[erlang-patches] DTrace patch, review draft #3

[Scott Lystig Fritchie]

Henrik Nord wrote a while ago, while I was working on another project:

hn> Been a while sens we got any news for this.
hn> The last branch you submitted, this draft #3 I think it was, did not 
hn> compile, and you where in contact with Bj<F6>rn and or Bj<F6>rn-Egil. Then

Hi, Henrik.  That branch worked fine for me.  :-)

hn> there was a release here and we did not really think about this for a
hn> while.

Ja, that branch didn't work so well on the R15B release.  There were
a couple of not-so-nice conflicts, due to a late addition of an Opts
property list to file:read_file_info() and a couple other file.erl
functions.

hn> So now im looking at my branches and there are no Dtrace here.

Well, nothing new in a while.  Sorry about that.  I've been working on
other stuff for the past while....

hn> Do you need assistance from us with something? there was talk about 
hn> loader support?

Last I heard, there was going to be some hackery done by the OTP team.
That hackery would load one of two versions of prim_file.beam, depending
on whether or not DTrace support was available.

I didn't hear anything about that hackery being done, or started, so I
figured y'all were busy working on other stuff.  So I didn't spend much
time on fixing the merge conflict.  To be honest, except for you, nobody
elseuntil today (09 Feb 2012) had asked me about an update for DTrace and
R15B.  {sigh}

hn> Let me know what version of the branch you want us to test and how/if we 
hn> may be of assistance so that we can add that to our plans.

So, I got blessed time from $DAYJOB to finish work on the DTrace stuff.
I've pushed two new branches to git://github.com/slfritchie/otp.git this
evening:

    * dtrace-review4-OTP_R15B, which is DTrace work rebased & conflicts
      fixed, based upon the "OTP_R15B" tag from the OTP team's GitHub
      repo.

    * dtrace-review4-upstream-maint, which is the same rebased & fixed
      stuff, based upon the 09 Feb 2012 "maint" branch of the OTP repo.

I haven't run all of the test suite yet, only the "kernel" portion.  And
only on OS X.  Using code based on the "OTP_R15B" tag, I see 6 test
failures.  Using the "dtrace-review4-upstream-maint" branch, I see 7 test
failures (code_SUITE, big_boot_embedded fails with a timeout).

So, I can do more testing on OS X, Linux, and Solaris boxes and letting
the whole test suite run overnight.

It'd be nice to know if/when y'all were thinking of doing the code loader
hacks.  I also haven't been paying much attention to any rumors of if/when
R15B01 might be frozen.

-Scott