[erlang-bugs] Internal error in sys_core_fold when matching on map is guaranteed to fail

[Björn-Egil Dahlberg]

Logging.

I had a private discussion with Anthony. The suggested solution is not the
way to solve it.
In fact, the same problem occurs with binaries as well.


2014-03-23 12:29 GMT+01:00 Anthony Ramine <n.oxyde@REDACTED>:

> https://github.com/erlang/otp/pull/306
>
> --
> Anthony Ramine
>
> Le 18 mars 2014 à 21:45, Björn-Egil Dahlberg <wallentin.dahlberg@REDACTED>
> a écrit :
>
> > In the future Maps should be considered as data but since it seemed like
> I needed a crowbar and some soap to whack it into the cerl opt structure
> "nicely" I decided against it for now, fearing more destruction than gains
> .. also I wanted literals in the master branch. =) The omission of maps as
> data will currently omit some core optimizations .. which is also totally
> ok .. for now.
> >
> > is_data(#c_literal{val=V}) when not is_map(V) ->
> >     true;
> >
> > Looks reasonable to me at this point. Give me a shout if you see/find
> any flaws with this plan.
> >
> > // Björn-Egil
> >
> >
> > 2014-03-18 19:58 GMT+01:00 Anthony Ramine <n.oxyde@REDACTED>:
> > Hello,
> >
> > Are map literals supposed to be considered data, as per cerl:is_data/1?
> can I currently fix this by changing
> >
> > is_data(#c_literal{}) ->
> >     true;
> >
> > to
> >
> > is_data(#c_literal{val=V}) when not is_map(V) ->
> >     true;
> >
> > or not?
> >
> > Regards,
> >
> > --
> > Anthony Ramine
> >
> > Le 18 mars 2014 à 18:53, Anthony Ramine <n.oxyde@REDACTED> a écrit :
> >
> > > Cool, will look at it.
> > >
> > > --
> > > Anthony Ramine
> > >
> > > Le 18 mars 2014 à 17:50, Ulf Norell <ulf.norell@REDACTED> a écrit :
> > >
> > >> That fixed the case when the map doesn't match but I get the same
> error when it's the map that matches and other patterns that don't:
> > >>
> > >> -module(bug).
> > >> f() ->
> > >>  case #{} of
> > >>    #{} -> 0;
> > >>    a   -> 1
> > >>  end.
> > >>
> > >> This is on master from today
> (aab1afa8a63081f2c30f83ebe8b6caaac979795c).
> > >>
> > >> / Ulf
> > >>
> > >>
> > >> On Tue, Mar 4, 2014 at 11:38 PM, Anthony Ramine <n.oxyde@REDACTED>
> wrote:
> > >> https://github.com/erlang/otp/pull/267
> > >>
> > >> --
> > >> Anthony Ramine
> > >>
> > >> Le 4 mars 2014 à 09:45, Ulf Norell <ulf.norell@REDACTED> a écrit :
> > >>
> > >>> This one crashes in a different place but likely for the same reason.
> > >>>
> > >>> -module(bug).
> > >>> f() ->
> > >>>  case true of
> > >>>    #{}  -> error;
> > >>>    true -> ok
> > >>>  end.
> > >>>
> > >>> $ erl
> > >>> Erlang/OTP 17 [RELEASE CANDIDATE 2] [erts-6.0] [source-a74e66a]
> [64-bit] [smp:8:8] [async-threads:10] [hipe] [kernel-poll:false]
> > >>>
> > >>> Eshell V6.0  (abort with ^G)
> > >>> 1> c(bug).
> > >>> Function: f/0
> > >>> bug.erl: internal error in core_fold_module;
> > >>> crash reason: function_clause
> > >>>
> > >>>  in function  sys_core_fold:opt_bool_clauses/3
> > >>>     called as sys_core_fold:opt_bool_clauses([],true,false)
> > >>>  in call from sys_core_fold:opt_bool_clauses/3 (sys_core_fold.erl,
> line 1815)
> > >>>  in call from sys_core_fold:opt_bool_clauses/1 (sys_core_fold.erl,
> line 1790)
> > >>>  in call from sys_core_fold:opt_bool_case/1 (sys_core_fold.erl, line
> 1779)
> > >>>  in call from sys_core_fold:expr/3 (sys_core_fold.erl, line 347)
> > >>>  in call from sys_core_fold:opt_simple_let_1/4 (sys_core_fold.erl,
> line 2562)
> > >>>  in call from sys_core_fold:expr/3 (sys_core_fold.erl, line 355)
> > >>>  in call from sys_core_fold:expr/3 (sys_core_fold.erl, line 279)
> > >>> bug.erl:2: Warning: function f/0 is unused
> > >>> error
> > >>>
> > >>> / Ulf
> > >>>
> > >>>
> > >>>
> > >>> On Tue, Mar 4, 2014 at 9:42 AM, Ulf Norell <ulf.norell@REDACTED>
> wrote:
> > >>> -module(bug).
> > >>> f() ->
> > >>>  case maybe_map of
> > >>>    #{}     -> ok;
> > >>>    not_map -> error
> > >>>  end.
> > >>>
> > >>> $ erl
> > >>> Erlang/OTP 17 [RELEASE CANDIDATE 2] [erts-6.0] [source-a74e66a]
> [64-bit] [smp:8:8] [async-threads:10] [hipe] [kernel-poll:false]
> > >>>
> > >>> Eshell V6.0  (abort with ^G)
> > >>> 1> c(bug).
> > >>> Function: f/0
> > >>> bug.erl: internal error in core_fold_module;
> > >>> crash reason: {case_clause,{none}}
> > >>>
> > >>>  in function  sys_core_fold:will_match/2 (sys_core_fold.erl, line
> 1748)
> > >>>  in call from sys_core_fold:clauses/5 (sys_core_fold.erl, line 1701)
> > >>>  in call from sys_core_fold:clauses/5 (sys_core_fold.erl, line 1718)
> > >>>  in call from sys_core_fold:expr/3 (sys_core_fold.erl, line 352)
> > >>>  in call from sys_core_fold:clause/4 (sys_core_fold.erl, line 1481)
> > >>>  in call from sys_core_fold:clauses/5 (sys_core_fold.erl, line 1699)
> > >>>  in call from sys_core_fold:expr/3 (sys_core_fold.erl, line 352)
> > >>>  in call from sys_core_fold:expr/3 (sys_core_fold.erl, line 279)
> > >>> bug.erl:2: Warning: function f/0 is unused
> > >>> error
> > >>>
> > >>> / Ulf
> > >>>
> > >>>
> > >>> _______________________________________________
> > >>> erlang-bugs mailing list
> > >>> erlang-bugs@REDACTED
> > >>> http://erlang.org/mailman/listinfo/erlang-bugs
> > >>
> > >> _______________________________________________
> > >> erlang-bugs mailing list
> > >> erlang-bugs@REDACTED
> > >> http://erlang.org/mailman/listinfo/erlang-bugs
> > >>
> > >
> >
> > _______________________________________________
> > erlang-bugs mailing list
> > erlang-bugs@REDACTED
> > http://erlang.org/mailman/listinfo/erlang-bugs
> >
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://erlang.org/pipermail/erlang-bugs/attachments/20140324/aad6d87d/attachment.htm>