Though I had already implemented passing GeglColor through the PDB, it was not
complete. In particular, the protocol was not able to pass GeglParamColor specs.
Fixes:
> LibGimp-WARNING **: 16:06:09.451: _gimp_gp_param_def_to_param_spec: GParamSpec type unsupported 'GeglParamColor'
This is part of the fix to issue #10811, though it's not complete yet.
I still see some limitations in GimpGradient, and in particular, they are still
always stored as RGB in GGR files. It would be nice if we could store the actual
color format. This way, if someone chooses a gradient stop as Lab or CMYK color,
that's what the gradient file would keep track of. But also even storing the
space of a color (instead of storing/loading always in sRGB, even though this
may still work fine as we store unbounded double values). This might warrant for
a v2 of GGR file format.
This commit also fixes loading of SVG gradient which was apparently broken
regarding hexadecimal color parsing.
Finally I improve gegl_color_set_alpha() by adding an alpha channel when the
initial format had none.
The invasion extended to some core widgets too, in particular GimpColorPanel (a
subclass of GimpColorButton). There was quite a lot of code depending on these
widgets.
We pass 2 GeglColor through the wire now. Since it is passed very early
(when sharing the configuration), I had some issues with initialization
order of GEGL, and in particular when calling gegl_init() before
gegl_config() inside _gimp_config(), I had a bunch of such criticals:
> Plugin script-fu: GLib-GObject: CRITICAL: Two different plugins tried to register 'GeglOpPlugIn-transform-core'
Anyway in the end, I store the passed colors as raw bytes and strings in
the GPConfig object, and re-construct the GeglColor last minute in
_gimp_config().
One of the big improvement in this commit is that text layers are now much
better at space accuracy. They were already space-aware, yet rendered as sRGB u8
only before being converted to the image's space. It means that text layers had
the following limitations:
* Any color out of sRGB gamut were trimmed.
* Precision was always 8-bit (even if the image was high-bit depth).
Now GimpTextLayout keeps track of its source space (for RGB and CMYK only, this
won't be as easy when we will support more backend, since Cairo has only RGB
support for image data) and the image TRC (in case it bypasses the color space's
TRB) and it draws within this gamut and space.
It means first that we are not limited to sRGB colors; we will draw text main
color in the full image gamut, with still 2 remaining limitations:
* Unbounded colors are impossible because Pango format (to color text) uses
hexadecimal (so even with half/float images, you can't draw out-of-gamut text
unfortunately).
* Main color precision is still 8-bit, yet a tiny bit better than before as we
at least follow TRC (so we avoid some of the precision loss when converting,
even though the bit-depth is still the biggest loss).
The outline color on the other hand is drawn through Cairo API entirely, in
float. This means that the outline color will now be without any precision loss.
Note that this depends on CAIRO_FORMAT_RGBA128F which is only available since
Cairo 1.17.2 which is not in Debian bookworm (our current baseline for GIMP
3.0). It means that the old precision will still happen with older Cairo
version, as determined by #if code at compilation.
- app: gimp_context_get_(foreground|background)() are now returning a GeglColor.
- libgimp: PDB functions named similarly in libgimp are returning a newly
allocated GeglColor too.
- A few other PDB functions (the ones using these functions) were updated and
their signature changed to use GeglColor too, when relevant. Plug-ins which
use any of the changed libgimp functions were fixed.
- GimpContext: signals "(foreground|background)-changed" are now passing a
GeglColor.
- libgimpconfig: new macro GIMP_CONFIG_PROP_COLOR using gegl_param_spec_color().
- GimpContext: properties "foreground" and "background" are now GeglParamColor
properties.
- app: All code interacting with GimpContext objects were updated to receive a
GeglColor (that they may still convert, or no, to GimpRGB for now).
- app: gimp_prop_gegl_color_button_new() was added as an alternative to
gimp_prop_color_button_new() when the property is a GeglParamColor. Eventually
the former should replace completely the latter.
- libgimpwidgets: gimp_prop_color_area_new() now works on GeglParamColor
properties only.
- libgimp: gimp_procedure_dialog_get_widget() will generate a GimpColorArea for
GeglTypeParamColor arguments.
Also the color is internally stored as GeglColor, though there are still get
APIs and signals using GimpRGB.
The equivalent PDB functions are also changed to use GeglColor, same as app/
functions.
This is a first commit to really getting rid of GimpRGB within core and
PDB/plug-in code. This will make color conversion reliability a lot better as
GeglColor will handle conversions for us. The goal is that we should keep origin
color space (for instance when picking colors in a GimpPickable, or when storing
in the FG/BG colors or in paletters) until the last second and convert at use
only.
It's still very much work-in-progress.
These 2 functions were removed in commit 89c359ce. They were in fact used and
clearly this historical API seems interesting (though we can likely do the same
thing using the drawable GeglBuffer, but this way is much easier).
This is now reimplemented using GeglColor instead of raw data.
Eventually this is meant to fully replace GimpRGB (as well as GimpHSV, GimpHSL
and GimpCMYK), both in libgimp and in core code, as part of both the space
invasion and the API rework. For this first commit, I keep this new object side
by side to GimpRGB.
MINGW64
- uses 0x601 as value for _WIN32_WINNT. No need for us to define
it to that value or even lower values in some places.
This also gets rid of: warning: "_WIN32_WINNT" redefined
- has 0x0502 for WINVER, so get rid of us setting it to 0x0500 in
gimp-app-test-utils.h. It also seems that the need to use G_OS_WIN32
has disappeared here.
- DIRECTINPUT_VERSION is 0x0800, no need for us to set it to that value.
- AI_ADDRCONFIG was apparently missing from the MINGW headers in the
past, but not anymore.
- Do not leak allocated return value of gegl_node_to_xml_full().
- When merging layer effects, use gimp_drawable_filter_commit(), making
sure we use the exact same code path as when applying layer effects
destructively from the start. This also ensures that filters are
properly removed from the filter stack (unlike
gimp_drawable_merge_filter()), which was the reason why the rendering
was wrong (hence getting the buffer without effects first, then
reapplying it after was only a workaround to an actual bug).
- When removing a filter, verify the object still exists before doing
anything with it. If this was the last reference, we don't want to
call functions on this object. In gimp_drawable_filter_commit(), we
set up a weak pointer. In gimp_drawable_filter_remove_filter() itself,
we save the pointer to the filter's drawable before actual removal (as
we don't want to dereference a freed object later on).
- export_merge_layer_effects() should merge filters recursively through
layer groups.
- clean up the XCF code:
* No need to wrap the effect pointers list into 2 zero offset. Only
have one zero offset to indicate the list end.
* Add the layer effect mask in the effect structure (not in the layer
structure), similar as for layer masks.
* Effect name and icon made as main data in the structure, not as
properties.
This patch implements an initial form of
non-destructive editing. Filters now stay active
instead of being immediately merged down.
A new column is added to the layer tree view, which
can be clicked to show a pop-over menu.
Filters can currently be hidden/shown, edited, reordered,
deleted, and merged down from this pop-over menu.
Currently, this works on layers and layer selections only.
Plenty of room for improvement!
Resolves#10651
The "Remove All Guides" script calls
gimp-image-find-next-guide, which per
its description can take in 0. However,
the parameter sets 1 as the minimum
value.
This patch fixes the range so that it can
accept 0, which enables the Remove All
Guides script to work again.
It also updates the script to the new
multi-layer aware API.
Due to GObject Introspection we can't have the last part of an
identifier start with a digit, since that part will be used in Python
as the identifier, and Python doesn't allow that to start with a digit.
e.g. GIMP_ROTATE_90 would be used in Python as
image.rotate(Gimp.RotationType.90)
To fix this we add DEGREES in front of the number, without a '_',
even though that looks ugly.
These are not usable by plug-ins anymore which should store their data between
runs as arguments or aux arguments (in case of values which should be stored
from one run to another but are not really usable for non-interactive scripts).
These are per-plug-in (not polluting the whole process space with just random
strings as identifiers which could be used by other plug-ins) and even survive
restarts of GIMP.
I still keep these functions, but only internally, as they are used to store
settings of GimpAspectPreview, GimpDrawablePreview and GimpZoomPreview across
plug-in runs. Still I changed their API to set and return a GBytes directly
(mimicking the private PDB functions' API).
Also I remove gimp_pdb_get_data_size() which is useless when exchanging GBytes
directly.
Note that the 2 functions are still exported in the library, and only not
advertized through headers (so they are not really internal, just hidden), on
purpose, because we need to call them in libgimpui. So it is still relatively
easy for a plug-in to use them. Nevertheless I made clear in the function
documentation that these must not be considered public and could end up deleted
at any time. Any plug-in still trying to call these takes the risk of having
their code relying on unreliable API.
gimp_procedure_new_return_values() takes ownership of the passed GError (it
allows, among other things, to call it directly as return value). So we must not
try and free it afterwards.
The gimp_procedure_run() already existed, though it was with an ordered
GimpValueArray array of arguments. Its usage feels redundant to the series of
gimp_pdb_run_procedure*() functions (which is confusing), but
gimp_procedure_run() was actually a bit more generic, because it does not
necessarily calls GimpProcedure-s through the PDB! For instance, it can runs a
local GimpProcedure, such as the case of one procedure which would want to call
another procedure in the same plug-in, but without having to go through PDB. Of
course, for local code, you may as well run relevant functions directly, yet it
makes sense that if one of the redundant-looking function is removed, it should
be the more specific one. Also gimp_procedure_run() feels a lot simpler and
logical, API wise.
A main difference in usage is that now, plug-in developers have to first
explicitly look up the GimpPdbProcedure with gimp_pdb_lookup_procedure() when
they wish to call PDB procedures on the wire. This was done anyway in the
gimp_pdb_run_procedure*() code, now it's explicit (rather than calling by name
directly).
Concretely:
* gimp_pdb_run_procedure(), gimp_pdb_run_procedure_config() and
gimp_pdb_run_procedure_valist() are removed.
* gimp_procedure_run() API is modified to use a variable args list instead of a
GimpValueArray.
* gimp_procedure_run_config() and gimp_procedure_run_valist() are added.
* gimp_procedure_run_config() in particular will be the one used in bindings
which don't have variable args support through a (rename-to
gimp_procedure_run) annotation.
Passing (name, type, value) triplets is actually useless because we can get the
type information from the procedure/config anyway. That only adds one more
verification to do. Let's just change the function so that we pass (name, value)
couples instead, pretty much like in `g_object_set()`.
As far as plug-in API is concerned, at least the calling API, order of arguments
when calling PDB procedures doesn't matter anymore.
Order still matters for creating procedures with standard arguments (for
instance, "run-mode" is first, then image, or file, drawables or whatnot,
depending on the subtype of procedure), but not for calling with libgimp.
Concretely in this commit:
- gimp_pdb_run_procedure_argv() was removed as it's intrinsically order-based.
- gimp_pdb_run_procedure() and gimp_pdb_run_procedure_valist() stay but their
semantic changes. Instead of an ordered list of (type, value) couple, it's now
an unordered list of (name, type, value) triplets. This way, you can also
ignore as many args as you want if you intend to keep them default. For
instance, say you have a procedure with 20 args and you only want to change
the last one and keep the 19 first with default values: while you used to have
to write down all 20 args annoyingly, now you can just list the only arg you
care about.
There are 2 important consequences here:
1. Calling PDB procedures becomes much more semantic, which means scripts with
PDB calls are simpler (smaller list of arguments) and easier to read (when
you had 5 int arguments in a row, you couldn't know what they refer to,
except by always checking the PDB source; now you'll have associated names,
such as "width", "height" and so on) hence maintain.
2. We will have the ability to add arguments and even order the new arguments in
middle of existing arguments without breaking compatibility. The only thing
which will matter will be that default values of new arguments will have to
behave like when the arg didn't exist. This way, existing scripts will not be
broken. This will avoid us having to always create variants of PDB procedure
(like original "file-bla-save", then variant "file-bla-save-2" and so on)
each time we add arguments.
Note: gimp_pdb_run_procedure_array() was not removed yet because it's currently
used by the PDB. To be followed.
This partially revert some of the changes in commit 652a1b4388 because the
Windows CI suddenly failed because of this (my local build on Linux didn't have
any problem though) with:
> /usr/bin/x86_64-w64-mingw32-ld: libgimp/libgimpui-3.0-0.dll.p/gimpproceduredialog.c.obj: in function `gimp_procedure_dialog_save_defaults':
> /builds/GNOME/gimp/_build/../libgimp/gimpproceduredialog.c:2570:(.text+0x633): undefined reference to `_gimp_procedure_config_save_default'
> /usr/bin/x86_64-w64-mingw32-ld: /builds/GNOME/gimp/_build/../libgimp/gimpproceduredialog.c:2576:(.text+0x644): undefined reference to `_gimp_procedure_config_has_default'
> /usr/bin/x86_64-w64-mingw32-ld: libgimp/libgimpui-3.0-0.dll.p/gimpproceduredialog.c.obj: in function `gimp_procedure_dialog_load_defaults':
> /builds/GNOME/gimp/_build/../libgimp/gimpproceduredialog.c:2549:(.text+0xa2f): undefined reference to `_gimp_procedure_config_load_default'
> /usr/bin/x86_64-w64-mingw32-ld: libgimp/libgimpui-3.0-0.dll.p/gimpproceduredialog.c.obj: in function `gimp_procedure_dialog_constructed':
> /builds/GNOME/gimp/_build/../libgimp/gimpproceduredialog.c:368:(.text+0x11b1): undefined reference to `_gimp_procedure_config_has_default'
This is because these functions are used not only inside libgimp but also
across inside libgimpui. As a consequence, the build fails when linking
libgimpui.
This goes with our planned change of not making GimpProcedure arguments order
relevant anymore regarding the PDB API. In particular, it means we don't want to
use GimpValueArray for various procedure arguments API, but directly
GimpProcedureConfig objects.
This change will allow to add or reorder arguments in the future, so that we
won't have to create new PDB procedures when adding new arguments, while still
keeping PDB API stability.
Some of these should not even be visible by libgimp and were just fine as static
as well! For the rest, I make them really private (not only with a private
header).
We cannot be 100% sure generically (i.e. for all possible bindings available
with GObject Introspection) if bindings add their own reference to objects or
not. Clearly we have cases when they always do (Lua, Javascript), cases when
they do only in certain conditions (global Python variables) and cases when they
don't (Vala). What we know for sure is that in these script languages,
developers don't manually manage memory anyway. So the additional reference is
not their fact.
So let's just maintain a list of automatic memory managed binding languages,
among the few we officially support (i.e. the ones for which we have working
test plug-ins) and verify by executable extension if the plug-in is written in
one of these.
Both keeping a manually-updated list and verifying by extension are not so
pretty solution, but for now it will do.
As explained in the comment above, the reference might actually be owned by the
binding code (not by the plug-in code) and therefore can still be released
afterwards. Freeing it now while we don't own the reference exposes us to
double-free crashes.
Until now, it was not really possible to delete a colormap color, but since we
now use GimpPalette, people would definitely try to do so. It just makes sense
to allow doing this, but only if the color is unused.
Additionally when we do this, all the pixels refering to bigger indexes will be
edited so that they continue to refer to the same color (bigger indexes are
shifted by -1). Therefore removing an unused color does not change the image
render.
I wondered if we might want more options, e.g. the ability to delete a color
without fixing indexes (i.e. that colors over the deleted color index would
shift to the next color). This would even allow to delete used colors (though
now the last index would have to be unused one, unless we cycle colors).
Yet I don't think this should belong to this basic API. The most expected
behavior when deleting a color from an image colormap is to fix all indexes
stored in pixels so that the image still shows the same. So that's what this
function will do in this generic usage.
This is meant to replace gimp_image_get_colormap() (see also #9477).
We likely won't need a gimp_image_set_palette() because we can simply edit the
image's colormap/palette with GimpPalette API now and it is directly updated.
For instance, the following code changes the first entry in the image palette to
red, immediately:
```python
i = Gimp.list_images()[0]
p = i.get_palette()
c = Gimp.RGB()
c.r = 1.0
p.entry_set_color(0, c)
```
For this to work fine, I added a new concept to GimpData, which is that they can
be tied to a GimpImage (instead of a GFile). Image palettes are not considered
internals, they are just tied to their image, therefore they can be edited by
scripts/plug-ins.
Additionally with this commit, editing an image's colormap from libgimp API also
generates undo steps now.
bootchk had the case in commit 6781a35668. I again had it with gfig. I think it
just makes sense to init GEGL, especially as the errors are not that explicit
and that the plug-in code may not even call GEGL code directly (so it makes it
harder to guess).
It returns all the fonts (possibly more than 1) with a given name. I left the
function gimp_font_get_by_name() as a utility when one don't want to choose (or
is not able anyway, e.g. a script with minimal information), though I wondered
if we should not simplify with a single function (the new one, which is the
correct one now that it is possible to have several fonts with a given name).
It is easy to test with fonts named the same. For instance I could find 2
different fonts, both named 'Holiday'. This call in the Python console returns
both:
> Gimp.fonts_get_by_name('Holiday')
As part of this commit, I also implemented resource arrays (or subtype arrays)
as PDB arguments and return types.
… description.
- The returned value as width/height/etc. of the glyph extents (or bounding
box), not "of the font" (which doesn't mean much).
- Adding some definition for ascent and descent. This text is straight out
copied from Pango documentation comments in pango/pango-types.h.
- I don't see why we were negating the descent value. Let's keep the value sign
as defined in Pango.
- Fonctions were renamed: s/gimp_text_fontname/gimp_text_font/ and
s/gimp_text_get_extents_fontname/gimp_text_get_extents_font/
- The size_type arguments were removed. Even in 2.10, this argument was marked
as "dead" and ignored. It was only kept for API compatibility.
- The font name (string) was replaced by a GimpFont argument.
gimp_text_font() is easily tested in the Python console with:
> Gimp.text_font(Gimp.list_images()[0], None, 10, 40, "Hello World!", 1.0, True, 100, Gimp.context_get_font())
And gimp_text_get_extents_font() with:
> Gimp.text_get_extents_font("Hello World!", 100, Gimp.context_get_font())
Fixing:
> [809/2421] Generating libgimp/GimpUi-3.0.gir with a custom command (wrapped by meson to set env)
> libgimpwidgets/gimppropwidgets.c:37: Warning: GimpUi: multiple comment blocks documenting 'SECTION:gimppropwidgets:' identifier (already seen at gimppropwidgets.c:23).
This function is not perfect and in particular doesn't seem usable with binding
because of GimpUnit being some weird mix between an enum and some kind of class.
So this will have to be fixed too. See #8900.
… function gimp_font_get_pango_font_description().
Also updating file-pdf-save which is the only plug-in using these right now.
Note that I am not fully happy with the new function
gimp_font_get_pango_font_description() because I experienced some weird behavior
in file-pdf-save which is that some fonts were wrong if this is called after
pango_cairo_font_map_set_resolution().
But let's say this is a first step looking for improvements.
I am using the same GimpDrawableChooser with an additional drawable_type
argument to only show the appropriate tab if we want to limit what can be
chosen.
None of our plug-ins actually use a GimpLayer or GimpChannel only arg so far,
but if we have some day, or if some third-party plug-ins want to have such arg,
now they quite easily can!
After testing, setting a window as transient to another from another process is
still broken on Windows and it's hard to diagnose without using Windows
directly. Since it's not just broken, but it even hangs the whole process, which
is quite a blocker issue, let's disable again the whole code on Windows.
This was not working properly and needed some external build script as well as
the stamp/bogus header trick like for other similar in-source generated code.
In the same time, I get rid of old meson code which was meant for when using
meson < 0.57.0 (since our requirement is now meson >= 0.59.0).
I believe it should not happen in normal GUI case (which is when you create a
GimpProcedureDialog). I had the issue while moving around some plug-in code and
moved dialog creation before gimp_ui_init() by mistake. The issue was not
obvious until I followed the trace inside libgimp. This would be even more
frustrating for plug-in developers so let's have a clear warning message giving
the possible plug-in crash reason.
Similarly to the various GimpResource select PDB calls, this allows to call a
core dialog in order to choose a drawable which will be returned back to the
calling plug-in.
This new GimpPickableSelect dialog is a subclass of GimpPdbDialog and uses the
same GimpPickableChooser widget as GimpPickablePopup, except that since it's
inter-process window management, it is harder to make a popup positioned
accurately relatively to a parent (especially on Wayland). This is why it's a
separate widget as a simpler dialog (which we will still try to make transient
as much as possible across platforms).
This name was really irking me because it's not a button (anymore? Maybe it used
to be just a button). Depending on the specific widget, it will have several
sub-widgets, including a label. And it can theoretically even be something else
than a button.
So let's just rename these widgets with the more generic "chooser" name.
Similar to the latest commits for GimpBrush:
- gimp_pattern_get_buffer() returns a GeglBuffer and allow getting a scaled
version of the pattern.
- Old gimp_pattern_get_pixels() is made private.
- Moved GimpPattern into its own file and store the buffer to avoid re-querying
it through PDB continuously.
No as for the widget to select a pattern:
- Preview frame ensured to be square.
- Default size increased.
- Drawing code using the new gimp_pattern_get_buffer().
- Cleaned up code.
So what I realized was that the core was sending contents without transparency.
Actually the mask was our transparency channel here. Since in most use cases,
what you want to do when you request a brush buffer is to be able to draw it
somewhere, having a buffer already with alpha is much better, even more because
by default, it looks like background color is black which is possibly not what
you expect usually from a brush preview.
If someone wants absolutely no-alpha, it's easy to get rid of the channel. It's
simply better that the default behavior is the most expected use case.
- Increase the default size to 40x40 and multiply it by the current window scale
factor to have decent preview size.
- Make the brush preview always square with a GtkAspectFrame: even though
brushes are not necessarily square, this is a much more obvious size rather
than letting GTK choose a random allocation size which ends up very weird
looking.
- Scale down the brush to the biggest possible dimensions which fit the square
preview area (if the brush native size is already smaller, I don't scale up
though) while keeping aspect ratio: previous implementation was really weird,
as we were only seeing a tiny corner of much brushes as we weren't scaling
them down. Obviously I use new gimp_brush_get_buffer|mask() functions for
this as it supports scaling.
- Implement drawing color brushes too: the previous implementation was only
drawing the brush mask, which was absolutely not what would be expected for
such brushes.
- Add a white background behind color brushes with transparency.
- Simplify and clean up the code.
One of the consequences of this new implementation is obviously that it's
mandatory to call gegl_init() when using this widget.
… and gimp_brush_get_mask().
gimp_brush_get_pixels() was a bit crappy, returning raw data with only
dimensions and bpp to go with (no color model/space, no bit depth…). So the
assumption is that we work with 8-bit per channel data, possibly with alpha
depending of number of channels as deduced from bpp, and very likely in sRGB
color space. It might be globally ok with many of the brush formats (and
historical brushes) but won't fare well as we improve brush capabilities.
- gimp_brush_get_pixels() is in fact made private.
- The 2 new functions are using this old PDB call _gimp_brush_get_pixels() to
construct buffers. This has some limitations, in particular that it returns
only 8-bit per channel sRGB data, but at least the signature won't change when
we will improve things in the future (so if some day, we pass fancy brushes in
high-bit depth, the method will stay the same).
- This new implementation also allows scaling down the brush (keeping aspect
ratio) which is useful when you need to fit a brush preview into a drawing
widget.
- Current implementation stores the buffers at native size in the libgimp's
GimpBrush object, hence save re-querying the core every time you need an
update. This can be improved as current implementation also means that you
don't get updates if the brush changed. This should handle most common use
cases for now, though.
- Also with this change, I move GimpBrush class implementation into its own
dedicated file.
- Move the property widget functions for GimpResource properties into a new
libgimp/gimppropwidgets.[ch] file. This mirrors the files
libgimpwidgets/gimppropwidgets.[ch] which are for more generic property types.
- Rename the functions gimp_prop_chooser_*_new() to gimp_prop_*_chooser_new().
- gimp_prop_chooser_factory() doesn't need to be public.
- Add a label to GimpResourceSelectButton, make so that the
gimp_prop_chooser_*_new() functions set the property nick to this label and
add this label to the size group in GimpProcedureDialog.
- Removing useless or redundant code.
- Simplifying various logics.
- Using GimpResource directly in temporary PDB procedures, not resource names.
- Better cleanup of the core resource chooser when the plug-in dialog quits (we
need it to ask core to close also any visible resource chooser dialog).
- Replace the "Close" button by more common OK/Cancel. In particular, the
GimpPdbDialog now properly keeps track of the initial object and when hitting
"Cancel" (or Escape key), this initial object is set back.
- Clean up some of the comments, especially when the code is self explanatory.
There is still much more to clean and improve, but it's a first welcome step.
Found by the definitely useful libgimp warnings:
> gimp_plug_in_destroy_proxies: ERROR: GimpPattern proxy with ID 13 was refed by plug-in, it MUST NOT do that!
Brush, font, gradient, palette and pattern choices are currently chosen through
a dialog created by the core, which then returns the user choice to the calling
plug-in. This has the unfortunate consequence of having a pile of likely at
least 3 windows (main GIMP window by core process, plug-in window by plug-in
process, then the choice popup by the core process) shared in 2 processes, which
often end up under each other and that's messy. Even more as the choice popup is
kinda expected to be like a sub-part of the plug-in dialog.
So anyway, now the plug-in can send its window handle to the core so that the
resource choice dialog ends up always above the plug-in dialog.
Of course, it will always work only on platforms where we have working
inter-process transient support.
Instead of passing a guint32, pass the proper type, since our the HANDLE type
can be 64-bit on Windows (according to links I found).
I was hoping it might be the reason for the breakage under Windows, though I
also found Microsoft documentation saying that the 64-bit handle can be safely
truncated: https://learn.microsoft.com/en-us/windows/win32/winprog64/interprocess-communication?redirectedfrom=MSDN
Nevertheless I'd appreciate testing again from NikcDC or anyone else, as I
reactivated setting transient between processes on Windows.
Note that I also pass the proper types on X11 now (Window), even though guint32
worked fine. Better be thorough.
Having windows ID as guint32 is a mistake. Different systems have
different protocols. In Wayland in particular, Windows handles are
exchanged as strings. What this commit does is the following:
In core:
- get_window_id() virtual function in core GimpProgress is changed to
return a GBytes, as a generic "data" to represent a window differently
on different systems.
- All implementations of get_window_id() in various classes implementing
this interface are updated accordingly:
* GimpSubProgress
* GimpDisplay returns the handle of its shell.
* GimpDisplayShell now creates its window handle at construction with
libgimpwidget's gimp_widget_set_native_handle() and simply return
this handle every time it's requested.
* GimpFileDialog also creates its window handle at construction with
gimp_widget_set_native_handle().
- gimp_window_set_transient_for() in core is changed to take a
GimpProgress as argument (instead of a guint32 ID), requests and
process the ID itself, according to the running platform. In
particular, the following were improved:
* Unlike old code, it will work even if the window is not visible yet.
In such a case, the function simply adds a signal handler to set
transient at mapping. It makes it easier to use it at construction
in a reliable way.
* It now works for Wayland too, additionally to X11.
- GimpPdbProgress now exchanges a GBytes too with the command
GIMP_PROGRESS_COMMAND_GET_WINDOW.
- display_get_window_id() in gimp-gui.h also returns a GBytes now.
PDB/libgimp:
- gimp_display_get_window_handle() and gimp_progress_get_window_handle()
now return a GBytes to represent a window handle in an opaque way
(depending on the running platform).
In libgimp:
- GimpProgress's get_window() virtual function changed to return a
GBytes and renamed get_window_handle().
- In particular GimpProgressBar is the only implementation of
get_window_handle(). It creates its handle at object construction with
libgimpwidget's gimp_widget_set_native_handle() and the virtual
method's implementation simply returns the GBytes.
In libgimpUi:
- gimp_ui_get_display_window() and gimp_ui_get_progress_window() were
removed. We should not assume anymore that it is possible to create a
GdkWindow to be used. For instance this is not possible with Wayland
which has its own way to set a window transient with a string handle.
- gimp_window_set_transient_for_display() and
gimp_window_set_transient() now use an internal implementation similar
to core gimp_window_set_transient_for(), with the same improvements
(works even at construction when the window is not visible yet + works
for Wayland too).
In libgimpwidgets:
- New gimp_widget_set_native_handle() is a helper function used both in
core and libgimp* libraries for widgets which we want to be usable as
possible parents. It takes care of getting the relevant window handle
(depending on the running platform) and stores it in a given pointer,
either immediately or after a callback once the widget is mapped. So
it can be used at construction. Also it sets a handle for X11 or
Wayland.
In plug-ins:
- Screenshot uses the new gimp_progress_get_window_handle() directly now
in its X11 code path and creates out of it a GdkWindows itself with
gdk_x11_window_foreign_new_for_display().
Our inter-process transient implementation only worked for X11, and with
this commit, it works for Wayland too.
There is code for Windows but it is currently disabled as it apparently
hangs (there is a comment in-code which links to this old report:
https://bugzilla.gnome.org/show_bug.cgi?id=359538). NikcDC tested
yesterday with re-enabling the code and said they experienced a freeze.
;-(
Finally there is no infrastructure yet to make this work on macOS and
apparently there is no implementation of window handle in GDK for macOS
that I could find. I'm not sure if macOS doesn't have this concept of
setting transient on another processus's window or GDK is simply lacking
the implementation.
… is set.
The order for thumbnail creation in gimp_imagefile_create_thumbnail() is now:
1. If there is a GimpThumbnailProcedure, it is run first.
2. Otherwise we check if a thumbnail is in the metadata.
3. As last resort, we just load the full image.
Part of the fix was to copy gimp_image_metadata_load_thumbnail() into the core
code. I have been wondering if we could not drop the same function from libgimp
and remove the GimpThumbnailProcedure frome file-jpeg, since it just uses the
metadata thumbnail and it is the only plug-in using this code.
Also it is much faster to run this in core and it's generic function which makes
thumbnail loading from Exif data working for every format supported by Exiv2.
On the other hand, the file-jpeg thumbnail procedure also gathers a few more
useful information, such as the color model (in a reliably manner, since based
on JPEG header, unlike from metadata which may be wrong).
The various information (width, height, image type and number of layers) are
those of the full image, not of the thumbnail. Make it clear in the docs of
GimpRunThumbnailFunc.
Additionally:
- file-xmc was returning the proper information but variables were wrongly
named, which was confusing.
- Fix file-ico thumbnail proc which was returning the thumbnail width/height.
- In file-darktable, initialize width/height to 0 so that we just don't show any
size when we don't get the information. It's better not to show anything than
completely wrong information (the thumbnail target size).
… than a GimpValueArray.
Similar to other GimpProcedure, move to using a config object. A difference is
that thumbnail procedures are always run non-interactively.
Also fixing WMF load thumbnail procedure: the dimension computation was wrong
when the image was wider than tall.
… a GimpProcedureConfig for arguments.
This also factorizes the code to load metadata. By default, a GimpLoadProcedure
will try and load metadata from a file (if Exiv2 knows the format). The run()
function will be allowed to edit the GimpMetadata object but also the load flags
before it is actually attached to the image, allowing plug-ins to have custom
metadata handling code when needed.
This is just a method to simplify transforming a GimpChoice argument into an
enum value, which is easier to deal with, in C. It also allows to benefit from
switch() warnings or the like to make sure no cases are missing.
Developers won't have to maintain manually a list of the possible values in the
help string. It can now be generated from the GimpChoice and will be therefore
ensured to always be up-to-date, and nicely formatted.
I also add some pango markup to the type helper texts to differentiate it from
the main argument docs.
These will replace the int arguments used in place of enums. The problem of int
arguments used as list of choices is that it makes calling PDB functions very
opaque. This is especially bad when a list is long, so you constantly have to
refer to the documentation to understand what a series of numbers mean in
argument lists.
And the second issue is that plug-in developers have to manually maintain a list
of values both in the GUI and in the documentation string. This help text may
get out-of-sync, may end up with missing values or whatnot. Also if it is used
as tooltips, it makes for very weird tooltips in the graphical interface, with
an overlong technical list of int-values mapping which should ideally only be
made visible in the PDB procedure browser listing.
Now gimp_procedure_dialog_get_label() can work both with an existing property ID
or a new property ID. In the former case, it will simply sync the label with the
procedure argument, which will make it easy to update the label contents. In the
latter case, it just initialize with the provided text.
The expectation of 2 references per object in gimp_plug_in_destroy_proxies() was
wrong. It is true during most of the plug-in life, because both the
GimpProcedure and the GimpPlugIn have a hash-table keeping their own reference
to it, except that in gimp_plug_in_pop_procedure(), we release the reference
owned by the procedure with _gimp_procedure_destroy_proxies() first. So at this
point of the object life, its reference count is supposed to be 1.
The source of the bug was in fact in _gimp_plug_in_get_*() (where * can be
display, image, item or resource) which was behaving differently the first time
it is called for an object with the successive calls. In the first call only, it
was creating then refing into the table (so the object started directly with 2
references) whereas on successive calls, it just returned the hashtable-looked
up reference. In other words, it behaved as a (transfer full) on the first call
and (transfer none) on successive calls. And so did all public API which were
making use of this infrastructure (in particular gimp_*_get_by_id() functions).
The widget_creator_func() given to gimp_prop_chooser_factory() will create an
object which will take its own reference. We must release the one we got with
g_object_get().
Metadata handling is also integrated in this API:
* while giving a possibility to disable metadata saving if you want to do it
yourself (e.g. in file-heif), by setting a NULL MimeType;
* and the GimpMetadata object is added as run() argument, allowing one to edit
the metadata during the run, while still letting the infrastucture handle the
save (e.g. in file-jpeg);
* or to save intermediate metadata with gimp_procedure_config_save_metadata()
(e.g. in file-tiff).
While we definitely should not use this inside app/, because having the private
structure easily accessible as a member is very convenient, it is clear that it
makes for a much nicer public signature. Also the priv member is of no help to
third-party developers using this API to make plug-ins and is better hidden in
such a case.
This will be useful for plug-in developers but also for us. Seeing we leak the
config object is often a good indication that something is wrong in our handling
of internal references (since everything relies on the config object in plug-ins
now, in particular all the GUI).
I was clearly confused when I wrote this. The sinking part matters to take
ownership of a reference in the widgets table, but we don't need to ref widgets
again before inserting them in containers. We were leaking widgets and as a
consequence the config object (and as a second consequence, some objects such as
resources for resource-selection widgets).
Same as with gimp_procedure_new2(), I will end up renaming the function to
gimp_image_procedure_new() once all usage of this function will have been ported
to the new function using GimpProcedureConfig instead of GimpValueArray
arguments.
The one in GimpResourceSelect is a very nice example of why using config objects
is much nicer, getting arguments by their name instead of an index (which in
this case had to be tracked down by a private function to handle different
cases).
Also in gimp_procedure_config_begin_run(), make sure we sync the arguments with
the config object first thing, even in interactive and with-last-vals case
(where the args may be further overridden). This was especially important for
Script-fu scripts as the image and drawable were not provided separately, so we
need to make sure that the config file has the right values.
Otherwise we will always try to reuse previous values or use the default,
bypassing the actual passed values.
I encountered this issue while porting file-glob and realizing that the
"pattern" argument was always passed to NULL, ignoring the explicitly set
pattern.
When a procedure has no run-mode argument, we should simply not assume anything
and use the passed arguments (which is what the non-interactive mode does).
This function allows to change the sensitivity of a widget depending on the
value of another property.
We already had gimp_procedure_dialog_set_sensitive() except it was only syncing
with a boolean property, whereas the new function can compare with any property
type.
This new function is meant to replace gimp_procedure_new() when all plug-in
usage will have been switched.
This function creates the GimpProcedureConfig object on behalf of the plug-in
and calls gimp_procedure_config_begin_run() and gimp_procedure_config_end_run().
This way we ensure that all plug-in calls with successful result are properly
stored without asking the developer not to forget to call these (if a "good
practice" is in fact something we request to do every time, especially for good
user experience, we might as well make it rather a core process).
Advantages:
* Better interactive experience: using any plug-in will result in saved
previously used settings.
* for developers, working on config objects is also much more comfortable than
working on GValueArray;
* step forward for the future macro infrastructure: if we can ensure that all
plug-in calls are properly logged, then we can replay plug-in actions, in
NON_INTERACTIVE with the same settings.
from plug-in
I get a critical error when calling gimp_text_layer_new and the function
doesn't return a GimpTextLayer:
(file-psd:47120): LibGimp-CRITICAL **: 16:00:59.035:
gimp_gp_param_to_value: type name GimpTextLayer is not registered
Adding the above line to libgimp/gimp.c fixes the problem.
Looking further, the @help is only used in gimp_proc_view_new() so far (for the
Procedure Browser) where the blurb and argument descriptions are already
localized. It makes no sense to only keep this in English. So let's ask to have
both arguments translated.
Now clearly we should not ask for @help to be mandatory. Very often, it makes no
sense to have a longer help string (the small blurb and the few arguments may be
very self-explanatory). So I make this argument nullable.
There is only the @help_id which I wonder if we could not have a simpler
function gimp_procedure_set_documentation_uri(). Indeed while having a unified
infrastructure with a XML summary and help IDs and whatelse makes sense for GIMP
as a whole, I think that many third-party plug-ins would work much better with a
very simple direct URL. Or it could even be a GFile to a local file (for
plug-ins which want to embed their documentation in the plug-in folder for
instance). To be continued…
I add a new class method deserialize_create() to GimpConfigInterface which
returns the GimpConfig object per deserialization, instead of modifying an
existing bare object.
This matters for cases like our GimpResource (and later our GimpItem) classes
which are fully managed by libgimp and should be unique objects per actual
resource. It should even be possible to compare the pointer itself for identity.
That's why we need to let GimpResource create the object (in reality request it
to the infra and only ref it) through this new class method.
With this commit and the previous ones, all GimpResource are now properly stored
as plug-in settings (e.g. the "film" plug-in has a font setting which is now
properly remembered).
These identifiers are not portable (across various installations and therefore
not for XCF either), but at least they are reasonably identifying data on a same
installation (unlike GimpResource's int ID which is only valid within a single
session) which makes them very fine for plug-in settings storage.
When a data file disappears, we fallback to the context default data instead.
This finds the core resource knowing its type, name, collection and internal
state (in other words, the values returned by _gimp_resource_get_identifiers()).
Rather than reimplementing the same checks for every possible resource data
type, just do it once and redirect to the correct factory container.
For the libgimp API, we leave per-type functions `gimp_*_get_by_name()` (where *
can be brush|gradient|font|palette|pattern so far), but internally they all use
gimp_pdb_get_resource().
Note that eventually we want these functions to return a list of resources as it
should be possible to have several resources of a given type with the same name
(since they are made by third-party who might have had the same idea of a name).
We use US English which uses behavior. So we replace all occurrences of
behaviour.
Most notable is File Open behavior in preferences. Besides that several
mentions in function documentation and a few in comments.
Resources are stored by the plug-in infrastructure and their memory should not
be managed by plug-in code.
My commit 4f69995b46 was crappy and modified a generated function. I was just
too tired with all the heat in here, I guess!
Similarly to how we handled image items, all resources are handled and stored by
the plug-in infrastructure and should not be destroyed. This wrong annotation
was triggering bindings to unref resources when going out of scope, hence
crashing plug-ins.
Since commit 89c359ce47, it's supposed to return a GBytes, yet it was returning
raw data (probably intermediate hacking state which was not properly cleaned
up).
It was crashing for instance GimpDrawablePreview widgets at drawing time.
It does absolutely nothing except sitting there, providing an is-a
relation (both ways because GimpData is its only subclass). This will
simplify having more libgimp API on GimpResource, without having to
add different PDB code for app and libgimp.
Much like for images and items. Change the PDB to transmit IDs
instead of names for brush, pattern etc. and refactor a whole
lot of libgimp code to deal with it.
modified: libgimp/gimpplugin-private.h
and Plug-in/Procedure Browser size.
gtk_widget_set_size_request () is applied to the dialog, as it does not
seem to bubble up from being applied to individual elements.
The existing 2.10 width/height values are used.
GLib has a specific type for byte arrays: `GBytes` (and it's underlying
GType `G_TYPE_BYTES`).
By using this type, we can avoid having a `GimpUint8Array` which is a
bit cumbersome to use for both the C API, as well as bindings. By using
`GBytes`, we allow other languages to pass on byte arrays as they are
used to, while the bindings will make sure to do the right thing.
In the end, it makes the API a little bit simpler for everyone, and
reduces confusion for people who are used to working with byte arrays
in other C/GLib based code (and not having 2 different types to denote
the same thing).
Related: https://gitlab.gnome.org/GNOME/gimp/-/issues/5919
Certain XMP metadata tags currently can't be saved by us until we get
support in gexiv2 for adding new structs not present in exiv2.
We remove these tags from the exported metadata because the XMPSDK in
exiv2 would otherwise fail to write all XMP metadata.
Examples are the newer (2019, 2021, 2022) sample images from the
IPTC Photo Metadata Standard.
We also add all the static functions present in gimpimagemetadata-save
at the top, which had been forgotten in the past.
Jehan