ports/140365: [ PATCH ] firebird20-client coredumps

[Dan Lukes]

>Number:         140365
>Category:       ports
>Synopsis:       [ PATCH ] firebird20-client coredumps
>Confidential:   no
>Severity:       critical
>Priority:       high
>Responsible:    freebsd-ports-bugs
>State:          open
>Quarter:        
>Keywords:       
>Date-Required:
>Class:          sw-bug
>Submitter-Id:   current-users
>Arrival-Date:   Sat Nov 07 18:20:01 UTC 2009
>Closed-Date:
>Last-Modified:
>Originator:     Dan Lukes
>Release:        FreeBSD 7.2-RELEASE-p3 i386
>Organization:
Obludarium
>Environment:
System: FreeBSD 7.2-RELEASE-p3 i386
but it seems to apply to 7.0-R and 7.1-R as well

databases/firebird20-client/Makefile,v 1.6 2009/08/23 10:40:40 erwin

>Description:
Program received signal SIGSEGV, Segmentation fault.
in ThreadData::restoreSpecific ()
from /usr/local/lib/libfbclient.so.2

src/jrd/ThreadData.cpp:233

For detailed description including back-trace see

http://bugs.php.net/bug.php?id=44711

>How-To-Repeat:
	just use the library. From example in php5-interbase

	The ibase_connect () call should be sufficient
>Fix:

	The bug is related to GCC and it's ISO-style handling of thread-specific data.

	It's the reason why the only 7.x are affected, not the 6.x (the GCC version doesn't support it here).

	I don't know if it is bug in gcc or the the feature is used in unexpected/improper way.

	Problem disappear when we force the code to use POSIX way of handling thread specific data.
--- patch-DAN-configure.in begins here ---
--- configure.in.ORIG	2009-11-07 17:47:10.000000000 +0100
+++ configure.in	2009-11-07 17:47:35.000000000 +0100
@@ -696,7 +696,7 @@
 rm -f conftest*])
 
 if test "$firebird_cv_gcc___thread" = yes; then
-  AC_DEFINE(HAVE___THREAD, 1, [Define it if compiler supports ISO syntax for thread-local storage])
+dnl  AC_DEFINE(HAVE___THREAD, 1, [Define it if compiler supports ISO syntax for thread-local storage])
 fi
 				  
 dnl ##################### DO NOT ADD ANY TESTS BELOW ###########################
--- patch-DAN-configure.in ends here ---


>Release-Note:
>Audit-Trail:
>Unformatted: