… gimp_image_policy_color_profile().
These functions allow a plug-in to explicitly execute the Rotation and
Profile conversion policies on an image (which may be any of
Rotating/Discarding/Ask or Converting/Keeping/Ask respectively). These
policies are automatically executed when loading an image from GIMP
interfaces, but they won't be when loading an image from the PDB. Then
it is up to the calling code to decide what to do (which can be either
some arbitrary code or following the user policy).
Orientation is now handled by core code, just next to profile conversion
handling.
One of the first consequence is that we don't need to have a non-GUI
version gimp_image_metadata_load_finish_batch() in libgimp, next to a
GUI version of the gimp_image_metadata_load_finish() function in
libgimpui. This makes for simpler API.
Also a plug-in which wishes to get access to the rotation dialog
provided by GIMP without loading ligimpui/GTK+ (for whatever reason)
will still have the feature.
The main advantage is that the "Don't ask me again" feature is now
handled by a settings in `Preferences > Image Import & Export` as the
"Metadata rotation policy". Until now it was saved as a global parasite,
which made it virtually non-editable once you checked it once (no easy
way to edit parasites except by scripts). So say you refused the
rotation once while checking "Don't ask again", and GIMP will forever
discard the rotation metadata without giving you a sane way to change
your mind. Of course, I could have passed the settings to plug-ins
through the PDB, but I find it a lot better to simply handle such
settings core-side.
The dialog code is basically the same as an app/dialogs/ as it was in
libgimp, with the minor improvement that it now takes the scale ratio
into account (basically the maximum thumbnail size will be bigger on
higher density displays).
Only downside of the move to the core is that this rotation dialog is
raised only when you open an image from the core, not as a PDB call. So
a plug-in which makes say a "file-jpeg-load" PDB call, even in
INTERACTIVE run mode, won't have rotation processed. Note that this was
already the same for embedded color profile conversion. This can be
wanted or not. Anyway some additional libgimp calls might be of interest
to explicitly call the core dialogs.
A log error can have a NULL domain (apparently equivalent to "" default
domain, according to g_log_set_handler() docs and we even explicitly
list the NULL domain in the log_domains array in app/gimp-log.c.
Yet our log handler was not expecting such possibility and was running a
g_str_has_prefix() on NULL. Not sure why it aborted there. It might be
because outputting a new warning inside the warning handler did not go
well. Anyway this seems to fix our side of the bug #5358. The main fix
will likely be on GEGL side (UMFPACK_ERROR_out_of_memory error).
Negative int values were not correctly handled because value.v_int is unsigned
causing a conversion to a large positive value.
To fix this we cast it to gint64 first before making it negative.
It seems to be an old file move error I did 2 years ago (commit
1e5cf10585). I wonder how I only realize this now. Anyway let's remove
this lonely file.
It's done, all Python plug-ins have been either ported to the new API +
Python 3, or they have been discarded (and moved to gimp-data-extras for
whoever wants to salvage them).
Let's get rid of the outdated pygimp directory (whose code has not been
built in the master branch for years now anyway)! Woohoo!
shadow_bevel.py and sphere.py are unfinished and have been in the
"Unstable" only builds for years. This is probably not worth porting
them. Let's just delete them.
I have actually move these (as well clothify and whirlpinch removed
yesterday) to the gimp-data-extras repository so if anyone wants to
revive them, and port them to GIMP 3, they can start off from there.
Following Elad Shahar advice (cf. #4368), let's delete whirlpinch.py and
clothify.py.
The rational is that whirlpinch.py does the same thing as the existing
"plug-in-whirl-pinch" which is itself a compat plug-in to
"gegl:whirl-pinch" operation. Also the Python file has an explicit
command saying it is exactly the same algorithm, yet with no preview and
slower. Finally it was not even installed on stable build. It doesn't
look like there is any reason to keep it (it was probably a demo/test
Python plug-in).
As for clothify.py, a quick look at the short code shows it is exactly
the same algorithm as clothify.scm, with the same arguments and
installed on the same `Filters > Artistic` menu (except that the Python
version is not installed on stable builds).
So let's just keep the script-fu version as it has been the used version
until now, and there is no deprecation going on in one side or another.
So let's keep what already works.
Since the patch was initially contributed, some parts of the
introspection changed. First all GUI-related code is in a GimpUi module
now.
Also Gimp.get_pdb().run_procedure() is now using a list instead of a
GimpValueArray.
Note by reviewer (Jehan): merging this port as it was in GIMP 2.10
anyway, but is this even still needed code? This plug-in is not even
available on stable release, it looks like for-development only
benchmark, and I'm not sure if it's relevant anymore, especially in our
GEGL-fueled new world.
Please anyone who knows a bit more on the history of this plug-in and
the evolution of our gimp_drawable_foreground_extract() algorithm, feel
free to weigh in and tell us what this was for exactly and if it's still
relevant.
This is a basic port without any UI.
Invoking the plugin will just sort the entire palette based on
default parameters.
The original plugin had several broken options, which I tried to fix.
Plug-ins that work from different bindings probably want to use their
own list-type to specify arguments, rather than working with a more
cumbersome `GimpValueArray`.
This new API should make it less verbose. For example:
```
args = Gimp.ValueArray.new(5)
args.insert(0, GObject.Value(Gimp.RunMode, Gimp.RunMode.NONINTERACTIVE))
args.insert(1, GObject.Value(Gimp.Image, image))
args.insert(2, GObject.Value(Gimp.Drawable, mask))
args.insert(3, GObject.Value(GObject.TYPE_INT, int(time.time())))
args.insert(4, GObject.Value(GObject.TYPE_DOUBLE, turbulence))
Gimp.get_pdb().run_procedure('plug-in-plasma', args)
```
becomes
```
Gimp.get_pdb().run_procedure('plug-in-plasma', [
GObject.Value(Gimp.RunMode, Gimp.RunMode.NONINTERACTIVE),
GObject.Value(Gimp.Image, image),
GObject.Value(Gimp.Drawable, mask),
GObject.Value(GObject.TYPE_INT, int(time.time())),
GObject.Value(GObject.TYPE_DOUBLE, turbulence),
])
```
The script `create_test_env.sh` was registered in meson as a run target
(i.e. to be run manually by `ninja create_test_env`), which is really
not useful. So a `ninja test` was outputting various:
> You have a writable data folder configured (/gimp/build/dir/app/tests/gimpdir-output/gradients),
> but this folder does not exist. Please create the folder or fix your
> configuration in the Preferences dialog's 'Folders' section.
Unfortunately run target are only meant to be run from command lines and
cannot be used in 'depends' argument of test() or 'dependencies' of
executable() because "in Meson all dependencies are to output files, not
to concepts" (cf. https://github.com/mesonbuild/meson/issues/1793).
So instead a run_target() just directly use a run_command() and make
this script run during configuration step. Also make the shell script
executable as it was not.
See also #5666 as it was one of the errors outputted by the reporter's
log (though probably not the main issue).
The XPM export plugin was saving the color `None` (transparency) to the
exported image palette for all images with an alpha channel, even if the
color was not used on any pixel.
Since it's nice to have `None` as the first color on the palette, mapped
to character ' ', and it could be too wasteful to scan all pixels twice,
the approach taken was to just undo it's premature insertion.
(cherry picked from commit 0a9bb2839e)
Existing code was returning only 1 out of 2 expected procedure names.
See !241 for the originally proposed patch by Rafał Mikrut (@qarmin).
Unfortunately the proposed patch also had another bug and was always
returning an empty list instead, and the contributor is unresponsive. So
let's just redo the patch and be done with it.
Also using g_list_prepend() instead of g_list_append() as it's usually
more efficient (to be fair here, for a size-2 list, it doesn't matter
much but it's good practice anyway) and order doesn't matter.
Thanks Rafał Mikrut for noticing the original bug!
The rotation was actually applied but the image had an orientation
metadata stored which is not visible in GIMP canvas (only checked at
import).
If GIMP had on-canvas viewing support of the orientation metadata, then
it would make sense to keep it between import and export, but since it
doesn't, we should assume that when someone asks to "Keep Original"
during import, they are actually asking to drop the metadata (which
actually can be wrong in various cases, in particular when you snap
pictures of the ground or the sky, then sensors are lost anyway and
regularly can't guess what orientation you wanted). This will make for
less unexpected exports.
Removed the WinMain specific stuff for the plug-in and determine
hInstance in twainMain which made porting a lot easier.
The _DEBUG enabled functions I have mostly left alone and
should be revised at a later time.
The proposed defaults for export should be the less destructive
possible. So for any 16 or 32 bpc GIMP images, since our HEIF plug-in
only supports up to 12 bpc, this should be what we export to.
… and have gimp_prop_int_combo_box_new() bind to "value" instead of
"active".
The "active" property is defined by GtkComboBox and is the index of the
combo box, not its values, whereas with gimp_prop_int_combo_box_new(),
we want to bind an int property to the combobox value. Therefore the
commit 0828a371c2 was only properly working when we were creating a
combo box with values starting at 0 and incremented by 1.
By adding a "value" property to GimpIntComboBox, I allow binding any
property to the int value rather than the index.
See also !265 where the issue was raised as it affected our HEIF
plug-in.
gimpimagemetadata.[ch] was built into libgimpui because GTK+ was used
for dialog query for rotation metadata. gimpimagemetadata-save.c only
was built into libgimp, which made no sense as the declaration for its
public function was inside gimpimagemetadata.h!
That was a weird situation and somehow only made visible in the build
system because GIR build was complaining about missing annotations to
gimp_image_metadata_save_prepare() (the annotation was actually present
but in the implementation which was not in the same library as the
header, how weird!):
> Warning: GimpUi: gimp_image_metadata_save_prepare: return value: Missing (transfer) annotation
Moreover it means that only plug-ins linking libgimpui had access to the
gimp_image_metadata*() API, which is obviously not cool (that should be
a core API).
Instead I moved everything into libgimp and replaced
gimp_image_metadata_load_finish() with
gimp_image_metadata_load_finish_batch(), which is essentially the same
function except that it's not interactive (it will proceed to rotate the
image without user confirmation, provided the right flag is present).
Then I add gimpimagemetadata-interactive.[ch] which contains only
gimp_image_metadata_load_finish() and is the alternative interactive
version of gimp_image_metadata_load_finish_batch(). Most plug-ins won't
even have to be changed (at least none in core GIMP) and would still
work as before, whereas now a non-interactive version exists, which
doesn't mandate to link GTK+.
Current dev code of GEGL is necessary as it fixes its VAPI dependency
(see gegl!83). Without this, with a recent meson version, Vala plug-in
build fails.
See !334 for some more background.