| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
As for Meson 0.43.0, no way to prevent a generated
GIR file to be installed in the system, even
if disable-introspection=true is set.
May this is not an issue as libgee generates
header, VAPI and GIR, all of them needed when
developing.
Just in case g-ir-compiler is not found
no typelib is generated and so introspection
is incomplete any way. g-ir-compiler is
installed with gobject-introspection package
in almost all distributions.
|
| |
|
|
|
|
|
|
|
|
| |
Building benchmark requires to recompile all
sources in order to get access to internal VAPI.
This is not a real issue, if taking in account
benchmark are not builded by default.
|
| |
|
| |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This needs:
https://github.com/mesonbuild/meson/pull/1780
to be merge before, because it adds internal VAPI and header
generation suport to Meson.
Now you can run:
ninja benchmark
in order to run Gee's benchmarks.
|
| |
|
|
|
|
|
| |
Here efectibaly remove header and VAPI for internals,
as reported by commit: 8e931bc334b2cba6d1311c30c8de034619bbfdbe
|
|
|
|
|
|
|
| |
* Removed internal headers and VAPI generation, because
no one is using them, they are not installed or distributed
* Unit Tests ported to Meson, pass all
|
|
|
|
|
|
|
|
|
|
|
|
| |
Meson has no support to pass --internal-header
and --internal-api, so a merge request has been sent
to meson in order to allow it.
https://github.com/mesonbuild/meson/pull/1780
This is required to finish Unit Test integration with
Meson. So initial meson.build is added, but requires
to fix it in future commits.
|
| |
|
| |
|
| |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Include Makefile.introspection instead of using a custom rule which will
result in the following build failure when cross-compiling on buildroot
because of missing --includedir:
/home/giuliobenetti/autobuild/run/instance-1/output-1/host/bin/../riscv32-buildroot-linux-gnu/sysroot/usr/bin/g-ir-compiler -l `/usr/bin/sed -nE "s/^dlname='([A-Za-z0-9.+-]+)'/\1/p" libgee-0.8.la` -o Gee-0.8.typelib Gee-0.8.gir libgee-0.8.la
Could not find GIR file 'GObject-2.0.gir'; check XDG_DATA_DIRS or use --includedir
error parsing file Gee-0.8.gir: Failed to parse included gir GObject-2.0
Fixes:
- http://autobuild.buildroot.org/results/884faa0f84c8dc43ed1ca6cde9caf21c731a4b35
Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
|
|
|
|
|
|
|
| |
It is possible that a spurious or stolen wakeup could occur. For that
reason, waiting on a condition variable should always be in a loop,
based on an explicitly-checked predicate.
Fixes https://gitlab.gnome.org/GNOME/libgee/issues/34
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
|
|
|
| |
This reverts commit da95e830524ffa309eb57925320666e5085b9d66.
|
|
|
|
|
|
|
| |
* one_match (Predicate<G>) check if exactly one element matches
* count_match (Predicate<G>) returns the count of items that matches
https://bugzilla.gnome.org/show_bug.cgi?id=781641
|
| |
|
| |
|
| |
|
| |
|
| |
|
|
|
|
|
|
|
|
|
|
| |
* first_match (Predicate<G>) returns the first item that matches
* any_match (Predicate<G>) checks if any element matches
* all_match (Predicate<G>) checks if all elements match
* max/min returns max/min value
* order_by to perform ordering on any Traversable
https://bugzilla.gnome.org/show_bug.cgi?id=776558
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Enumerations and flags are classed types for Vala, not integers, so
they don't fall in the `typeof(G) == typeof(int)` kind of tests. This
leads to using the generic code in which Vala assumes pointer-sized
elements, which is often not true for enumerations and flags.
Add special case for those to use the `int` converters for enumerations
and flags.
This is most generally correct, but not always: the compiler will
likely chose a larger type for a specific enumeration if one of its
value is larger than `int`. It would be tempting to use the
enumeration's class minimum and maximum values to determine the
appropriate type, but unfortunately the API for this uses int itself,
so doesn't help.
https://bugzilla.gnome.org/show_bug.cgi?id=774669
|
| |
|
|
|
|
|
|
|
|
| |
Discussed this briefly with upstream on IRC, and it was concluded that
this should probably have been forbidden by the Vala compiler in the
first place.
https://bugzilla.gnome.org/show_bug.cgi?id=772417
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Same issue in HashMap and TreeMap:
```
==3251==ERROR: AddressSanitizer: heap-use-after-free on address 0x604000000870 at pc 0x000108be666b bp 0x7fff571e62b0 sp 0x7fff571e62a8
WRITE of size 8 at 0x604000000870 thread T0
#0 0x108be666a in g_nullify_pointer gutils.c:2051
#1 0x108b8c906 in weak_refs_notify gobject.c:2638
#2 0x108bbb17c in g_data_set_internal gdataset.c:407
#3 0x108b887db in g_object_unref gobject.c:3148
#4 0x108a4b0ec in map_tests_test_entry_weak_pointer_lifetime testmap.c:1358
0x604000000870 is located 32 bytes inside of 40-byte region [0x604000000850,0x604000000878)
freed by thread T0 here:
#0 0x1090f0e29 in wrap_free (libclang_rt.asan_osx_dynamic.dylib+0x4ae29)
#1 0x108ace566 in gee_hash_map_unset_helper hashmap.c:1692
#2 0x108acc534 in gee_hash_map_real_unset hashmap.c:1520
#3 0x108a4b0df in map_tests_test_entry_weak_pointer_lifetime testmap.c:1357
previously allocated by thread T0 here:
#0 0x1090f0c60 in wrap_malloc (libclang_rt.asan_osx_dynamic.dylib+0x4ac60)
#1 0x108bce848 in g_malloc gmem.c:95
#2 0x108bd6585 in g_slice_alloc gslice.c:1012
#3 0x108bd6bee in g_slice_alloc0 gslice.c:1038
#4 0x108acdc27 in gee_hash_map_node_new hashmap.c:2084
#5 0x108acc277 in gee_hash_map_real_set hashmap.c:1494
#6 0x108a4b032 in map_tests_test_entry_weak_pointer_lifetime testmap.c:1311
https://bugzilla.gnome.org/show_bug.cgi?id=772418
|
| |
|
|
|
|
| |
https://bugzilla.gnome.org/show_bug.cgi?id=772149
|
| |
|