GNOME Bugzilla – Bug 564740
configure error on python module detection
Last modified: 2009-01-12 20:32:31 UTC
Detection of some python modules now fails with: ** (.:11887): WARNING **: g_thread_init() has not been called. ORBit2 would like to use threads, so the application should have called g_thread_init(NULL) before any GLib or GLib-using API. Calling g_thread_init() now here inside the libORBit-2 library is too late, so instead threads won't be used by ORBit2. This might have bad side-effects. Fix the application! GLib-ERROR **: The thread system is not yet initialized. aborting...
Created attachment 124797 [details] [review] init threads in python test script This fixes the build issue; however I didn't actually run Orca and it may well be this is also needed at runtime.
(In reply to comment #0) > Detection of some python modules now fails with: > > ** (.:11887): WARNING **: g_thread_init() has not been called. ORBit2 would > like to use threads, so the application should have called g_thread_init(NULL) > before any GLib or GLib-using API. Calling g_thread_init() now here inside the > libORBit-2 library is too late, so instead threads won't be used by ORBit2. > This might have bad side-effects. Fix the application! > > GLib-ERROR **: The thread system is not yet initialized. > aborting... > Thanks Frederic! Do you know which specific modules this fails with? I'm curious why this hasn't happened before and if a specific module import (e.g., ORBit?) is causing this. The reason I ask is that I cannot reproduce the problem on my machine, but I notice there are a couple places in acinclude.m4 that we might need to modify in a manner similar to your patch.
checking for python module bonobo... ./configure: line 23706: 9228 Trace/breakpoint trap (core dumped) $PYTHON -c "$prog" 1>&5 2>&5 no configure: error: Could not find python module bonobo - make sure you have gnome-python with support for libbonobo-2.0 >= 2.14.0 installed The thing is I suddenly got all those g_thread_init() warnings and couldn't find the culprit (as I spent a few days away of my jhbuild environment); and asking other people I couldn't find any body with the same problem; so it may well be something local and not a real issue. I'll get to rebuild an environment soon, to confirm or invalidate this bug report. (I'm marking it as NEEDINFO so I remember to get back).
So the thread init issue was related to a change in ORBit2, that has been reverted; this bug can be closed.