commit 6a629e23ff7b0d6f532184017577c7427d577e28 Author: Martin Pitt Date: Mon Aug 20 23:05:49 2012 +0200 release 3.3.90 NEWS | 13 +++++++++++++ configure.ac | 2 +- 2 files changed, 14 insertions(+), 1 deletion(-) commit 5cd18c9bd59a60b930ced0b35d728c12bb3291c7 Author: Mathieu Duponchelle Date: Mon Aug 20 22:54:52 2012 +0200 Implement marshalling for GParamSpec https://bugzilla.gnome.org/show_bug.cgi?id=681565 Co-Authored-By: Martin Pitt gi/pygi-marshal-to-py.c | 16 ++++++++++++---- tests/test_gi.py | 14 ++++++++++++++ 2 files changed, 26 insertions(+), 4 deletions(-) commit 16462de3f025f14706ec23fa9b3653feb66ad57f Author: Martin Pitt Date: Mon Aug 20 15:24:10 2012 +0200 Fix pep8/pyflakes invocation Fix regression from commit 1e056e4f4a: Do fail the tests if pyflakes/pep8 exist, but fail. tests/Makefile.am | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) commit 1bee194274bcda9ba5f6751fa921218a92c8ac72 Author: Simon Feltman Date: Thu Aug 16 16:05:52 2012 -0700 Fix erronous import statements for Python 3.3 Update pygobject-external.h to use GType (which is what GTypeWrapper is exposed as) instead of GTypeWrapper when attempting import. Catch ImportError around attempted imports of a typelibs override file which don't always exist (GObject...). This is a behavioural change in Python 3.3 (http://bugs.python.org/issue15715), but let's fix it anyway. https://bugzilla.gnome.org/show_bug.cgi?id=682051 gi/module.py | 7 +++++-- gi/pygobject-external.h | 2 +- 2 files changed, 6 insertions(+), 3 deletions(-) commit 1e056e4f4a19fd1139187467677c2592c2722290 Author: Martin Pitt Date: Mon Aug 20 11:52:08 2012 +0200 Do not fail tests if pyflakes or pep8 are not installed These tools might not be desirable in restricted build environments or backports, and e. g. Fedora patches those out. So let the tests work without these tools. tests/Makefile.am | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) commit c219fa6da89a7d55c5c111751684aae6876a9fe3 Author: Martin Pitt Date: Mon Aug 20 11:42:47 2012 +0200 gtk-demo: Fix some PEP-8 whitespace issues demos/gtk-demo/demos/Icon View/iconviewbasics.py | 3 +-- demos/gtk-demo/demos/dialogs.py | 2 +- demos/gtk-demo/demos/rotatedtext.py | 3 +-- 3 files changed, 3 insertions(+), 5 deletions(-) commit 0ac2a85cae368c046839b5619a96efc9e0b91ba3 Author: Martin Pitt Date: Mon Aug 20 11:38:50 2012 +0200 test_overrides.py: Fix PEP8 whitespacing tests/test_overrides.py | 48 ++++++++++++++++++++++++----------------------- 1 file changed, 25 insertions(+), 23 deletions(-) commit 631a9cd05cbc7dc3d0f743a84b948ef7d93c0ed4 Author: Martin Pitt Date: Mon Aug 20 11:36:19 2012 +0200 Ignore E124 pep8 error This is "closing bracket does not match visual indentation" which is really stupid. We do want the closing bracket at the same indentation level as the opening bracket, not the indentation level of the whole statement. tests/Makefile.am | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) commit 266d37719bb54e6f04d23ff21bcceb9514e20ff2 Author: David Malcolm Date: Mon Aug 20 11:27:52 2012 +0200 Fix unmarshalling of gssize Do not assume that the v_int union member always corresponds to a gssize. This is not true on big-endian 64 bit machines like ppc64, so add a new gi_argument_to_gssize() and use it properly. https://bugzilla.gnome.org/show_bug.cgi?id=680693 https://bugzilla.redhat.com/show_bug.cgi?id=842880 gi/pygi-argument.c | 53 +++++++++++++++++++++++++++++++++++++++++++--- gi/pygi-argument.h | 1 + gi/pygi-closure.c | 2 +- gi/pygi-info.c | 4 ++-- gi/pygi-signal-closure.c | 2 +- 5 files changed, 55 insertions(+), 7 deletions(-) commit 1c5d497d3c354f4d02f1d4570df2c61d6f47300c Author: David Malcolm Date: Mon Aug 20 11:19:27 2012 +0200 Fix various endianess errors Fix code which assumed little endian behaviour when mixing different types of ints, putting ints into pointers, etc. https://bugzilla.gnome.org/show_bug.cgi?id=680692 https://bugzilla.redhat.com/show_bug.cgi?id=841596 gi/pygi-argument.c | 64 ++++++++++++-- gi/pygi-argument.h | 6 ++ gi/pygi-cache.c | 2 + gi/pygi-closure.c | 57 ++++++++++++- gi/pygi-marshal-from-py.c | 203 +++++++++++++++++++++++++++++++++++---------- gi/pygi-marshal-from-py.h | 3 + gi/pygi-marshal-to-py.c | 137 +++++++++++++++++++++++------- 7 files changed, 391 insertions(+), 81 deletions(-) commit ee6da6f1aa2cd6e55834f9edc17f785613d00031 Author: Paolo Borelli Date: Wed Aug 15 13:16:11 2012 +0200 Add unit test for the TreeModelSort override tests/test_overrides.py | 7 +++++++ 1 file changed, 7 insertions(+) commit 9f027daa5737107b5959964b699c0089aec8ab1e Author: Simon Feltman Date: Thu Aug 9 03:33:06 2012 -0700 Gtk overrides: Add TreeModelSort.__init__(self, model) This adds "model" as a required argument to TreeModelSort instead of it being a hidden keyword argument. This is needed because the model property is set to construct only and the default value of None/NULL makes the object useless anyhow. https://bugzilla.gnome.org/show_bug.cgi?id=681477 gi/overrides/Gtk.py | 8 ++++++++ 1 file changed, 8 insertions(+) commit c8424c2bb19356679e250e73542682dd5f4c74a5 Author: Manuel QuiƱones Date: Fri Aug 10 09:38:24 2012 -0300 Convert Gtk.CellRendererState in the pygi-convert script Signed-off-by: Manuel QuiƱones https://bugzilla.gnome.org/show_bug.cgi?id=681596 pygi-convert.sh | 5 +++++ 1 file changed, 5 insertions(+) commit 54d829b34a0d32d852db370f61cc7f25c149f373 Author: Paolo Borelli Date: Mon Aug 6 16:19:28 2012 +0200 More updates to the HACKING file module-install has been replaced with 'ftpadmin install' and other minor changes HACKING | 26 ++++++++++---------------- 1 file changed, 10 insertions(+), 16 deletions(-) commit 0788880c6cf4070d3db09896c165fe470d2ec186 Author: Paolo Borelli Date: Mon Aug 6 16:00:39 2012 +0200 Post-release version bump to 3.3.6 configure.ac | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)