ports/92218: openoffice.org-2.0 build fails with shlibsign core dump

[Mike Durian]

The following reply was made to PR ports/92218; it has been noted by GNATS.

From: Mike Durian <durian at shadetreesoftware.com>
To: bug-followup at freebsd.org
Cc:  
Subject: Re: ports/92218: openoffice.org-2.0 build fails with shlibsign core dump
Date: Sun, 29 Jan 2006 12:16:48 -0700

 I've done a bit of digging.  It looks like the
 work/OOA680_m1/moz/unxfbsd.pro/misc/build/mozilla/configure
 script is failing to detect that -lpthread contains the
 pthread_create() function.  It then finds -lc_r and uses it instead.
 
 The config.log file shows:
 configure:8578: checking for pthread_create in -lpthread
 gcc-ooo -o dummy dummy.c    -I/usr/X11R6/include  -I/usr/X11R6/include 
 -lpthread -lm
 /usr/lib/libpthread.so: undefined reference to `__usleep'
 collect2: ld returned 1 exit status
 configure:8600: checking for pthread_create in -lc_r
 gcc-ooo -o dummy dummy.c    -I/usr/X11R6/include  -I/usr/X11R6/include -lc_r  
 -lm
 
 
 I created the dummy.c file by hand and tried compiling it.  It compiles
 just fine (without the missing __usleep reference) if I use the system
 cc.  It fails when I try gcc-ooo.
 
 > cc -o dummy dummy.c -I/usr/X11R6/include -I/usr/X11R6/include -lpthread -lm
 > gcc-ooo -o dummy dummy.c -I/usr/X11R6/include -I/usr/X11R6/include -lpthread 
 -lm
 /usr/lib/libpthread.so: undefined reference to `__usleep'
 collect2: ld returned 1 exit status
 
 
 My gcc-ooo appears to be relatively old:
 > ls -l /usr/local/bin/gcc-ooo
 -r-xr-xr-x  3 root  wheel  77216 May 10  2005 /usr/local/bin/gcc-ooo
 
 
 It also looks like it is not a required dependancy:
 > pkg_info -W /usr/local/bin/gcc-ooo
 /usr/local/bin/gcc-ooo was installed by package gcc-ooo-3.4.1_2
 > pkg_info -R gcc-ooo-3.4.1_2
 Information for gcc-ooo-3.4.1_2:
 
 
 
 I'm going to un-install lang/gcc-ooo and see if that fixes my problem.
 If gcc-ooo is required by openoffice.org-2.0, then it will get rebuilt.
 
 Assuming a stale gcc-ooo is my problem, I'm not sure what a real fix
 is.  Others might experience this same problem.  I'm thinking the
 gcc-ooo port revision needed a bump due to a change in a system library or
 header file so time ago.
 
 mike