Since introducing gimp_display_shell_canvas_tick() we were setting
shell->disp_width and shell->disp_height in the tick callback, which
was too late and caused redundant rendering and scrolling, but was
never noticed.
Now we clear the render cache and its valid region in
gimp_display_shell_canvas_size_allocate() directly and set
shell->disp_width and shell->disp_height immediately, so other places
that listen to the canvas' size-allocate get the right values.
The old size of the canvas gets to the tick callback using a small
struct as user data.
Revert the previous commit 786686a541 and comes up with a better fix.
Let's actually change the image base type and add its colormap as close
as possible without any GUI calls in-between. I also add an explicative
comment so that people are aware of this call proximity requirement to
avoid future problems when the code gets remixed.
It should be better than hacking around with exception in GUI code, and
should (hopefully) avoid other similar bugs.
Fixes the GIMP-CRITICAL:
> gimp_babl_format_get_trc: assertion 'format != NULL' failed".
During an indexed conversion, we are in a weird limbo state between the
time we changed the base-type to GIMP_INDEXED and when we actually set
the new palette. If during this time, we hit a context switch (which
typically happens during GUI code; in this specific cases, the various
progress updates would call gimp_widget_flush_expose() which does
trigger context switch), then gimp_display_shell_update_title_idle()
might be idly called. And when it does, it sees an indexed function with
neither format nor builtin profile.
Must initialize shell->render_buf_width,height before realize(), so
move the code to gimp_display_shell_init(), it doesn't depend on the
shell being realized.
When set to FAST we do nearest neighbor from the next bigger mipmap
level instead of linear or box filtering - this gives a slight and
permanent boost to painting, and all updates, having one that combines
best of both worlds and reblits in high quality after a timeout would
be even more desirable.
Introduce a render cache that keeps the result of scaling, color
management, display filters and shell mask (for tools like fuzzy
select).
Change gimpdisplayshell-render.[ch] to only render to the cache and
manage a cairo region of the cache's valid area. Call cache
invalidation functions form various places. Change the API of all
render functions to be in display coordinates.
Also get rid of gimpdisplayxfer.[ch] because we now have a
canvas-sized cairo surface which is a surface similar to the
destination surface.
Step one: get rid of all those deprecation warnings that make
it hard to see any other warnings:
- add a lot of dummy API to GimpAction, GimpActionGroup, GimpUIManager
etc. which simply forwards to the deprecated GTK functions, they
will all go away again later
- rename GimpAction to GimpActionImpl
- add interface GimpAction that is implemented by all action classes,
creates a common interface and allows to remove some duplicated
logic from GimpToggleAction and GimpRadioAction, and at the same
time adds more features
Connect GimpDisplayShell to GimpImage::mode-changed and update its
cached color transforms. Also get rid of a duplicate conection to
GimpImage::precision-changed.
This is useful to be able to support file formats other than image
formats. In particular I will use this in the next commit to support a
"GIMP extension" format. When GIMP will open such file, it will
install an extension (not open an image on canvas).
This is an internal flag only, i.e. only usable from core GIMP. File
formats which a plug-in can register are still only image file formats.
Set the priority of the image window's UI-manager update idle
slightly higher than the projection idle priority, so that the
image actions are updated during projection rendering.
In GimpTransformGridTool, avoid producing non-finite coordinate
and angle values. In particular, this fixes a crash in
gimp_transform_grid_tool_get_cursor() as a result of a NaN angle
value being converted to a negative integer, hitting an assert.
In gimp_tool_gui_set_response_sensitive(), silently ignore non-
existent response IDs, instead of emitting a CRITICAL, to match the
behavior of GtkDialog and GimpOverlayDialog. This simplifies code
with optional dialog responses.
In the unified-transform, scale, and perspective tools, take the
maximal transform-grid handle size into account when readjusting
the transform, so that the handles themselves are fully within view
under arbitrary rotation, rather than just the corners.
In GimpToolTransformGrid, use the transformed center-point of the
original polygon as the position of the center-point handle, and as
a snapping point for the pivot when "cornersnap" is TRUE, instead
of using the center-point of the transformed polygon. These two
points are different for non-affine transformations.
Furthermore, when "use-pivot-handle" is FALSE, use the center-
point as the reference point when transforming around the pivot,
instead of the pivot itself, which might not be set. In
particular, this fixes transformation around the pivot in the
perspective tool.
Don't explicitly set the center-point as the pivot in the scale
tool, since this commit makes it unnecessary.
gimp_widget_flush_expose() has been removed since commit 3089a20167.
I now reimplemented it by simply checking if event sources are waiting
to be processed.
This was heavily needed as the statusbar was not showing any progress
(at least on highly demanding process, such as saving or loading files),
and therefore we were stuck with a seamingly frozen GUI.
Despite the name, this does not apply to a widget in particular anymore,
but to the whole program events.
`icon_space_width` is set when GtkWidget::style-updated signal is
emitted. In some cases, it was possible that gimp_statusbar_set_text()
is run before this happens, in which case, a division by zero would
crash the software.
I encountered this case in some rare occasions when duplicating an image
with ctrl-d, then as Ctrl was hold, the message "Click in any image to
pick the foreground color" was pushed on the brand new statusbar as it
is created (hence a race condition occurs with the signal handler and
this message). This was therefore not reproducible every time, but easy
enough to reproduce with multiple tests.
In GimpToolCompass, add a read-only "effective-orientation"
property, which returns the actual orientation of the compass; in
particular, if the "orientation" property is set to AUTO,
"effective-orientation" returns HORIZONTAL or VERTICAL, depending
on the current compass direction. In 3-point mode, the property
always returns AUTO.
In GimpToolRectangle, fix the type of the cornder_radius field, so
that non-integer radii are properly displayed.
In GimpRectangleSelectOptions and GimpToolRectangle, increase the
maximal corner radius.
We were doing it all the wrong way, fixing one combo box object at a
time. So this commit basically reverses commits 68a33ab5bd, 6dfca83c2a
and a9a979b2d0 and instead runs the same code in the class code. This
way, all objects based on these base classes will have the fix from
scratch.
These improved various other drop-down lists (I found some of them, and
probably not all) as I fixed all GIMP custom widgets based on
GtkComboBox.
Note that it has to be run after filling the list apparently (I had the
problem especially with GimpIntComboBox if running in the _init() code,
then the list widget showed wrong).
Add mew gimp_tool_widget_{get,set}_visible() functions, which allow
setting the visibility of a tool widget. While the widget is
invisible, it ignores all events.
Add a GimpToolWidget::message signal, which can be emitted by tool
widgets to display a message, instead of using the ::status signal.
Add corresponding gimp_tool_widget_message[_literal]() functions.
Though the layer list will also show updated, it is much easier to look
at the layer name in the status bar whose position never changes.
Anyway it makes sense to just show a temporary status info message
giving the picked layer name, making it all the easier to find the layer
you are looking for.
In gimp_display_empty(), clear the image of all contexts whose
display is the current display, so that, in particular, when
subsequently updating the action groups, which causes certain
actions to be activated, the image that used to be bound to the
display is not found through the user context. This avoids re-
validating the image projection when closing the last image,
postponing image destruction.
Instead of having layer picking only on paint tools with alt-click, make
it available everywhere with alt-middle click. Moving through layers is
also a way to navigate an image, so it actually makes sense to be with
other modifiers (panning, zooming, rotating), while making the feature
more generic (this is definitely useful whatever the selected tool).
...closing this image while the file is being loaded
Ref the image around all calls to file_open_layers() and
gimp_image_add_layers() so it stays around even if the user closes the
display in the meantime.
Set the label to ELLIPSIZE_END and add some packing options, and
switch from using pango attributes to simply prepending the right
number of spaces to leave enough room for the icon. The attributes
were somehow breaking ellipsation.
In all tools, when the current item can't be edited due to its lock
mask, use gimp_tools_blink_lock_box(), added in the previous
commit,to blink the lock box of the corresponding dockable, in
addition to showing an error message in the status bar, to hint at
the source of the error.
In gimp_tool_compass_update_angle(), use fuzzy comparisson when
determining whether to update the angle properties, to avoid
infinite recursion due to floating-point inaccuracies. In
partcicular, on x86, when using the x87 FPU rather than SSE, the
floating-point registers are 80-bit, while the properties are
stored as 64-bit, which can create small discrepancies between the
calculated angles and the stored values.
Michael Natterer