Skip to content
Commits on Jan 26, 2011
  1. Revert "attempt to fix "undefined symbol: pyglib_init" on python -c "…

    committed
    …import gio""
    
    This reverts commit 4b59558.
    
    Breaks pyglib with undefined errors.
Commits on Jan 25, 2011
  1. @dieterv
  2. @dieterv

    also allow libtool 2.4

    dieterv committed
Commits on Jan 23, 2011
  1. @dieterv
Commits on Jan 22, 2011
  1. Install pygsource system wide

    committed
  2. Make pyglib system wide

    committed
Commits on Jan 21, 2011
  1. @sebp

    pygi-convert.sh: Convert gdk.GRAB_*

    sebp committed
  2. @nacho
  3. @nacho
  4. @martinpitt
  5. @martinpitt

    [gi] Add Variant construction/unpack support for boxed Variants

    martinpitt committed
    Construction uses a GVariantBuilder for now, as the new_variant() constructor
    currently does not work (see https://bugzilla.gnome.org/show_bug.cgi?id=639952)
  6. @dieterv
Commits on Jan 20, 2011
  1. @phomes

    pygi-convert.sh: GdkPixbuf methods

    phomes committed
    GNOME bug #639880
  2. @phomes

    pygi-convert.sh: Gdk.COLORSPACE_RGB

    phomes committed
    GNOME bug #639880
  3. @martinpitt

    [gi] Support nested objects and empty sequences in GLib.Variant building

    martinpitt committed
    The GVariant constructor (in the overrides) previously did not support empty
    arrays/dictionaries or nested structures. Rewrite the VariantCreator class to
    be fully recursive and determine the element types of arrays/dictionaries.
    
    This now also allows you to use actual tuples as input values for GVariant
    tuple types. Taking values from the flat argument list is still supported for
    convenience, though.
    
    https://bugzilla.gnome.org/show_bug.cgi?id=639939
  4. @tomeuv
  5. @nud @tomeuv

    Fix reference leaks for GInitiallyUnowned objects

    nud committed with tomeuv
    References were leaked for GInitiallyUnowned objects which got their
    wrappers created several times, because someone else holds reference
    on it and it got out of python scope at some point.
    
    https://bugzilla.gnome.org/show_bug.cgi?id=639949
  6. @nud @tomeuv

    Add tests for refcount of a GObject owned by a library

    nud committed with tomeuv
    When the object is constructed, its refcount is 2 because the library
    refs it once. It should remain around until we ask the library to
    release its reference.
    
    https://bugzilla.gnome.org/show_bug.cgi?id=639949
  7. @nud @tomeuv
  8. @martinpitt

    [gi] Update TreeView.enable_model_drag_{source,dest} to current GTK

    martinpitt committed
    GTK master now landed a lot of annotation fixes which also correctly marks the
    array length argument of Gtk.TreeView.enable_model_drag_{source,dest}(). Thus
    drop the explicit array length argument from the call in the override.
  9. @nud

    Fix a typo in a private symbol name.

    nud committed
  10. @martinpitt
  11. @martinpitt
  12. @martinpitt
  13. Modify override for Gtk.Adjustment to allow position or keyword argum…

    Laszlo Pandy committed
    …ents in __init__().
    
    Previously passing no arguments was not working, because the default value for each parameter was None, and GObject.__init__() refuses to allow None for integer properties. This patch does not pass None up to GObject.__init__. Instead it does not pass the parameter at all, and uses the class's default values.
    
    https://bugzilla.gnome.org/show_bug.cgi?id=639934
Commits on Jan 19, 2011
  1. @martinpitt

    [gi] Fix small typo in previous commit

    martinpitt committed
    The GVariant signature of the self test had a trailing 'i'. The current
    GVariant builder doesn't mind, but the new implementation proposed in bug
    639939 does.
  2. @martinpitt

    [gi] Add pythonic iterator and indexing for string GVariants

    martinpitt committed
    This extends commit b1a9808 to also work for strings.
  3. @tomeuv

    Construct structs using default API constructor

    tomeuv committed
    If the struct has something that looks like a default constructor,
    use it instead of trying to directly allocate it, as it will fail
    if the struct fields are not exposed.
    
    https://bugzilla.gnome.org/show_bug.cgi?id=627444
  4. @martinpitt
  5. @martinpitt

    pygi-convert.sh: Handle .window attributes

    martinpitt committed
    In general, convert them to .get_window(). For some of them, prefer calling the
    GtkWidget methods instead.
  6. @tomeuv

    Also deal with foreign boxed structs

    tomeuv committed
    cairo.Context has been boxed and our tests started failing
    
    https://bugzilla.gnome.org/show_bug.cgi?id=639967
  7. [gi] Convert GErrors to GObject.GError exceptions, and throw them upo…

    Laszlo Pandy committed
    …n returning from calling the C function.
    
    This changes gi to make use of pyglib_error_check() which already exists in pyglib.
    
    The included tests make use of the other patch attached to this bug, to check that the right exception is thrown from the new function in GIMarshallingTests.
    two Gtk C functions.
    
    https://bugzilla.gnome.org/show_bug.cgi?id=639834
  8. @martinpitt

    pygi-convert.sh: Don't convert glib -> GLib for now

    martinpitt committed
    This currently leads to a load of crashes, MemoryErrors, etc, as GLib is not
    very well introspectable, due to the low-level operations that it performs.
    
    John Palmieri confirms that using the static "glib" binding is preferred for
    now, so disable the replacement rules.
  9. @tomeuv
  10. Fix attributes 2BUTTON_PRESS and 3BUTTON_PRESS of Gdk.EventType.

    Laszlo Pandy committed
    This puts an underscore in front of 2BUTTON_PRESS and 3BUTTON_PRESS because in Python attributes starting with a numeral causes a syntax error.
Something went wrong with that request. Please try again.