trunk not building on OSX

trunk not building on OSX

[Randal L. Schwartz]

Writing LC_LOAD_DYLIB     command
Writing LC_LOAD_DYLIB     command
Writing LC_LOAD_DYLIB     command
Writing LC_LOAD_DYLIB     command
Writing LC_LOAD_DYLIB     command
Writing LC_FUNCTION_STARTS command
unexec: not enough room for load commands for new __DATA segments
make[1]: *** [emacs] Error 1
make: *** [src] Error 2

I've seen this before, and it sometimes gets fixed if I completely wipe
back to the distro, because some intermediate build product is throwing
it off.  No luck today.

Here's the commit I'm at.  If anyone wants me to try something, let me
know:

Author: Glenn Morris <rgm@gnu.org>
Date:   Mon May 14 09:33:11 2012 -0700

    Make a self-contained ns build able to find its own libexec
    directory
    
    Ref:
    http://lists.gnu.org/archive/html/emacs-devel/2012-05/msg00298.html
    
    * src/nsterm.m (ns_init_paths):
    Put "libexec" before "bin", for the sake of init_callproc_1.


-- 
Randal L. Schwartz - Stonehenge Consulting Services, Inc. - +1 503 777 0095
<merlyn@stonehenge.com> <URL:http://www.stonehenge.com/merlyn/>
Smalltalk/Perl/Unix consulting, Technical writing, Comedy, etc. etc.
See http://methodsandmessages.posterous.com/ for Smalltalk discussion

Re: trunk not building on OSX

[Jan Djärv]

Hello.

I do not have this problem. How do you configure Emacs?

      Jan D.

15 maj 2012 kl. 17:10 skrev merlyn@stonehenge.com (Randal L. Schwartz):

> 
> Writing LC_LOAD_DYLIB     command
> Writing LC_LOAD_DYLIB     command
> Writing LC_LOAD_DYLIB     command
> Writing LC_LOAD_DYLIB     command
> Writing LC_LOAD_DYLIB     command
> Writing LC_FUNCTION_STARTS command
> unexec: not enough room for load commands for new __DATA segments
> make[1]: *** [emacs] Error 1
> make: *** [src] Error 2
> 
> I've seen this before, and it sometimes gets fixed if I completely wipe
> back to the distro, because some intermediate build product is throwing
> it off.  No luck today.
> 
> Here's the commit I'm at.  If anyone wants me to try something, let me
> know:
> 
> Author: Glenn Morris <rgm@gnu.org>
> Date:   Mon May 14 09:33:11 2012 -0700
> 
>    Make a self-contained ns build able to find its own libexec
>    directory
> 
>    Ref:
>    http://lists.gnu.org/archive/html/emacs-devel/2012-05/msg00298.html
> 
>    * src/nsterm.m (ns_init_paths):
>    Put "libexec" before "bin", for the sake of init_callproc_1.
> 
> 
> -- 
> Randal L. Schwartz - Stonehenge Consulting Services, Inc. - +1 503 777 0095
> <merlyn@stonehenge.com> <URL:http://www.stonehenge.com/merlyn/>
> Smalltalk/Perl/Unix consulting, Technical writing, Comedy, etc. etc.
> See http://methodsandmessages.posterous.com/ for Smalltalk discussion
> 

Re: trunk not building on OSX

[Randal L. Schwartz]

>>>>> "Jan" == Jan Djärv <jan.h.d@swipnet.se> writes:

Jan> Hello.
Jan> I do not have this problem. How do you configure Emacs?

This was a weird problem, and I think there's still a problem somehow.

I did a git bisect between the version the day before, and the broken
version.  Each time, I did a "git clean -f -x" which wipes everything
but the commit, and:

  ./autogen.sh
  ./configure --with-ns --without-x --without-dbus \
  --with-jpeg=no --with-gif=no --with-tiff=no --prefix=/tmp/emacs

to configure, followed by "make clean bootstrap install" to build.

Now the weird part.  Every single one I did, worked just fine.
So the bisect told me the "last good commit" was in fact the one before
the origin/master.

On a whim, I built origin/master again... and this time it worked.

It should not have.  I was unable to get it to *fail* as it had earlier,
even though I was building exactly the same way.

Is there some part of the build running in parallel, and I was running
into timing issues?  I've actually seen this error before, but only
sporadically.

-- 
Randal L. Schwartz - Stonehenge Consulting Services, Inc. - +1 503 777 0095
<merlyn@stonehenge.com> <URL:http://www.stonehenge.com/merlyn/>
Smalltalk/Perl/Unix consulting, Technical writing, Comedy, etc. etc.
See http://methodsandmessages.posterous.com/ for Smalltalk discussion

Re: trunk not building on OSX

[Jan Djärv]

Hello.

I had a similar experiece on GNU/Linux about a year ago.  Dumping Emacs would randomly crash with a segmentation violation.  Just running make again usually worked.

Turns out bad RAM was the cause.  Could this be similar?

	Jan D.

16 maj 2012 kl. 20:33 skrev Randal L. Schwartz:

>>>>>> "Jan" == Jan Djärv <jan.h.d@swipnet.se> writes:
> 
> Jan> Hello.
> Jan> I do not have this problem. How do you configure Emacs?
> 
> This was a weird problem, and I think there's still a problem somehow.
> 
> I did a git bisect between the version the day before, and the broken
> version.  Each time, I did a "git clean -f -x" which wipes everything
> but the commit, and:
> 
>  ./autogen.sh
>  ./configure --with-ns --without-x --without-dbus \
>  --with-jpeg=no --with-gif=no --with-tiff=no --prefix=/tmp/emacs
> 
> to configure, followed by "make clean bootstrap install" to build.
> 
> Now the weird part.  Every single one I did, worked just fine.
> So the bisect told me the "last good commit" was in fact the one before
> the origin/master.
> 
> On a whim, I built origin/master again... and this time it worked.
> 
> It should not have.  I was unable to get it to *fail* as it had earlier,
> even though I was building exactly the same way.
> 
> Is there some part of the build running in parallel, and I was running
> into timing issues?  I've actually seen this error before, but only
> sporadically.
> 
> -- 
> Randal L. Schwartz - Stonehenge Consulting Services, Inc. - +1 503 777 0095
> <merlyn@stonehenge.com> <URL:http://www.stonehenge.com/merlyn/>
> Smalltalk/Perl/Unix consulting, Technical writing, Comedy, etc. etc.
> See http://methodsandmessages.posterous.com/ for Smalltalk discussion
> 

Re: trunk not building on OSX

[Randal L. Schwartz]

>>>>> "Jan" == Jan Djärv <jan.h.d@swipnet.se> writes:

Jan> I had a similar experiece on GNU/Linux about a year ago.  Dumping
Jan> Emacs would randomly crash with a segmentation violation.  Just
Jan> running make again usually worked.

Jan> Turns out bad RAM was the cause.  Could this be similar?

Unlikely.  I'd have a lot of other crashes randomly then as well.

-- 
Randal L. Schwartz - Stonehenge Consulting Services, Inc. - +1 503 777 0095
<merlyn@stonehenge.com> <URL:http://www.stonehenge.com/merlyn/>
Smalltalk/Perl/Unix consulting, Technical writing, Comedy, etc. etc.
See http://methodsandmessages.posterous.com/ for Smalltalk discussion