Commit Graph

110 Commits

Author SHA1 Message Date
24ac040563 build: bump versions 2022-04-01 13:16:31 -07:00
5e55f5f2cc build: switch build to GTK 4 2022-04-01 13:14:51 -07:00
a1f8f7e9f8 build: remove dazzle reference 2022-04-01 13:14:51 -07:00
b307550ea5 build: fix various meson deprecation warnings 2022-04-01 13:09:56 -07:00
3c5540047f build: bump for development 2022-04-01 12:41:14 -07:00
7d44e2db75 release 3.44.0 (GNOME 42.0) 2022-03-18 15:46:32 -07:00
c6dfe76810 Release 3.43.90 2022-03-07 13:43:12 -08:00
75364415b3 build: require libunwind unless disabled
Only some situations can use this properly without libunwind, so make
sure that we have it available unless -Dlibunwind=false.

Distributors, please be kind, use libunwind!
2021-11-04 19:48:29 -07:00
2903f08d02 release 3.42.1 2021-11-04 10:26:31 -07:00
f49681cb23 Release 3.42.0 (GNOME 41.0) 2021-09-21 16:41:52 -07:00
c6840d48d1 Release 3.40.1
This release fixes an issue created during GNOME 40 development where the
libsysprof-capture-4.a library could leak symbols into a shared library
that consumed it (such as Pango, GTK 4, and GLib).

This fixes that by removing symbol visibility from libsysprof-capture.a.
Distributions are encouraged to rebuild their libraries that consume
Sysprof's libsysprof-capture-4.a.
2021-03-23 10:58:57 -07:00
3f2b486741 build: remove symbol visibility from capture static library
The visibility of symbols was getting extracted from the .a into the
libraries that link_whole the archive. We can simply disable the visibility
from the .a with a custom config.h.meson which avoids it for the capture
library. Since we already double-compile those sources for the shared
library, the shared library visibility is not affected.

This was tested by compiling a simple library which calls
sysprof_clock_init() and ensuring that no extra symbols were found with
`nm libfoo.so | grep ' T '`.

Fixes #60
2021-03-22 14:49:00 -07:00
5463c965b1 Release 3.40.0 2021-03-19 17:17:46 -07:00
311405fd84 Release 3.39.94 (40.beta2) 2021-02-25 14:43:14 -08:00
902397917b release 3.39.92 2021-02-23 15:52:36 -08:00
4d332dc68a about: add symbolic information to release 2021-02-23 15:40:32 -08:00
b77523eb3d use g_memdup2() 2021-02-23 15:39:25 -08:00
14ae3e7c77 Release 40.alpha1 (3.39.90) 2021-01-14 15:39:08 -08:00
b8ac99dd48 meson.build: fix build without stack-protector
Check for -fstack-protector-strong availability through
has_link_argument as some compilers could missed the needed library
(-lssp or -lssp_nonshared) at linking step resulting in a build failure.

Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
2020-11-25 08:25:22 +01:00
04eacbcfe2 build: make polkit optional again for sysprof-cli 2020-11-16 14:57:04 -08:00
621fd70e81 build: avoid some compile checks on clang 12 2020-11-16 14:08:47 -08:00
c82b0a1677 build: specify -D_DARWIN_C_SOURCE on darwin
Needed for access to things like RTLD_NEXT
2020-11-16 13:58:42 -08:00
309ec5b091 Release 3.38.1 2020-10-15 17:12:53 -07:00
50439c9ca9 compat: add fallback implementation for reallocarray(3) 2020-09-15 20:30:48 +07:00
510ca74f23 Release 3.38.0 2020-09-12 11:42:36 -07:00
56b6dc55f6 release 3.37.92 2020-09-06 13:33:52 -07:00
d6aabaf1ff build: use current_build_dir() instead of build_root()
This can help with subprojects.
2020-08-21 15:20:04 -07:00
e656c334d8 release 3.37.90 2020-08-11 08:57:01 -07:00
857d91624c build: bump meson version 2020-07-28 09:33:40 -07:00
4f4674c54f Release 3.37.2 2020-07-28 09:21:20 -07:00
50679dba04 build: Add feature test flags to project-wide preprocessor flags
Rather than defining them piecemeal in each C and H file, and having
problems with headers being included in different orders so that
`_POSIX_C_SOURCE` gets a value which is too low, just define the feature
test flags in `meson.build`.

See `man 7 feature_test_macros`.

This is useful for when building sysprof as a subproject beneath a
project which doesn’t define these macros, or which targets an older C
standard than `gnu11`.

Signed-off-by: Philip Withnall <withnall@endlessm.com>
2020-07-06 11:27:06 +01:00
24033d586e build: make C++ language conditional
We don't need C++ unless we are building the C++ demangler.
2020-07-04 15:46:03 -07:00
3c8a9e4152 build: bump for new ABI 2020-07-04 11:30:11 -07:00
5d1678974a build: Simplify config file setup
This introduces no functional changes, just simplifies the code.

As described in the meson manual
(https://mesonbuild.com/Configuration.html#configuring-without-an-input-file),
passing `false` to `configuration_data.set()` will cause the specified
value to be `#undef`fed. Passing `true` will cause it to be defined to
1.

Signed-off-by: Philip Withnall <withnall@endlessm.com>
2020-07-03 22:01:33 +01:00
3c7b76c6ba build: Move Meson dependency() calls to where they’re used
This avoids making the project depend on all its dependencies, some of
which are optional, when being built as a Meson subproject.

Signed-off-by: Philip Withnall <withnall@endlessm.com>
2020-07-03 22:01:33 +01:00
6e281dca1f libsysprof-capture: Use strlcpy() instead of g_strlcpy()
If the system doesn’t provide `strlcpy()` (FreeBSD does, Linux doesn’t),
use an inbuilt copy instead.

Signed-off-by: Philip Withnall <withnall@endlessm.com>

Helps: #40
2020-07-03 22:00:34 +01:00
2cc79dd444 build: bump ABI to 4 2020-07-02 21:07:11 +01:00
1bfdcca9d1 build: Update bug report define in meson.build
GitLab!

Signed-off-by: Philip Withnall <withnall@endlessm.com>
2020-06-25 14:58:42 +01:00
4a2aaf7ebd bump version for development 2020-03-13 10:18:51 -07:00
dcaeba54ee release 3.36.0 2020-03-06 14:03:33 -08:00
82c2ee5a73 build: fix building in jhbuild
Not sure why we are hitting this, but we are so just force it to work.
2020-03-06 13:43:15 -08:00
a6c39af553 preload: move to libdir from libexecdir
This isn't an executable, it just belongs in libdir.
2020-03-05 15:46:03 -08:00
d6fef722a0 build: add meson subproject with libdazzle.wrap
When building on older systems such as RHEL 7, it can be convenient
to have access to libdazzle directly.

This also has to work around a problem that RHEL 7 does not support
the version of git that Meson requires for [wrap-git] by using a
specific release.
2020-03-02 11:35:14 -08:00
0add5eb035 release 3.35.92 2020-02-28 09:27:40 -08:00
5b88e9c3aa build: check for unw_set_cache_size()
This may not be available on older libunwind versions such as 1.2.
2020-02-20 11:22:56 -08:00
454b192a19 build: relax GLib requirement to 2.56 2020-02-20 11:08:26 -08:00
cd8a99402f build: fallback to __sync_synchronize()
If we're running on a GCC older than 4.9, then we won't have the
stdatomic.h available. We can just use a full barrier instead using
__sync_synchronize() to get the same effect, albeit slower.
2020-02-20 10:38:35 -08:00
e06638d665 build: make libunwind optional 2020-02-17 12:02:44 -08:00
33c81a3a9c memprof: add memory profiling using LD_PRELOAD
This brings over some of the techniques from the old memprof design.
Sysprof and memprof shared a lot of code, so it is pretty natural to
bring back the same callgraph view based on memory allocations.

This reuses the StackStash just like it did in memprof. While it
would be nice to reuse some existing tools out there, the fit of
memprof with sysprof is so naturally aligned, it's not really a
big deal to bring back the LD_PRELOAD. The value really comes
from seeing all this stuff together instead of multiple apps.

There are plenty of things we can implement on top of this that
we are not doing yet such as temporary allocations, cross-thread
frees, graphing the heap, and graphing differences between the
heap at to points in time. I'd like all of these things, given
enough time to make them useful.

This is still a bit slow though due to the global lock we take
to access the writer. To improve the speed here we need to get
rid of that lock and head towards a design that allows a thread
to request a new writer from Sysprof and save it in TLS (to be
destroyed when the thread exits).
2020-02-07 19:00:33 -08:00
21fc5609f2 build: ignore post-install unless gtk app is installed 2020-02-07 11:04:03 -08:00