> Has anyone investigated interactions between ExtensionClass objects and GC? > I've encountered segfaults with 2.1 in certain situations when using the > latest PyGtk stuff. The gdb traceback (appended) sort of suggests the two > intersect somewhere. PyGtk provides a Python interface to the Gtk widget > get using ExtensionClasses. Any ideas how I should approach the problem? I > don't know either piece of code at all and the code that generates the > segfault isn't particularly small, not to mention which it uses the bleeding > edge Gtk stuff (which I doubt anyone on this list will have installed) and a > version of ExtensionClass patched by James Henstridge, the PyGtk author. > > Here's what I know: > > 1. Disabling gc gets rid of the segfault > 2. I only see the problem with importing a specific module that > subclasses the GtkTextView widget from the Python command line. If I > run it as a script from the shell prompt, I get no segfault. > 3. If I first import the gtk module, then import my module, I get no > segfault. > 4. Most changes I make to the module causing the problem cause the > problemm to disappear. > > All told, all this really tells me is I'm probably dealing with a > malloc/free problem of some sort. > > Neil and/or Jim (and/or anyone else willing to look into this problem), I > can give you access to my development machine via ssh if you think that > would help debug the problem. AFAIK, the latest version of Zope (which uses ExtensionClass extensively if not exclusively :-) works fine with Python 2.1. This suggests pointing a finger towards the PyGtk code... :-( --Guido van Rossum (home page: http://www.python.org/~guido/)
RetroSearch is an open source project built by @garambo | Open a GitHub Issue
Search and Browse the WWW like it's 1997 | Search results from DuckDuckGo
HTML:
3.2
| Encoding:
UTF-8
| Version:
0.7.4