Closures use a 16-bit atomic reference count, which is really slow
on certain ARM64 CPUs such as the Cortex-A57 (glib#1316). This is
non-trivial to solve, since the public struct field cannot be enlarged
to 32-bit while preserving ABI, and 16-bit atomic operations would be new
(and rather niche) API.
Until this can be solved properly (hopefully in GLib 2.59.x), cut down
the number of signal emission cycles and bump up the timeout in the
Meson build system, so that builds won't time out. We can't just take
another zero off the number of signal emission cycles, as was done in the
original version of this patch in Debian, because if we do that it can
result in test failures when the main thread starves the other threads.
ARM64 CPUs are backwards-compatible with 32-bit ARM, and the same
slowdown can be seen when building and testing 32-bit code on these
CPUs, so check for both 32- and 64-bit ARM.
Bug-Debian: https://bugs.debian.org/880883
Co-authored-by: Iain Lane <laney@debian.org>
Signed-off-by: Simon McVittie <smcv@debian.org>
This reverts commits:
• 9ddcc79502
• ae02adc3c3
g_date_time_format() supports a few non-standard format placeholders:
• %:z
• %::z
• %:::z
These are all gnulib strtime() extensions, and hence are not recognised
by the compiler when the function is annotated with G_GNUC_STRFTIME.
However, this wasn’t noticed when we originally merged this change
because the errors were disabled in the tests which covered those
placeholders.
* Wait for adder threads before deallocating crawler_array and
context_array to avoid use after-free and data race.
* Handle spurious wakeups around g_cond_wait.
* Avoid starting recurser_idle without context.
Fixes issue #1530.
If we tell clang that we're passing the arguments on to a compatible
printf call, it can see that we're doing it right, and won't warn us
about -Wformat-nonliteral. This follows up from commit abac67c0.
Signed-off-by: Simon McVittie <smcv@collabora.com>
This makes it easier to debug test failures, by ensuring that g_debug()
and g_test_message() are printed as TAP diagnostics.
Fixes: https://gitlab.gnome.org/GNOME/glib/issues/1528
Signed-off-by: Simon McVittie <smcv@collabora.com>
As we recently established, G_G*_FORMAT macros are for glib
functions *only*. It's not guaranteed that they will work for
C runtime printf() or scanf() variants, and indeed - in this case
they do not (GCC thinks that MSVCRT sscanf() requires %I64 prefix
for 64-bit values; whether that is true or not is irrelevant at this
point - we need to make the werror go away).
LPARAM is LONG_PTR which is the same as INT_PTR on Windows.
WPARAM is UINT_PTR.
HWND is a pointer.
(%p prints pointers without 0x prefix, hence the extra 0x).
These keep on taking just longer than 30s on my local machine when run
in parallel with the rest of the tests (i.e. with `ninja test`). Testing
them individually, they do terminate correctly.
Signed-off-by: Philip Withnall <withnall@endlessm.com>
It is a bug if we distribute files which are generated at build time —
they should be built on the machine which is compiling GLib, not be
shipped in the tarball.
This brings the autotools-generated tarball in line with the
ninja-generated one, with the exception of man pages and gtk-doc HTML
output.
Signed-off-by: Philip Withnall <withnall@endlessm.com>
It's been 4 years and 8 development cycles since we introduced
G_ADD_PRIVATE and offset-based private data access. It is now
time to finally deprecate the old mechanism.
Closes: #699
Meson has the ability to classify tests according to "suites", a list of
tags. This is especially useful when we want to run specific sets of
tests — e.g. only GLib's tests — instead of the whole test suite. It
also allows us to classify special tests, like "slow" ones, so that we
can only run them when needed.
This fixes:
glib/tests/testglib.c: In function ‘test_paths’:
glib/tests/testglib.c:955:3: error: this ‘if’ clause does not guard... [-Werror=misleading-indentation]
if (g_test_verbose ())
^~
glib/tests/testglib.c:958:5: note: ...this statement, but the latter is misleadingly indented as if it were guarded by the ‘if’
{
^
https://bugzilla.gnome.org/show_bug.cgi?id=796385
The existing implementation was completely incorrect (despite the fix in
commit 566e64a66) — it always compared GVariants by pointer, rather than
by value.
Reimplement it to compare them by value where possible, depending on
their type. The core of this implementation is g_variant_compare(). See
the documentation and tests for further details of the new sort order.
This adds documentation and tests.
Signed-off-by: Philip Withnall <withnall@endlessm.com>
https://bugzilla.gnome.org/show_bug.cgi?id=795735
Getting the canonical filename is a relatively common
operation when dealing with symbolic links.
This commit exposes GLocalFile's implementation of a
filename canonicalizer function, with a few additions
to make it more useful for consumers of it.
Instead of always assuming g_get_current_dir(), the
exposed function allows passing it as an additional
parameter.
This will be used to fix the GTimeZone code to retrieve
the local timezone from a zoneinfo symlink.
(Tweaked by Philip Withnall <withnall@endlessm.com> to drop g_autofree
usage and add some additional tests.)
https://bugzilla.gnome.org/show_bug.cgi?id=111848
All those logging functions already add a newline to any message they
print, so there’s no need to add a trailing newline in the message
passed to them.
Signed-off-by: Philip Withnall <withnall@endlessm.com>
Reviewed-by: nobody
It seems that when GLib is compiled without CFLAGS=-g, gdb can’t work
out the size of __glib_assert_msg, so assumes it’s 4 bytes — even on
64-bit systems. This causes it to not read the most significant 4 bytes
of the assertion message pointer, and hence it can’t print the stored
assertion message. This causes assert-msg-test to fail.
The upstream gdb bug is
https://sourceware.org/bugzilla/show_bug.cgi?id=22501.
Work around that by referencing and dereferencing __glib_assert_msg so
that gdb treats it as a pointer of sizeof(char*) rather than of the size
it incorrectly calculated from the library’s symbol table (or through
some other mystical process).
Signed-off-by: Philip Withnall <withnall@endlessm.com>
https://bugzilla.gnome.org/show_bug.cgi?id=782057
It is outdated and no longer effectively used. It was originally in
place to prevent rebuilding generated files (from a tarball) if the
right build tools (awk, Perl, indent) were not available. However, we no
longer use indent, we have hard-required awk for a while, and the only
places the @REBUILD@ substitution was still used were for
glib-genmarshal, which has recently been rewritten in Python (so no
longer depends on whether Perl is available).
Drop the whole lot.
Signed-off-by: Philip Withnall <withnall@endlessm.com>
https://bugzilla.gnome.org/show_bug.cgi?id=694723
After the build system rework in commit f9eb9e testgobject fell through
the cracks and was not built since then.
Re-enable it, even if it is currently failing due to commit 31fde56.
(Tweaked by Philip Withnall to add meson.build support.)
https://bugzilla.gnome.org/show_bug.cgi?id=701156
Commit 31fde56 changed the way the private data is laid out in memory by
putting it *before* the instance data to keep the offsets fixed
regardless of the number of many subclasses.
This means that the invariant testgobject was verifying is no longer
true and the failing tests can be safely dropped.
https://bugzilla.gnome.org/show_bug.cgi?id=701156
ssize_t is supported widely, but not universally, so use gssize instead.
Currently only one piece of code actually *needs* this change to be compilable
with MSVC, the rest are mostly in *nix parts of the code, but these are changed
too, for symmetry.
https://bugzilla.gnome.org/show_bug.cgi?id=788180
Prevent the situation where errno is set by function A, then function B
is called (which is typically _(), but could be anything else) and it
overwrites errno, then errno is checked by the caller.
errno is a horrific API, and we need to be careful to save its value as
soon as a function call (which might set it) returns. i.e. Follow the
pattern:
int errsv, ret;
ret = some_call_which_might_set_errno ();
errsv = errno;
if (ret < 0)
puts (strerror (errsv));
This patch implements that pattern throughout GLib. There might be a few
places in the test code which still use errno directly. They should be
ported as necessary. It doesn’t modify all the call sites like this:
if (some_call_which_might_set_errno () && errno == ESOMETHING)
since the refactoring involved is probably more harmful than beneficial
there. It does, however, refactor other call sites regardless of whether
they were originally buggy.
https://bugzilla.gnome.org/show_bug.cgi?id=785577
The Meson build has fallen a bit behind the Autotools one, when it comes
to the internally built tools like glib-mkenums and glib-genmarshals.
We don't need to generate gmarshal.strings any more, and since the
glib-genmarshal tool is now written in Python it can also be used when
cross-compiling, and without indirection, just like we use glib-mkenums.
We can also coalesce various rules into a simple array iteration, with
minimal changes to glib-mkenums, thus making the build a bit more
resilient and without unnecessary duplication.
gen-casefold-txt.pl and gen-casemap-txt.pl are licensed under GPLv2+, so
they are not touched by this commit.
A lot of *.c files in tests/ don't have a license header.
https://bugzilla.gnome.org/show_bug.cgi?id=776504
It's unnecessary, and only adds visual noise; we have been fairly
inconsistent in the past, but the semi-colon-less version clearly
dominates in the code base.
https://bugzilla.gnome.org/show_bug.cgi?id=669355
Signed integer overflow is undefined behaviour: if a compiler
detects signed integer overflow, it is free to compile it to absolutely
anything.
Signed-off-by: Simon McVittie <smcv@debian.org>
Bug: https://bugzilla.gnome.org/show_bug.cgi?id=775510
Reviewed-by: Colin Walters
We don't need to run binaries we just built in order to successfully
build GLib and friends any more.
Since commit b74e2a7, we don't need to run glib-genmarshal when building
GIO; since commit f9eb9eed, all our tests (including the ones that do
need to run binaries we just built) are only built when running "make
check", instead of unconditionally at every build.
This means that we don't need to check for existing, native binaries
when cross-compiling, and fail the configuration step if they are not
found — which also means that you don't need to natively build GLib for
your toolchain, in order to cross-compile GLib.
We can also use the cross-compilation conditional, and skip those tests
that require a binary we just built in order to build.
https://bugzilla.gnome.org/show_bug.cgi?id=753745
I searched all files that mention g_test_run, and replaced most
g_print() calls. This avoids interfering with TAP. Exceptions:
* gio/tests/network-monitor: a manual mode that is run by
"./network-monitor --watch" is unaffected
* glib/gtester.c: not a test
* glib/gtestutils.c: not a test
* glib/tests/logging.c: specifically exercising g_print()
* glib/tests/markup-parse.c: a manual mode that is run by
"./markup-parse --cdata-as-text" is unaffected
* glib/tests/testing.c: specifically exercising capture of stdout
in subprocesses
* glib/tests/utils.c: captures a subprocess's stdout
* glib/tests/testglib.c: exercises an assertion failure in g_print()
Bug: https://bugzilla.gnome.org/show_bug.cgi?id=725981
Reviewed-by: Colin Walters <walters@verbum.org>
Signed-off-by: Simon McVittie <simon.mcvittie@collabora.co.uk>
The test is to remove all the odd values with my_hash_callback_remove(),
then iterate over all values and verify that they are even. However,
failing this check would just print "bad!" instead of failing the test.
Bug: https://bugzilla.gnome.org/show_bug.cgi?id=725981
Reviewed-by: Colin Walters <walters@verbum.org>
Signed-off-by: Simon McVittie <simon.mcvittie@collabora.co.uk>
gdb is run in batch mode, and can leave leave the program being
executed/debugged running when the batchfile is finished. Explicitly
"quit"ing the subprocess prevents it from leaving the stray subprocess
when gdb finishes.
https://bugzilla.gnome.org/show_bug.cgi?id=731366
It turns out that due to a recent gdm change, the inherited
signal mask has SIGUSR1 blocked - which is bad news for
tests using SIGUSR1. Fix the test by explicitly checking the
signal mask before using SIGUSR1.
The mapping-test is failing under gnome-continuous. I suspect this
is simply due to running many tests in parallel, and mapping-test
being racy. Replace the blind sleep by signals, to avoid the
races.
g_time_val_from_iso8601 was attempting to parse strings
having only a date, but failed to actually set the timeval
despite returning TRUE. Since the docs state that the function
only parses strings containing a date and a time, just return
FALSE in this case.
Also remove an incomplete testcase for this behaviour that was
just checking the boolean return value, but not timeval.
In Windows development environments that have it, <unistd.h> is mostly
just a wrapper around several other native headers (in particular,
<io.h>, which contains read(), close(), etc, and <process.h>, which
contains getpid()). But given that some Windows dev environments don't
have <unistd.h>, everything that uses those functions on Windows
already needed to include the correct Windows header as well, and so
there is never any point to including <unistd.h> on Windows.
Also, remove some <unistd.h> includes (and a few others) that were
unnecessary even on unix.
https://bugzilla.gnome.org/show_bug.cgi?id=710519
...and only include unistd.h when we are on *NIX.
Newer Visual C++ runtimes (8.0/2005 and later) will cause the program to
crash with an internal abort() call when they detect instances of close()
being called on an invalid fd, such as when the fd is -1, and these should
be purged anyways.
https://bugzilla.gnome.org/show_bug.cgi?id=711047
