Drop the glib-dbus dependency and port stuff to GDBus. Use it
unconditionally if not on Windows or OSX. It seels to work just like
before but I have no idea if everything is really correct.
This is a basic implementation of an OpenEXR loader. This
"infrastructure" is required for any further work. It consists of:
* The build system changes.
* A C wrapper around the OpenEXR library, which is necessary as it's not
possible to intermix GIMP's code with C++ code.
* A basic image loader. Chroma is not supported currently, and some
other weird files like multi-view files are unsupported. These can be
added when necessary. There is no UI, but it should be straightforward
to add new features like this on top of this work.
Without the dependencies of file-compressor (currently zlib, libbzip2
and liblzma), configure fails now. file-compressor is referenced in too
many places for its functionality to be optional.
Recent Cairo uses SHM transports when available, and exposes the ability
for its users to manage images shared between it and the display.
This allows us to eliminate copies, and if the architecture supports it
even to upload directly into GPU addressable memory without any copies
(all in normal system memory so we suffer no performance penalty when
applying the filters). The caveat is that we need to be aware of the
synchronize requirements, the cairo_surface_flush and
cairo_surface_mark_dirty, around access to the transport image. To
reduce the frequency of these barriers, we can subdivide the transport
image into small chunks as to satisfy individual updates and delay the
synchronisation barrier until we are forced to reuse earlier pixels.
Note this bumps the required Cairo version to 1.12, and please be aware
that the XSHM transport requires bug fixes from cairo.git (will be
1.12.12)
v2: After further reflections with Mitch, we realized we can share the
transport surface between all canvases by attaching it to the common
screen.
v3: Fix a couple of typos in insert_node() introduced when switching
variables names.
v4: Encapsulating within an image surface rather than a subsurface was
hiding the backing SHM segment from cairo, causing it to allocate
further SHM resources to stream the upload. We should be able to use a
sub-surface here, but it is more convenient to wrap the pixels in an
image surface for rendering the filters (and conveniently masking the
callee flushes from invalidating our parent transport surface).
Cc: Michael Natterer <mitch@gimp.org>
It will never hold high bit depths using JPEG compression, and nobody
is going to port it to layer groups and whatever either. Wolfgang
says it's obsolete, whoever needs to convert old files can use 2.8.
New configuration directory scheme, consistent across platforms, and
following standards.
UNIX platforms (except OSX): $XDG_CONFIG_HOME/GIMP/{GIMP_APP_VERSION}
Windows: %APPDATA%/GIMP/{GIMP_APP_VERSION}
OSX: NSApplicationSupportDirectory/GIMP/{GIMP_APP_VERSION}
This commit causes a compiler warning because string literals are of
type signed char, and we use a character out of the signed range in the
mimetype for xz. It can be fixed by changing the data type of the mime
string to (unsigned char *), and not using string literals, but that's
a bit too much for this commit.
And along with it a lot of stuff like the drawable preview cache, the
gegl tile manager backend, temporary gimp_gegl_buffer_foo() stuff, and
the remaining bits of performance.
The projection is in an evil semi-ported state which makes it work
ok-ish for stuff like layer moving, but absolutely unbearable for
painting, there is also an off-by-one rendering glitch at some zoom
levels.
gimpdir goes to ~/Library/Gimp/x.y
thumbnails go to ~/Library/Caches/org.freedesktop.thumbnails
The thumbnail location is not standardized yet, but is the only
location that makes sense. Also fix user install to search old
gimpdirs to migrate in both Library and the classic location.
Remove the obsolete CABON_CFLAGS from all makefiles.
This commit will temporarily cause the build to break, because
file-compressor doesn't link -lz and -lbz2 after this commit. I'll fix
it soon, but if any of you want to kill me for breaking the build, feel
free to revert it. The commit will stay and can be used later.
This change was made so that file-compressor isn't crappy on Windows. It
also removes win32-specific code.
Gegl >= 0.1.8, Babl >= 0.1.6, Gdk-Pixbuf >= 1.24.0, Pango >= 1.29.4,
GLib >= 2.28.8, GTK+ 2.24.7
Which means depending on a gazillion of bug fixes, which means less
pain for GIMP 2.8 users, and less useless bugzilla traffic eating
developer resources.
This only helps to maintain proper includes in app/ and shouldn't
affect plug-ins at all, because these are supposed to only include the
main headers from libgimp/ since the beginning of time.
The gimpfootypes.h files do not have these guards, so we can continue
to maintain app/'s include policy that is very likely to error out if
wrong things are included.
This flag warns when we use deprecated API. It's added in configure.ac
instead of someplace like plugin-defs.pl because libpng is used
in several places.
Move the pixbuf layer and image thumbnail function from libgimpui to
libgimp and move gimp_layer_new_from_pixbuf() to gimplayer.[ch] where
it belongs. Change gimp-2.0.pc accordingly, adapt plug-in Makefiles
and update devel-docs.
and completely separate configure and sanity checks for gdk-pixbuf
from GTK+, because it's now distributed as a separate package. Remove
all sorts of conditional compiling based on GDK_CHECK_VERSION() and
CAIRO_VERSION.
Also fix the errors that were discovered by it. Multiple
definitions of global variables are no longer silently combined by
the linker, and will be reported as errors.
If you see GIMP crash strangely (i.e, apart from the usual crashes),
please go see a head doctor.
Add the automake distdir to POTFILES.skip so that a make distcheck
followed by a failed make distcheck doesn't fail with
[...]
gimp-2.7.2/tools/gimp-remote-x11.c
gimp-2.7.2/tools/gimp-remote.c
If some of these files are left out on purpose then please add them to
POTFILES.skip instead of POTFILES.in. A file 'missing' containing this list
of left out files has been written in the current directory.
Bump the "too new" test to GTK+ 2.26 which will never be released, but
this way we can keep the configure check around for later reuse after
the switch to GTK+ 3.x. Also bump the "too new" GLib and Pango
versions to really future versions so we early-catch deprecations
in the current development cycle.
Add a configure check for xvfb-run and run tests with that if
available. This makes it possible to easily run UI tests in headless
environments such as a server that does nightly builds; one just needs
to make sure xvfb-run is available.
Uodate Pango required version and stop using the deprecated
pango_cairo_font_map_create_context(). Compile with
PANGO_DISABLED_DEPRECATED for pango < 1.30.
Created a resource file for plug-ins and a rule to link it to them.
The application icon file is smaller than the default one, there's no
point in doubling the size of a plug-in executable just for an icon.
If no problems turn up, this will close the bug.
Eventually all executables will have the manifest and version resources
linked. In this initial step, the current resources are moved from the
app/ directory to a global build/windows directory.
Also set ACLOCAL_AMFLAGS in the top-level Makefile.am and update
the .gitignore file in the m4macros directory. This fixes the build
from a fresh checkout for me. Please revert if it causes problems.
Depend on the libpng version that is in debian testing (1.2.42) and
use png_set_expand_gray_1_2_4_to_8() instead of the deprecated
png_set_gray_1_2_4_to_8().
To give us experience with Glade + GtkBuilder, use it for the save
dialog in the PNG plug-in. The layout is as good as
identical. Mnemonics still works and strings are still translated.
Bump required GTK+ version to 2.18.1 so we get
e.g. gtk_widget_set_visible() and offscreen widget hierarchies. So far
we don't rely on bug fixes in any later releases.
Michael Henning