ports/187472: lang/ruby21 -- configure has mistakes/errors relating to pthread detection
Jeremy Chadwick
jdc at koitsu.org
Wed Mar 12 05:30:01 UTC 2014
>Number: 187472
>Category: ports
>Synopsis: lang/ruby21 -- configure has mistakes/errors relating to pthread detection
>Confidential: no
>Severity: non-critical
>Priority: medium
>Responsible: freebsd-ports-bugs
>State: open
>Quarter:
>Keywords:
>Date-Required:
>Class: sw-bug
>Submitter-Id: current-users
>Arrival-Date: Wed Mar 12 05:30:00 UTC 2014
>Closed-Date:
>Last-Modified:
>Originator: Jeremy Chadwick
>Release: FreeBSD 9.2-STABLE amd64
>Organization:
>Environment:
System: FreeBSD icarus.home.lan 9.2-STABLE FreeBSD 9.2-STABLE #0 r262477: Tue Feb 25 01:04:30 PST 2014 root at icarus.home.lan:/usr/obj/usr/src/sys/X7SBA_RELENG_9_amd64 amd64
>Description:
This may affect other ruby ports (ruby19, ruby20), I haven't checked.
During configure-time, the following lines are shown (and easily
missed because they scroll by so fast):
checking for pthread_kill in -lpthread... yes
./configure: pthread,: not found
checking for pthread_np.h... yes
The problem appears to be around line 19613 of the configure script (not
configure.in, I haven't reviewed that):
19608 if test x"$rb_with_pthread" = xyes; then
19609 $as_echo "#define _REENTRANT 1" >>confdefs.h
19610
19611 $as_echo "#define _THREAD_SAFE 1" >>confdefs.h
19612
19613 pthread, MAINLIBS="-pthread $MAINLIBS",
19614 $as_echo "#define HAVE_LIBPTHREAD 1" >>confdefs.h
This looks quite broken. Since autoconf is used, the "source" for
this comes from configure.in. The code for that:
2431 if test x"$rb_with_pthread" = xyes; then
2432 AC_DEFINE(_REENTRANT)
2433 AC_DEFINE(_THREAD_SAFE)
2434 [pthread], [MAINLIBS="-pthread $MAINLIBS"],
2435 AC_DEFINE(HAVE_LIBPTHREAD)
Line 2434 shown above comes directly from files/patch-configure.in,
so it appears to me someone needs to investigate the following:
1) If files/patch-configure.in is TRULY doing the correct thing
for ruby21 (and possibly the other ports) -- or if it's even
necessary (someone may want to talk to the Ruby folks), and,
2) Examination of post-patch in Makefile to make sure it's doing the
right thing as well.
I tried looking at the SVN annotations for this port, but the commit
lfor the section where post-patch is is so old that it precedes use
of Subversion thus the commit reason is unknown.
>How-To-Repeat:
See above.
>Fix:
Correct fix unknown at this time.
>Release-Note:
>Audit-Trail:
>Unformatted:
More information about the freebsd-ports-bugs
mailing list