Bug 160717

Summary: port: lang/python27 causing other ports to hang during configure?
Product: Ports & Packages Reporter: Kris Moore <kris>
Component: Individual Port(s)Assignee: freebsd-python (Nobody) <python>
Status: Closed FIXED    
Severity: Affects Only Me    
Priority: Normal    
Version: Latest   
Hardware: Any   
OS: Any   

Description Kris Moore 2011-09-13 22:30:07 UTC
I'm running into some very weird problems when trying to build xmms2 on a 9.0-BETA2 system (also saw this on BETA1). What's happening is that the port begins to build, and just "halts" during the configure phase, as shown below:

===>  Configuring for xmms2-0.7_2
cd /usr/ports/audio/xmms2/work/xmms2-0.7DrNo && ./waf configure --conf-prefix=/usr/local  --prefix=/usr/local --mandir=/usr/local/man --without-optionals="python"  --without-optionals="ruby"  --without-optionals="xmmsclient-ecore"  --without-optionals="xmmsclient++"  --without-optionals="xmmsclient++-glib"  --without-optionals="perl" --without-plugins="airplay" --without-plugins="ao" --without-plugins="mac" --without-plugins="avcodec" --without-plugins="cdda" --without-optionals="medialib-updater" --without-plugins="gme" --without-plugins="gvfs" --without-plugins="ices" --without-plugins="jack" --without-optionals="avahi" --without-optionals="avahi" --without-plugins="mms" --without-plugins="modplug" --without-plugins="mpg123" --without-plugins="musepack" --without-plugins="ofa" --without-plugins="pulse" --without-plugins="samba" --without-plugins="sid" --without-plugins="speex" --without-plugins="tremor" --without-optionals="vistest" --without-plugins="vocoder" --without-plu
 gins="wavpack" --without-plugins="xml" --without-plugins="rss"  --without-plugins="xspf"
Checking for program gcc,cc              : ok /usr/bin/gcc 
Checking for program cpp                 : ok /usr/bin/cpp 
Checking for program ar                  : ok /usr/bin/ar 
Checking for program ranlib              : ok /usr/bin/ranlib 
Checking for program g++,c++             : ok /usr/bin/g++ 
Checking for program windres             : not found 
Checking for git commit id               : ok 4bd0e653c2fd8b36a5f2ae4203ff0b37b03180f0 
Checking for uncommitted changes         : ok  
Checking for compiler flags -Wall        : ok 
Checking for compiler flags -Wno-format-extra-args : ok 
Checking for compiler flags -Wno-format-zero-length : ok 
Checking for compiler flags -Wformat-nonliteral     : ok 
Checking for compiler flags -Wformat-security       : ok 
Checking for compiler flags -Wformat=2              : ok 
Checking for compiler flags -Wmissing-prototypes    : ok 
Checking for compiler flags -Wstrict-prototypes     : ok 
Checking for compiler flags -Wempty-body            : fail 
Checking for compiler flags -Wignored-qualifiers    : fail 
Checking for compiler flags -Wtype-limits           : fail 
Checking for compiler flags -Wwrite-strings         : ok 
Checking for platform code for posix                : ok  
Checking for glib-2.0 >= 2.8.0                      : ok 
Checking for library m                              : ok 
Checking for header sys/soundcard.h                 : ok 
Checking for library ossaudio                       : not found 
Checking for vorbisfile                             : ok 
Checking for library m                              : ok 
Checking for function fstatat                       : ok 
Checking for function dirfd                         : not found 
Checking for libcurl                                : ok 
Checking for function curl_multi_strerror           : ok 
Checking for avahi-glib                             : ok 
Checking for avahi-client                           : ok 
Checking for sc68 >= 2.3.0                          : sc68-config: not found
fail 
Checking for alsa                                   : not found 
Checking for header AudioUnit/AudioUnit.h           : not found 
Checking for neuros-cooler-media                    : not found 
Checking for mad                                    : ok 
Checking for flac                                   : ok 
Checking for library m                              : ok 
Checking for library m                              : ok 
Checking for header faad.h                          : ok 
Checking for library faad                           : ok 
Checking for library m                              : ok 
Checking for header windows.h                       : not found 
Checking for header CUnit/CUnit.h                   : not found 
Checking for header ['stdio.h', 'readline/readline.h'] : 




The config.log looks pretty typical:
---------------------------------------------------------

# tail _build_/config.log

==>

int main() {
	return 0;
}

<==
[1/2] cc: _build_/.conf_check_0/test.c -> _build_/.conf_check_0/testbuild/default/test_1.o



When I look through "ps" I noticed some hanging "python" processes, killing these sometimes causes the configure to resume, kill the wrong one and we get a signal 9, and the configure fails. 
-------------------------------------------------------------------------

root   81812   0.0  0.4 108616  18568   1  I+    8:26PM   0:00.35 [python]
root   81823   0.0  0.4 104520  17312   1  I+    8:27PM   0:00.00 [python]


At first I thought this was a xmms2 bug, but I'm seeing it in devel/talloc as well which makes me think its a python issue. Here's the top of config.log which has some python stuff:

# project xmms2 configured on Tue Sep 13 17:19:15 2011 by
# waf 1.5.10 (abi 7, python 20702f0 on freebsd9)
# using ./waf configure --conf-prefix=/usr/local --prefix=/usr/local --mandir=/
usr/local/man --without-optionals=python --without-optionals=ruby --without-opt
ionals=xmmsclient-ecore --without-optionals=xmmsclient++ --without-optionals=xm
msclient++-glib --without-optionals=perl --without-plugins=airplay --without-pl
ugins=ao --without-plugins=mac --without-plugins=avcodec --without-plugins=cdda
 --without-optionals=medialib-updater --without-plugins=gme --without-plugins=g
vfs --without-plugins=ices --without-plugins=jack --without-optionals=avahi --w
ithout-optionals=avahi --without-plugins=mms --without-plugins=modplug --withou
t-plugins=mpg123 --without-plugins=musepack --without-plugins=ofa --without-plu
gins=pulse --without-plugins=samba --without-plugins=sid --without-plugins=spee
x --without-plugins=tremor --without-optionals=vistest --without-plugins=vocode
r --without-plugins=wavpack --without-plugins=xml --without-plugins=rss --witho
ut-plugins=xspf


Here's my make.conf
---------------------------------------------------------------------

# Keep KDE4 in /usr/local, fixes sharing of icons / mime and others
KDE4_PREFIX=/usr/local
# added by use.perl 2011-09-13 05:21:33
PERL_VERSION=5.12.4

How-To-Repeat: Install 9.0-BETA2 (amd64) and build audio/xmms2, or possibly devel/talloc from ports.
Comment 1 Edwin Groothuis freebsd_committer freebsd_triage 2011-09-13 22:30:20 UTC
Responsible Changed
From-To: freebsd-ports-bugs->freebsd-python

Over to maintainer (via the GNATS Auto Assign Tool)
Comment 2 Ruslan Makhmatkhanov freebsd_committer freebsd_triage 2013-05-18 12:31:53 UTC
State Changed
From-To: open->feedback

Kris, is this still the case? I wasn't able to reproduce this on 
different versions of FreeBSD with different python versions (2.7.3 and 
2.7.5) for both ports
Comment 3 Ruslan Makhmatkhanov freebsd_committer freebsd_triage 2013-06-10 09:28:46 UTC
State Changed
From-To: feedback->closed

Thanks for the report. Closing this pr because it's believed that this 
is fixed in r318353, that reverts incorrect threading patch. I wasn't 
able to reproduce the breakage and we have submitter's response timeout.