4d116ed0f0
[ci] Remove cmake
...
Apparently(?) our cmake build never bothered setting up pthreads.
So when I removed the busy-loop mutex impl it now doesn't build.
711c241f6chttps://github.com/harfbuzz/harfbuzz/pull/2874#issuecomment-782778263
2021-02-22 17:14:45 -07:00
094406b67b
[ci] Disable failing Arch Linux job
...
Some g-ir-scanner failure that does not seem to be related to us. The
job does not seem to be testing anything interesting anyway.
2021-02-10 23:22:02 +02:00
e8c15b7a1c
[ci] Install gobject-introspection on distcheck job
...
To fix missing --enable-introspection from release tarballs.
Fixes https://github.com/harfbuzz/harfbuzz/issues/2779
2020-12-26 23:17:08 +02:00
bac8a856f8
Revert "[CI] Don’t install ragel on circleci jobs"
...
This reverts commit 9a9da110a5
2020-11-21 08:41:06 -07:00
a082fe400d
ci: Install meson 0.56.0 in crossbuild-win32
...
It is required for using cairo as subproject.
2020-11-17 11:52:35 -05:00
5adc1344a4
[ci] Upload meason logs on some CircleCI jobs
2020-10-23 19:02:54 +02:00
71376f88ef
[ci] CircleCI removed xcode:9.0.1 image
...
The replacement run newer versions of macOS (10.13.3), making this job
obsolete as it tests nothing the other jobs don’t test.
2020-10-23 19:02:41 +02:00
9a9da110a5
[CI] Don’t install ragel on circleci jobs
...
Latest ragel version is broken (https://github.com/adrian-thurston/ragel/issues/56 ), but we also don’t need to regenerate state machine on CI jobs.
2020-10-17 09:24:54 -07:00
84dd65a874
[test] Remove timeout from test runners
...
See https://github.com/harfbuzz/harfbuzz/issues/2707#issuecomment-707744079
This wasn’t inconsistent as well, HB_TEST_SUBSET_FUZZER_TIMEOUT defaulted
to 12 in the test runner, but it was overridden to 50 in meson.build,
and then meson has its own test timeout.
2020-10-15 00:49:02 -07:00
e43b9a29a9
[ci] Trigger dependencies of publish jobs on tags
2020-09-11 22:21:59 +03:00
deb09e0d75
[ci] Tweak job names so config is valid Circle 2.1 schema
2020-09-11 20:12:35 +03:00
f4dea33f6d
[ci] Move repeated artifact upload job code to script
2020-09-11 20:12:24 +03:00
65be538fff
[ci] Add job to publish autotools dist file to GitHub
2020-09-11 20:12:14 +03:00
42bcc47d01
[ci] Use executor with Circle's image for autotools job
2020-09-11 20:11:52 +03:00
c70f509022
[ci] Use `ghr` CLI tool to post Win32 artifact to GitHub
2020-09-11 20:11:38 +03:00
610ed3b42c
[ci] Add job to publish Win32 releases and pass it artifacts
2020-09-11 20:11:22 +03:00
a2d45cd5f7
[ci] Use Circle's customized image for workspace support
2020-09-11 20:10:43 +03:00
6e32145dc9
[meson] Make compatbile with 0.47.0
2020-08-13 18:28:42 +04:30
58209c86a0
[meson] Make compatbile with 0.49.0
...
Contains a just put together summary feature polyfill and workaround
to broken ternary operator.
2020-08-13 18:28:42 +04:30
2af58c3518
Revert "Remove cmake build files"
...
This reverts commit 4e572c9407
2020-08-12 01:00:33 +04:30
5193357832
Revert "Remove autotools build support"
...
This reverts commit 01ac32aab2
2020-08-11 23:51:59 +04:30
c0107c1e6e
[ci] test experimental APIs in Arch Linux and sanitizers
2020-08-10 14:30:34 +04:30
01ac32aab2
Remove autotools build support
2020-08-07 23:28:12 +04:30
50151567c1
[ci] Disable Travis CI
2020-08-02 11:08:12 +04:30
4e572c9407
Remove cmake build files
...
As discussed in https://github.com/harfbuzz/harfbuzz/pull/2252
2020-07-30 16:19:15 +04:30
bec3db49bd
[ci] meson-gcc-mingw to crossbuild-win32
2020-07-25 13:45:34 +04:30
024fc665da
[ci] disable clang nightly installation
...
they don't like to work apparently for some reason https://circleci.com/gh/harfbuzz/harfbuzz/149164
better to use oss-fuzz docker images, gcr.io/oss-fuzz-base/base-clang, but the fact it uses 16.04 is showstopper for us.
2020-07-20 23:57:27 +04:30
3163e9fedb
[ci] Remove HB_TEST_SUBSET_FUZZER_TIMEOUT
...
meson is passing it itself
Fortunately oss-fuzz has its own timeout value, I just like to be conservative here
and not involve outsiders with flakiness of tests specially as meson which runs
them concurrently so consuming more time is expected on meson.
2020-07-12 23:11:45 +04:30
f8d345ee06
[ci] Remove ambiguous-reversed-operator warning from whitelist
...
Should be fixed by 6a45df0
2020-07-09 10:57:51 +04:30
5ea11cb817
[ci] don't use harfbuzz.cc with -Weverything
...
Causes
In file included from src/harfbuzz.cc:13:
src/hb-ot-cff1-table.cc:36:17: error: using namespace directive in global context in header [-Werror,-Wheader-hygiene]
using namespace CFF;
^
In file included from src/harfbuzz.cc:14:
src/hb-ot-cff2-table.cc:35:17: error: using namespace directive in global context in header [-Werror,-Wheader-hygiene]
using namespace CFF;
^
2 errors generated.
2020-07-08 19:29:47 +04:30
e4f9969108
[ci] migrate to meson
...
two bots, one bot here (distcheck) and one in travis still run autotools and
won't be removed till we decide about autotools
2020-07-08 19:18:31 +04:30
471ae786ce
[ci] Install a newer meson version
2020-07-06 22:19:59 +04:30
694a0c6bbc
[meson] Always add icu deps
2020-07-06 22:19:59 +04:30
3a8d137c2c
[ci] migrate macOS bots to meson
2020-07-06 22:19:59 +04:30
702847aadd
[meson] Turn benchmark to a feature
2020-06-29 12:54:58 +04:30
fd8b0a44c4
[ci] Limit number of processes ninja uses
...
Our CricleCI bots run with 32 cores last time I check which makes random compile issues
like this https://circleci.com/gh/harfbuzz/harfbuzz/146030 to happen. Guess that has
something to do with the way virtualization is done in Circle CI so let's limit number of processes.
2020-06-29 00:19:38 +04:30
ca8202208a
[ci] retire psvita and djgpp bots
...
till update and overhaul
2020-06-28 11:34:01 +04:30
95b1081be4
Add performance benchmark for shaping, get extents and draw
2020-06-25 01:30:07 +04:30
b5ae8eab44
[ci] raise msan bot subset timeout
2020-06-21 08:51:46 +04:30
47dedc7500
[ci] increase tsan and ubsan subset fuzzer timeout
...
as https://circleci.com/gh/harfbuzz/harfbuzz/144966 and https://circleci.com/gh/harfbuzz/harfbuzz/144970
2020-06-21 01:54:57 +04:30
f31e3d426d
[ci] raise subset fuzzer timeout for asan and valgrind bots
2020-06-12 17:33:11 +04:30
be278acf44
[ci] raise subset fuzzer cases runner timeout
...
the bots are either not optimized ones or sanitizer ones so not much value
on trying to optimize them further.
2020-06-12 10:41:07 +04:30
7e74aa6a78
[ci] fix meson-gcc-mingw bot
...
keep build folder for dist testing
2020-06-04 09:32:20 +04:30
b35764334e
[ci] check workability of meson based dist
2020-06-04 04:10:11 +04:30
870304df76
[meson] Remove amalgam build option
...
Fixes #2444
2020-06-04 03:26:19 +04:30
6058ede3ae
Build win32 artifacts using meson
...
As a part of meson migration, this builds hb-shape dependencies
using meson subprojects and removes different unrelated mingw
files we've put on source root about things that aren't really our
concern. This also simplifies all the instructions we've put in
README.mingw.md and makes it easy to build the project on
distros that no mingw libraries (such i686-win32-{freetype,glib,...}
are packaged. The known catch is however lack of hb-view as its need
for build of cairo which apparently isn't that straightforward.
2020-06-01 12:58:31 +04:30
9d13124cfd
[ci] fix alpine build issue
2020-05-30 14:45:22 +04:30
cc53fd158c
[meson] Normalize options names
...
As https://github.com/harfbuzz/harfbuzz/pull/2412#discussion_r426275027
2020-05-21 19:33:18 +04:30
99e2342a6b
[ci] Install newer meson from pip
2020-05-21 18:52:31 +04:30
eac2c3bdb1
[ci] Enable sanitizer bots again
2020-05-21 08:41:02 +04:30
Behdad Esfahbod
Khaled Hosny
Khaled Hosny
Xavier Claessens
Khaled Hosny
Caleb Maclennan
Ebrahim Byagowi