This results in an infinite loop and there is no change done, so no reason to
flush anything. Mitch had a fix for this (probably the same, per the IRC
discussion), but I really wanted to just release GIMP 2.99.16. So Mitch, if you
read this, sorry for not waiting for you to push this one!
We already search for a compatible Python version in the root meson file, no
need to look up Python 3 again in the PATH, each time we run an external Python
script in the build.
This should hopefully fix#9687.
See: https://gitlab.gnome.org/GNOME/gimp/-/issues/9653#note_1780587
Looking at MSYS2 logs, it looks like they very recently "fixed" the search paths
for lua files, which in turn broke our workaround (searching in subdirectories
of bin/ instead of share/ and lib/).
This should work better (though untested) with the workaround removed now.
Commit on luajit package at MSYS2:
703c7bae2f
In commit 48c27770 some unicode related changes were made. As a result of
that on Windows display_device, which was previously a duplicated string,
is now referenced directly.
However, the g_free was not removed, causing a crash.
We resolve this by removing the obsolete g_free.
Switching to the cfitsio library's fits_read_img () function in 03ea666f
caused RGB FITS images to be read in upside down per SiriL developers and
NASA standards (odd, since cfitsio is maintained by NASA). This flips them back.
The Microsoft style guide and other places online suggest that using a
hyphen is recommended between a number and the related word, when modifying
a noun, so let's do that in our Image -> Encoding submenu too.
The developers of SiriL sent me some RGB FITS images that rendered incorrectly
after d4f42076. These must be read in as
planes rather than individual rows.
The creation of the BMP welcome images for the Windows installer (part of
-Dwindows-installer=true build option) fails in the Windows job. After much
debugging, I could run GIMP, yet it was not enough. One of my hypothesis so far
is that the environment variables for DLLs won't work, since all the DLLs must
be in the same directory as the main binary (though with the WSL thing, I am
unsure, maybe it is still supposed to work), which only happens once GIMP is
installed. So GIMP runs successfully but not plug-ins.
Anyway I wasted too much time working on this and without a local Windows, it
just takes too long (mostly testing thanks to the CI) and is frustrating. Let's
just move to building both the localization files and the images on the main
Debian job (gimp-meson-debian), then use these as dependencies of the
win-installer-nightly job, i.e. when building the installer.
The common order logic for list of directories in environment variables is that
left paths have precedence. This is at least the case for LD_LIBRARY_PATH (and
probably GI_TYPELIB_PATH too).
Make sure that our local libraries and introspected binaries (in the build
directory) are used and not any version installed on the system or by previous
`ninja install` calls.
This was a first attempt at fixing this error on the CI:
> Cannot open display:
Though it was not enough (see next commit calling plug-ins as non-interactive
when called without interface), it is still a useful change overall.
We had the following warning:
In function 'make_remap_table',
inlined from 'gimp_image_convert_indexed' at
../../gimp/app/core/gimpimage-convert-indexed.c:1057:7:
D:/msys64/mingw64/include/glib-2.0/glib/gmem.h:261:19: warning: argument 1
range [18446744071562067968, 18446744073709551615] exceeds maximum object
size 9223372036854775807 [-Walloc-size-larger-than=]
This is apparently caused by inlining in combination with using a signed
int.
See also: https://gcc.gnu.org/bugzilla//show_bug.cgi?id=85783
Casting to (guint) silences the warning here.
… being installed.
There is already most of the main code logic for this, though now plug-ins need
to be in their own subdirectories, which breaks for plug-ins/common/ and
plug-ins/python/, while I needed plug-ins in both these categories to generate
the Windows installer welcome images (file-png, and python-fu-eval in
particular).
Once again, meson was not very helpful, since all its functions still refuse to
output generated files in subdirectories, so I end up duplicating plug-in files
with a custom Python script.
This should fix the CI. It was working on my machine as GIMP was installed, but
such a build rule should work even without GIMP installed.
This will also be useful in the future when we'll want to run unit tests of
plug-ins through the finale GIMP binary itself.
After discussion with Jernej, InnoSetup should now work better with rescaling
a big image properly to the window size, yet the ratio should still matter.
Apparently the welcome image is a hack and this is why it requires specific
ratio images. We don't use the big size yet, but since Jernej told me which
dimensions are expected, I already added the code for it to make it easier
later.
So anyway this code would allow us not to have to commit welcome images each
time, which are basically resized copy in BMP of the splash screen, slowly yet
surely filling up our repository with image duplicates.
After all, we develop a scriptable image editor! We should use it to edit images
and export in expected formats!
I only use this script for the devel installer for now, for testing and see how
it goes.
While some packages may be needed only when building (and others only when
packaging), we should probably have a shared list of packages needed for both
steps so that we avoid discrepancies which lead to missing libraries in our
installer.
See: https://gitlab.gnome.org/GNOME/gimp/-/issues/9653#note_1777596
Jehan