After an evaluation, GNOME has moved from Bugzilla to GitLab. Learn more about GitLab.
No new issues can be reported in GNOME Bugzilla anymore.
To report an issue in a GNOME project, go to GNOME GitLab.
Do not go to GNOME Gitlab for: Bluefish, Doxygen, GnuCash, GStreamer, java-gnome, LDTP, NetworkManager, Tomboy.
Bug 694862 - gtype array leak
gtype array leak
Status: RESOLVED FIXED
Product: pygobject
Classification: Bindings
Component: introspection
Git master
Other Linux
: Normal normal
: ---
Assigned To: Nobody's working on this now (help wanted and appreciated)
Python bindings maintainers
Depends on:
Blocks: 693111
 
 
Reported: 2013-02-28 12:26 UTC by Martin Pitt
Modified: 2013-10-11 06:28 UTC
See Also:
GNOME target: ---
GNOME version: ---


Description Martin Pitt 2013-02-28 12:26:01 UTC 
This code

G_SLICE=debug-blocks PYTHONPATH=. valgrind --tool=memcheck --leak-check=full --show-possibly-lost=no python3 -c 'from gi.repository import Gtk, GObject; st = Gtk.ListStore.new([GObject.TYPE_INT])'

shows the following leak:

==16974== 24 bytes in 1 blocks are definitely lost in loss record 1,744 of 7,999
==16974==    at 0x4C2CD7B: malloc (in /usr/lib/valgrind/vgpreload_memcheck-amd64-linux.so)
==16974==    by 0x9376C27: pyg_type_wrapper_new (pygtype.c:342)
==16974==    by 0x792E6A5: _pygi_marshal_to_py_gtype (pygi-marshal-to-py.c:284)
==16974==    by 0x79208E6: _invoke_marshal_out_args (pygi-invoke.c:553)
==16974==    by 0x7920F2A: pygi_callable_info_invoke (pygi-invoke.c:667)
==16974==    by 0x7921075: _wrap_g_callable_info_invoke (pygi-invoke.c:685)
==16974==    by 0x4814AD: PyEval_EvalFrameEx (ceval.c:4374)
==16974==    by 0x47A90D: PyEval_EvalCodeEx (ceval.c:3433)
==16974==    by 0x4818A6: PyEval_EvalFrameEx (ceval.c:4160)
==16974==    by 0x47A90D: PyEval_EvalCodeEx (ceval.c:3433)
==16974==    by 0x47A9BA: PyEval_EvalCode (ceval.c:771)
==16974==    by 0x47811C: builtin_exec (bltinmodule.c:850)

I tried to reduce this further to a method which just takes a GType, but this doesn't leak:

LD_LIBRARY_PATH=tests GI_TYPELIB_PATH=tests G_SLICE=debug-blocks PYTHONPATH=. python3 -c 'from gi.repository import GIMarshallingTests, GObject; GIMarshallingTests.gtype_in(GObject.TYPE_NONE)'
Comment 1 Martin Pitt 2013-02-28 13:34:00 UTC 
I was eliminating the overrides. This snippet does not leak:

python3 -c 'from gi.module import get_introspection_module as gim; Gtk=gim("Gtk"); GObject=gim("GObject"); st = Gtk.ListStore.new([GObject.type_from_name("guint")])'

But when I save the GType variable as I, it does leak:

python3 -c 'from gi.module import get_introspection_module as gim; Gtk=gim("Gtk"); GObject=gim("GObject"); I = GObject.type_from_name("guint"); st = Gtk.ListStore.new([I])

when not using the I reference, but recreating it again, it does not leak:

python -c 'from gi.module import get_introspection_module as gim; Gtk=gim("Gtk"); GObject=gim("GObject"); I = GObject.type_from_name("guint"); st = Gtk.ListStore.new([GObject.type_from_name("guint")])'

But if I move the GObject and Gtk overrides away from gi/overrides/, the leak again does not happen.
Comment 2 Simon Feltman 2013-10-11 06:28:06 UTC 
I was unable to reproduce any of the noted leaks with what is currently in master. These were most likely fixed with stuff that has been going in for bug 693402.