Andrzej Hunt [Sat, 4 Sep 2021 11:43:29 +0000 (13:43 +0200)]
babl_format_new: also free doc to plug leak
doc and name are allocated via babl_strdup, hence if we don't use them
we need to free both of them - therefore add a free(doc) to match the
existing free(name).
format_new (called just below) transfers ownership of doc, hence it's
correct not to free doc in the case where we're instantiating a new
format - whereas name is copied within format_new, and therefore needs to be
free'd in either scenario.
Leak seen while running gimp+babl built with ASAN:
Direct leak of 99 byte(s) in 1 object(s) allocated from:
#0 0x5e87cd in malloc /home/abuild/rpmbuild/BUILD/llvm-12.0.0.src/build/../projects/compiler-rt/lib/asan/asan_malloc_linux.cpp:145:3
#1 0x7f038605e2b8 in babl_malloc /home/ahunt/git/babl/_build/../babl/babl-memory.c:115:9
#2 0x7f038605eec7 in babl_strdup /home/ahunt/git/babl/_build/../babl/babl-memory.c:296:9
#3 0x7f038605110f in babl_format_new /home/ahunt/git/babl/_build/../babl/babl-format.c:376:17
#4 0x7f037313a432 in init /home/ahunt/git/babl/_build/../extensions/cairo.c:569:25
#5 0x7f03860412a1 in babl_extension_load /home/ahunt/git/babl/_build/../babl/babl-extension.c:226:7
#6 0x7f0386040d1c in babl_extension_load_dir /home/ahunt/git/babl/_build/../babl/babl-extension.c:268:19
#7 0x7f03860409a8 in babl_extension_load_dir_list /home/ahunt/git/babl/_build/../babl/babl-extension.c:338:17
#8 0x7f038607d6cd in babl_init /home/ahunt/git/babl/_build/../babl/babl.c:154:7
#9 0x7f0386bce5ab in gegl_post_parse_hook /home/ahunt/git/gegl/_build/../gegl/gegl-init.c:542:3
#10 0x7f03844aad8a in g_option_context_parse /home/ahunt/git/glib/_build/../glib/goption.c:2228:12
#11 0x7f0386bce381 in gegl_init /home/ahunt/git/gegl/_build/../gegl/gegl-init.c:189:8
#12 0xef1f6b in gimp_init_for_testing /home/ahunt/git/gimp/app/tests.c:69:3
[...]
SUMMARY: AddressSanitizer: 99 byte(s) leaked in 1 allocation(s).
Øyvind Kolås [Wed, 9 Feb 2022 19:03:46 +0000 (20:03 +0100)]
babl: fix mutex lock/unlocks symmetry in babl_fish()
In the case where the fish we are to look up is the memcpy fish we were
not locking the mutex. Whereas the common cleanup after the branches
assumes the lock is held. This is probably the cause of gimp#7632.
Jehan [Mon, 24 Jan 2022 14:11:34 +0000 (15:11 +0100)]
tools: use rand() as fallback of random().
This fixes building babl on Windows as the random() API is not available
there (unlike rand()). Windows docs says it's not cryptographically
secure but for this tool usage, it's probably enough and I see we use
the same function replacement tricks on other tool sources.
Øyvind Kolås [Sun, 23 Jan 2022 10:54:17 +0000 (11:54 +0100)]
babl: add lut for some u8 conversions
The fish path now marks created fishes that shall be monitored for
usage and have full 24bit LUTs generated on demand when limits are
exceeded. This 3-4x increases the throughput.
Øyvind Kolås [Sun, 23 Jan 2022 03:18:01 +0000 (04:18 +0100)]
babl: adjust default path parameters to a path_length of 3
Also adjust so that the max additional bits of depth is 2, this makes us
catch some fast paths that are missed by very promising slow two-step
converters.
Øyvind Kolås [Sun, 23 Jan 2022 01:18:46 +0000 (02:18 +0100)]
extensions: reduce amount of competing conversions
Avoid loading extensions if a variant with better SIMD support
is expected to exist. We know all variants that are expectd to
exist and thus can register only the fastest known variant.
This avoids combinatorial explosion of conversions to test.
Øyvind Kolås [Fri, 29 Oct 2021 01:34:36 +0000 (03:34 +0200)]
Oklab: disable only Oklch not Oklab
Oklab itself is enabled by default now for testing, but does not seem to
produce correct values with the Oklab test vectors, even with manual
attempt at converting the test vector (presumed) CIE XYZ D65 to CIE XYZ
D50 before running the conversion.
The failure to pass symmetry tests as mentioned in issue #63 seems to be
limited to Oklch, Oklab itself is fine.
Jehan [Wed, 3 Mar 2021 10:03:38 +0000 (11:03 +0100)]
babl: re-creation of palette conversions must always set data.
If we allow conversions to be recreated for palettes, it will create a
new identical conversion with collision number incremented in the name
and no user data. Yet the user data is extremely important for palette
conversions, because it contains the palette itself which is required by
the palette conversion functions. Hence we end up crashing.
Moreover, although I think that right now, only palette conversions use
the data field, it might be used for other type of data in the future.
In any case, we should always save this data along as it has been stored
there for a reason.
Finally in babl-palette code, make sure the user data is added to the
models before creating the formats and conversions as we also ended up
in crash-happy code when we didn't.
Jehan [Wed, 3 Mar 2021 09:54:38 +0000 (10:54 +0100)]
babl: do not append space to the format name when explicitly set.
When a format is named explicitly by the caller, we should just keep the
same name, not tweak it. Generating a name is only valid when no
explicit name is given. Moreover this is the behavior as described in
the docs of babl_format_new():
> If no name is provided a (long) descriptive name is used.
(which usually implies that with a name provided, this one will be used
instead of a generated name)
This is especially important for palette formats for which name is
mostly not descriptive anyway (either caller-set or generated by babl
with incremental number) and this fixes a bug when calling
babl_new_palette_with_space() with the same name and space would fail to
find the already created formats (yet would find the models and return
NULL for both formats) because it would search without the space
description appended. Yet the docs of babl_new_palette*() clearly says
as well:
> If you pass in the same name the previous formats will be provided
> again.
Moreover we actually already tweak a bit the palette format name by
appending the space pointer. No need to tweak it twice appending the
same information (space) in a different way.
See also: https://gitlab.gnome.org/GNOME/gimp/-/issues/6501
John [Fri, 26 Feb 2021 07:19:44 +0000 (07:19 +0000)]
build: subproject improvements:
- don't expose source in dependency to avoid name clashes
- add babl path variables to dependency object
- add babl path variables to uninstalled pc
Recently Arch has changed its glibc package to require kernel
4.4[1] as a poor man's attempt to disable faccessat2 syscall that
is known to cause problems with old libseccomp and Docker. Downgrade
glibc to the pkgrel before the change and ensure it's not upgraded.
Additionally switch to the official Arch Linux image in the library
namespace. The base image has been deprecated.