361940 - CVS checkout finish: Mon Nov 27 02:37:21 EST 2006 compiles but will not run

Status: RESOLVED INVALID

(SeaMonkey :: General, defect)

Description

[Rinaldi J. Montessi]

User-Agent:       Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.8.0.7) Gecko/20060909 Firefox/1.5.0.7
Build Identifier: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.8.0.7) Gecko/20060909 Firefox/1.5.0.7

Program ./seamonkey-bin (pid = 2824) received signal 11.
Stack:
__kernel_sigreturn+0x00000000 [ +0x00000420]
nsTHashtable<nsBaseHashtableET<nsVoidPtrHashKey, nsRefPtr<nsThread> > >::PutEntry(void const*)+0x0000005D [./libxpcom_core.so +0x000FF5D9]
nsBaseHashtable<nsVoidPtrHashKey, nsRefPtr<nsThread>, nsThread*>::Put(void const*, nsThread*)+0x0000001E [./libxpcom_core.so +0x000FF200]
nsThreadManager::RegisterCurrentThread(nsThread*)+0x00000088 [./libxpcom_core.so +0x000FEB4C]
nsThread::InitCurrentThread()+0x0000005C [./libxpcom_core.so +0x000FC80A]
nsThreadManager::Init()+0x000000BB [./libxpcom_core.so +0x000FE8FB]
NS_InitXPCOM3_P+0x00000043 [./libxpcom_core.so +0x000A1209]
UNKNOWN [./seamonkey-bin +0x00009D14]
__libc_start_main+0x000000D4 [/lib/tls/libc.so.6 +0x00014E14]

Reproducible: Always

Comment 1

[Andrew Schultz]

trunk or branch?

Comment 2

[Rinaldi J. Montessi]

Should be trunk, yes?

CVSROOT=":pserver:[e-mail]:/cvsroot"
cvs checkout mozilla/client.mk
mk_add_options MOZ_CO_PROJECT=suite

Comment 3

[Andrew Schultz]

this isn't a build config bug (at least there's not enough info to say it is)

What was the last CVS timestamp that started up OK?

Comment 4

[Rinaldi J. Montessi]

I build every 7 - 10 days so a last working timestamp would be meaningless.

What I've done since reporting is nuked my moz tree and object_dir, pulled a fresh copy of the source (Tue Nov 28 09:24:19 EST 2006), simplified my ~/.mozconfig 

mk_add_options MOZ_CO_PROJECT=suite 
mk_add_options MOZ_OBJDIR=/home/Rinaldi/tmp/Mozilla/suite_objdir
mk_add_options MOZ_MAKE_FLAGS=-j4
ac_add_options --enable-application=suite 
ac_add_options --disable-tests
ac_add_options --enable-debug
ac_add_options --enable-crypto 
ac_add_options --enable-xft

and rebuilt.  Still no joy.  I didn't run a diff, but it looks the same as before:

Program ./seamonkey-bin (pid = 22803) received signal 11.
Stack:
__kernel_sigreturn+0x00000000 [ +0x00000420]
nsTHashtable<nsBaseHashtableET<nsVoidPtrHashKey, nsRefPtr<nsThread> > >::PutEntry(void const*)+0x0000005D [./libxpcom_core.so +0x000FF5B9]
nsBaseHashtable<nsVoidPtrHashKey, nsRefPtr<nsThread>, nsThread*>::Put(void const*, nsThread*)+0x0000001E [./libxpcom_core.so +0x000FF1E0]
nsThreadManager::RegisterCurrentThread(nsThread*)+0x00000088 [./libxpcom_core.so +0x000FEB2C]
nsThread::InitCurrentThread()+0x0000005C [./libxpcom_core.so +0x000FC7EA]
nsThreadManager::Init()+0x000000BB [./libxpcom_core.so +0x000FE8DB]
NS_InitXPCOM3_P+0x00000043 [./libxpcom_core.so +0x000A11E9]
UNKNOWN [./seamonkey-bin +0x00009D14]
__libc_start_main+0x000000D4 [/lib/tls/libc.so.6 +0x00014E14]

If I'm the only one with this problem I'm willing to accept it's a PEBCAK situation, but I'd like at least a "works for me" from someone before falling on my sword :-)

Comment 5

[Andrew Schultz]

It "works for me" and all of these:
http://tinderbox.mozilla.org/showbuilds.cgi?tree=SeaMonkey
But your .mozconfig should work.

Did anything related to your toolchain change between the last good and first bad build?

Best plan would be to pull by date between the last known good build and the first known bad build.  Edit client.mk and add a line like

MOZ_CO_DATE=Nov 20, 2006

and then do "make -f client.mk"

Do a binary search on dates to narrow down the regression window.

Comment 6

[Rinaldi J. Montessi]

****.  Upgraded to freetype-2.1.10.  I'll put the blame their until further review.  I know it's been a problem in the past.

Comment 7

[Rinaldi J. Montessi]

I finally had a chance to track this down.  Compiler error.  Upgrading to gcc (GCC) 3.4.6 fixed the problem.