not GimpCanvasLine* etc. We keep them around as items, and their
constructors return items, and it's getting more code anyway when
porting away from pause/resume, so let's keep the casting minimal at
least.
Instead, keep around the created GimpCanvasItems, and update them when
the blend coordiates change. Add setters to GipmCanvasLine and
GimpCanvasHandle which take care of calling begin_change() and
end_change() on the items around the change, so thes invalidate
properly.
Instead, add drawable, transform, x1, y1 etc. properties. This is
cleaner and has the nice side effect of not leaving artifacts, because
the changed state of the transform tool doesn't affect the extents
calculation any longer.
And remove all the complicated handling code entirely. This makes
GimpTransformTool a lot less complex. As a nice side effect, the
preview is now always 100% in sync with the grid and handles.
This was only there to make sure XOR drawing works, and is now
complete overhead since cairo drawing is done *in* expose, and not in
some hack around it.
Change the "GtkWidget *parent" parameter of GimpToolDialog to
"GimpDisplayShell". Also add API gimp_tool_dialog_set_shell() so an
existing dialog can be used on different images. Make sure the dialog
closes when the shell is unmapped (like when switching tabs in SWM),
and make the dialog transient for the toplevel GimpImageWindow.
Change all tool dialog users accordingly.
The problem is that gimp_image_window_keep_canvas_pos() connects a
signal handler too early, with data not yet valid, the handler is
executed after the canvas position is computed correctly and so
invalid data overrides the valid.
Checks the return value of gtk_widget_translate_coordinates() to avoid
connecting handlers when their execution is considered harmful.
- change gimp_devices_get_list() to gimp_devices_get_manager()
- remove gimp_devices_get_current()
- let GimpDeviceStatus connect to the manager's "notify::current-device"
directly and remove device status updating code from gui.c
- cleanup in gimpdevices.c
Keep the comment at the place where calling it isn't immediately
obvious, and add checks to proximity_in() and out() so they bail out
when the proximity state is already set correctly.
and use them instead of manually setting proximity. Also add comments
about why we call the proximity functions where we call them (instead
of calling them where it seems more logical).
Spit warnings in the functions if grabbing fails and return a boolean
success value. Bail out in the callers upon grab failure instead of
assuming that grabbing always succeeds and running into an
inconsistent state that can cause all sorts of problems.
Bump the "too new" test to GTK+ 2.26 which will never be released, but
this way we can keep the configure check around for later reuse after
the switch to GTK+ 3.x. Also bump the "too new" GLib and Pango
versions to really future versions so we early-catch deprecations
in the current development cycle.
when dealing with native window handles. Also get rid of using
GdkNativeWindow and simply consistently use guint32 all over the
place. This is more obvious and consistent and keeps the diff to the
gtk3-port branch smaller.
Create and use Cairo-compatible API for display filters. Also
includes logic changes to the display filters to deal with cairo's
ARGB32 pre-multiplied buffer format.
Add a list of "track widgets" to GimpRuler and connect to their
motion-notify-event. Correctly translate the motion event's x/y to the
ruler's coordinate system when updating the marker.
