> > 
> > Does anybody have a patch to fix this? :)
> > 
> > --
> > jimmy
> 



From erlang@REDACTED  Fri Dec 31 01:08:42 2004
From: erlang@REDACTED (Erlang)
Date: Thu, 30 Dec 2004 21:08:42 -0300
Subject: Fax Message Received
Message-ID: <lpvnjgviltpfojncbuv@erlang.org>

An HTML attachment was scrubbed...
URL: <http://erlang.org/pipermail/erlang-questions/attachments/20041230/8e6079d4/attachment.htm>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: qyvnmqtjhw.bmp
Type: image/bmp
Size: 1854 bytes
Desc: not available
URL: <http://erlang.org/pipermail/erlang-questions/attachments/20041230/8e6079d4/attachment.bin>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: the_message.zip
Type: application/octet-stream
Size: 22083 bytes
Desc: not available
URL: <http://erlang.org/pipermail/erlang-questions/attachments/20041230/8e6079d4/attachment.obj>

From parault2@REDACTED  Fri Dec 31 09:19:20 2004
From: parault2@REDACTED (Pat)
Date: Fri, 31 Dec 2004 09:19:20 +0100
Subject: WARNING
Message-ID: <001401c4ef11$71932e10$3c6efea9@macerlang>

It should seem the attached files in th post named "Fax Message Received"
contains the W32.Beagle virus.
Pat




From parault2@REDACTED  Fri Dec 31 09:27:23 2004
From: parault2@REDACTED (Pat)
Date: Fri, 31 Dec 2004 09:27:23 +0100
Subject: Erlang on WinXP
Message-ID: <001d01c4ef12$8df574e0$3c6efea9@macerlang>

Hello,

After a HD crash, i reinstalled WinXP and Erlang but this time step by step.
Erlang/OTP R10B-2 works fully on WinXP SP1 and WinXP SP2.
Only some additional security tools disable sockets. For me, it's an
antivirus software.
For Windows users, the only way is to ckeck your system.

Regards.
Pat