AFAIU the issue which it tried to fix only happens on custom builds and it looks
to me that it's just about multi-prefixes developer builds which were not
configured with -Drelocatable-bundle=no as they should.
When you compile GIMP yourself and intend to run it in a multi-prefix
environment, without ever relocating the build, like on Unix-like systems, then
disable this feature.
In any case, we should never do this kind of system file moving hacks in our
build system. The MSYS2 project even had to patch this out:
https://gitlab.gnome.org/GNOME/gimp/-/merge_requests/1171#note_2022406
The indexed conversion code was originally written for very low-memory systems,
this makes the mapping it does quantized and noisy, for examples see issue
#5159 this commit removes quality/memory use trade-offs, modern machines can
take a 2-300mb temporary table allocation.
This is similar code to what I added for the Windows builds, except that here
it's because deps-debian-x64 will also end up feeding dev-docs job too.
Right now our tagged build of GIMP produces docs for unstable versions of babl
and GEGL.
Since !1171 we can use the CI scripts locally, but the initial implementation,
after some commits, revealed to be troublesome. So, now reworked to be simple.
As noted by Jehan, the way of defining the (many) GIMP versions can change.
So, let's take the GIMP version directly from the generated config.h file.
Also, changes the declaration of some variables to conform with "best pratices".
This list is referenced in the gimp-web-devel site, and my initial
implementation, in terms of ease to use, was a throwback.
Also, make the CI code more simple without changing the list file.
As noted in a comment on issue 10873,
choosing Selection for Fixed Gear throws
a TypeError. This is because with the change
from GimpRGB to GeglColor, we can no longer
directly access the red component with an index.
Instead, we need to call Gegl.Color.get_rgba() to get
that value, which is what this patch does.
Resolves#10882.
Before the color space invasion, INPUT_COLOR () returned
either a GimpRGB or NULL (if alpha == 0) that was used
directly in functions.
Now it's used to set a GeglColor in gegl_color_set_pixel (),
and calling that with a NULL color resulted in a Gegl-Critical.
This patch checks if INPUT_COLOR () returned NULL before
trying gegl_color_set_pixel () to prevent the issue.
The alternative solution would be to call:
> image_new_dialog_set (dialog, NULL, NULL);
This would have reset to default template. But it's still not exactly like the
"new image" action which defaults to the active image's size if there is an
opened image.
In order to avoid complicating the code further, as well as code duplication,
hence code divergence, let's call the "image-new" action, making sure that this
button will always behave exactly like the "File > New" menu item.
Then since we need to process the return value of this dialog (either re-showing
the welcome dialog in case of new image creation cancelation, or destroying the
hidden welcome dialog otherwise), I check for the singleton pointer and connect
the handlers to it.
Moreover I made the "response" signal of ImageNewDialog be handled as
G_CONNECT_AFTER otherwise we nearly never had the possibility to handle its
responses properly (because the base handler was doing it first, then often
destroying the dialog).
The rest of the time, we want to use HEAD of their main branches.
The last ctx update seem to have broken the CI build on Windows. While we'll
have to fix this, let's do this after we release GIMP.
Anyway we obviously don't want to use unstable code for actual point releases,
so it's a good thing that this made me think of adding such rules.
Use the same amount of dithering on R,G and B if they initially were all equal.
This allows us to keep achromatic gradients achromatic, while providing the
same amount of dither as before for colored gradients. Discrepancy uncovered in
issue #10756.
While the buffer used by Cairo may have a rowstride bigger than 3 * width, our
internally stored buffer has the exact size required for drawing the scale.
We were not properly duplicating the left and right colors when duplicating the
existing segments in gradient_editor_save_selection(). Therefore we ended up
freeing original GeglColor on exiting the GimpColorDialog.
Additionally to the reported crash, I add a few GeglColor freeing which were
leaked and some safeguard in gimp_gradient_get_flat_color() to ensure it always
returns a color (or raise a CRITICAL to be investigated easily).
Finally a few signals in GimpColorDialog had to be ensured disconnected when the
dialog is destroyed.
