Note that it does not actually work, though. Maybe because
g_io_channel_set_flags() is not implemented for file descriptor based
GIOChannels on Windows.
This is a minimal patch-out of the galias functionality. We will do a
release like this so that we can easily back it out if there are
reported problems.
A more substantial cleanup (mostly removing #includes from every file)
will follow if there are no issues.
* glib/gtestutils.c: don't treat non-fatal messages as errors
* glib/tests/.gitignore: silence
* glib/tests/Makefile.am: updated
* glib/tests/protocol.c: add a test for the message treatment
GBinding is a simple, opaque object that represents a binding between a
property on a GObject instance (source) and property on another GObject
instance (target).
https://bugzilla.gnome.org/show_bug.cgi?id=348080
This patch adds subunit support to gtester-report via a -s switch. Subunit
(https://launchpad.net/subunit) is a language neutral test activity protocol.
This can be used to integrate gtester tests with the Hudson CI tool amongst
other things.
Bug #611869.
Take advantage of our knowledge that GVariant strings are always valid
utf8 when printing and parsing:
- allow valid printing unicode characters to pass through unescaped
- escape non-printing characters using \uxxxx or \Uxxxxxxxx format
- do the same in the parser
- update existing test cases to use utf8, add a new test case
Add GObject introspection annotations so that the length parameter is
correctly detected for g_variant_new_strv(), g_variant_get_strv() and
g_variant_dup_strv(). Also specify that it can be a NULL pointer in
g_variant_get_strv() and g_variant_dup_strv().
For g_settings_set_strv(), detect that a NULL value is allowed, meaning
empty array.
Closes bug #620384.
Signed-off-by: Ryan Lortie <desrt@desrt.ca>
This adds static markers for dtrace, which are also usable
by systemtap. Additionally it adds a tapset for systemtap
that makes it easier to use the static markers.
These are enabled by default.
This initial set of probes is rather limited:
* allocation and free using g_malloc & co
* allocation and free using g_slice
* gquark name tracking (useful for converting quarks to strings in probes)
Notes on naming:
Its traditional with dtrace to use probe names with dashes as
delimiter (slice-alloc). Since dashes are not usable in identifiers
the C code uses double underscores (slice__alloc) which is converted
to dashes in the UI. We follow this for the shared lowlevel probe
names.
Additionally dtrace supports putting a "provider" part in the probe
names which is essentially a namespacing thing. On systemtap this
field is currently ignored (but may be implemented in the future), but
this is not really a problem since in systemtap the probes are
specified by combining the solib file and the marker name, so there
can't really be name conflicts.
For the systemtap tapset highlevel probes we instead use names that
are systemtapish with single dashes as separators.
https://bugzilla.gnome.org/show_bug.cgi?id=606044
Rather make it branch to get the due sequence length for the resulting
character code, we can as well get the minimum code value in the initial
branching.
FreeBSD's malloc() sometimes returns unaligned memory if you are
requesting small sizes. This can get GVariant into trouble. For
example, consider the type "mmi" containing the value "just nothing".
According to the type signature, the memory containing this should be
aligned to a boundary of 4 since it might contain an int. The
serialised size of this value is 1 byte, however, and when you ask
FreeBSD to allocate memory of that size, it knows you can't put an int
into it so it doesn't bother aligning it.
This patch modifies the GVariant serialiser to not assert the alignment
constraint in the case that the size of the serialised data is smaller
than its own alignment requirement.
The GVariant serialiser works well with non-8-aligned memory, but the
comparison serialiser in the test case depends on memory being
8-aligned. Use posix_memalign() to get the memory used by this
serialiser.
These allow applications to give meaningful names to their sources.
Source names can then be used for debugging and profiling, for
example with systemtap or gdb.
https://bugzilla.gnome.org/show_bug.cgi?id=606044
These allow applications to give meaningful names to their sources.
Source names can then be used for debugging and profiling, for
example with systemtap or gdb.
https://bugzilla.gnome.org/show_bug.cgi?id=606044
It's possible that a given binary may conditionally decided not to run
any test cases (e.g. since they are all slow but -m=quick is currently
in use) In this case the xml may contain <testbinary> nodes with no
<testcase> children. This was resulting in a divide by zero when
calculating the green → red color interpolation.
https://bugzilla.gnome.org/show_bug.cgi?id=617914
- modify serialiser validation function to enforce utf8 encoding
- add documentation to g_variant_new_string(), g_variant_get_string(),
g_variant_dup_string()
- add 2 new test cases to check that it works
It would be good, error reporting-wise, to be able to signal which
function should be used instead of a deprecated one. GCC 4.5 added an
optional "message" payload to the deprecated attribute, so that:
void f1 (void) __attribute__((deprecated("Use f2 instead")));
Will expand to:
warning: f1 is deprecated: Use f2 instead
Instead of just printing:
warning: f1 is deprecated
Since we already have a G_GNUC_DEPRECATED macro we should provide a
G_GNUC_DEPRECATED_FOR macro defined as:
G_GNUC_DEPRECATED_FOR(bar)
Which would expand the deprecation message to "Use bar instead"
automatically. The deprecation message should probably be similar
to what we use in gtk-doc to match up with the documentation.
https://bugzilla.gnome.org/show_bug.cgi?id=614965
Define GStatBuf as the type used by g_stat() and g_lstat(). Replaces
the non-public struct tag _g_stat_struct. Mostly relevant for Windows
where there are several variants of stat-style structs. On POSIX, is
just another name for struct stat.
Actually, also on many POSIX systems there are in fact several
variants of struct stat and corresponding stat() and lstat()
functions, but as g_stat and g_lstat are normally on POSIX just macros
that expand to stat and lstat, this should not cause a problem. It's
only when it's the actual g_stat() or g_lstat() implementation inside
GLib that gets called that one needs to be sure the passed struct is
the same as what GLib expects.)
Don't keep the lists of source files for libglib, libgobject and
libgio in the VS project files in addition to the canonical location,
the corresponding Makefile.am files.
Instead, generate the corresponding .vcproj files at make dist time
using the C preprocessor, from template files called .vcprojin. We
still list explicitly in the .vcprojin files some of the
Windows-specific source files, and the sources files of gnulib and
pcre.
NaN floating point values get mangled when passing across the function
call ABI on x86 so avoid using them to get rid of spurious failures.
Reported by Christian Persch and reliably reproduced by Emilio Pozuelo
Monfort.
Use g_io_add_watch(), g_io_add_watch_full() and
g_source_remove() instead the deprecated gtk_input_add_full(),
gtk_input_remove(), gdk_input_add(), gdk_input_add_full() and
gdk_input_remove()
Merge GVariant variable arguments support and put it under tests.
Also, remove the hack of the test case directly '#include'ing .c files
from glib/. Instead, create a non-installed gvariant-internal.h that
the tests can include and make the symbols in it visible on the symbol
table of the shared library. These symbols (as they are present in no
installed header files) are not part of the API of GLib.
Increase test coverage in a few other areas.
Remove the macros for the g_malloc_n family -- calls directly to those
functions now always go directly to those functions.
Reimplement the macros for g_new and friends.
Remove the branch that checked for calling g_new() with a constant
n_structs == 1. With the struct size always known this case will now be
caught under the case that does the inline multiplication and the
multiplication by 1 will be optimised away.
Don't check that a GVariantBuilder is fully-valid for use when
performing refcounting operations, as these operations may be occuring
after the builder has already been cleared.
Issue caught by Michael Meeks. This patch based on his.
GVariantTypeInfo was dropping the reference count on the TypeInfo
structure before removing it from the hash table. This means that
another thread could come along and grab the value from the hash table
in the meantime.
Solve this by holding the lock on the table before dropping the
reference.
Also: move the hash table initialisation inside of the lock to remove
the standard double-initialisation race plus a more insidious issue
caused by the fact that we free the hash table once it becomes empty.
1) always emit ending type FIXED for fixed sized items
The serialiser needs to know if it dealing with a fixed sized item
in case the serialised data is corrupt and the item has the wrong
size.
2) add a macro that will be used by test cases to check that all
memory has been freed.
Fixes: Bug 599197 - array ref and unref functions crash on NULL
array.
* garray.c: Added safety guards to all public functions, which did
not have them earlier. Now when NULL is passed to them, they will
issue a warning and return, instead of segfaulting.
Add support for a mutex lock that consumes only one bit of storage
inside of an integer on systems that support futexes. Futex is emulated
(at a higher cost) on systems that don't have it -- but only in the
contended case.
In the case of an empty file, return NULL instead of "". This means
that the return result of g_mapped_file_get_contents() will always be
page-aligned.
Re-using glibc's __abort_msg symbol causes linking problems, since the symbol
is declared private. Always use our own__glib_abort_msg symbol to store
assertion messages, to avoid compatibility and linking problems.
Also fix the test case to work with out of tree builds (such as "make
distcheck"), and re-enable it.
https://bugzilla.gnome.org/show_bug.cgi?id=594872
Fixes: Bug 568760 - nautilus freezes due to a bug in garray.c:322
* glib/garray.c: increase the size of potential return values by
using an unsigned result; properly check if we still handle valid size
proposals, return the original request if there's no usable size left
* tests/array-test.c: reproduce the error condition of the bug report
Since gobject now initialized threads unconditionally if threads are
available it makes no sense to have each thread operation
conditionally check if threads are enables, so allow this to be avoided.
https://bugzilla.gnome.org/show_bug.cgi?id=606775
Crash interception/debugging systems like Apport or ABRT capture core dumps for
later crash analysis. However, if a program exits with an assertion failure,
the core dump is not useful since the assertion message is only printed to
stderr.
glibc recently got a patch which stores the message of assert() into the
__abort_msg global variable.
(http://sourceware.org/git/?p=glibc.git;a=commitdiff;h=48dcd0ba)
That works fine for programs which actually use the standard C assert() macro.
This patch adds the same functionality for glib's assertion tests. If we are
building against a glibc which already has __abort_msg (2.11 and later, or
backported above git commit), use that, otherwise put it into our own field
__glib_assert_msg.
Usage:
$ cat test.c
#include <glib.h>
int main() {
g_assert(1 < 0);
return 0;
}
$ ./test
**ERROR:test.c:5:main: assertion failed: (1 < 0)
Aborted (Core dumped)
$ gdb --batch --ex 'print (char*) __abort_msg' ./test core
[...]
$1 = 0x93bf028 "ERROR:test.c:5:main: assertion failed: (1 < 0)"
https://bugzilla.gnome.org/show_bug.cgi?id=594872
One of the option context tests relied on the ability to
call g_set_pgrname() more than once. Thats no longer possible
(without incurring a warning that wrecks the test).
If g_io_win32_sock_close() has been called on a socket channel, don't
later in g_io_win32_free() call WSAEventSelect() on its fd which has
been set to -1.
The patch ensures that multibyte character sets are properly
nul-terminated. This is an issue because the documentation claims to
return a "nul-terminated" string and users of the API assume that means
a proper nul-termination in the resulting character set.
I looked at Pidgin and GStreamer code and found at least 3 cases where
this was happening.
This patch also reverts the documentation change from
5a633f82db as that is now no longer
necessary.
Implement g_test_log_set_fatal_handler which is a function similar to
g_log_set_default_handler but for use in unit tests where certain
errors have to be ignored because it is not possible to fix or avoid
them otherwise. A unit test is added.
This function exposes more variables than g_mkstemp() and therefor
allows more flexibility when creating temporary files.
The intended use is gio's code for g_file_replace() (see next patch)
This means we don't sync in the case where we created an (empty)
temp file and now replace it with the data.
This fixes (among other things) the performance of trashing files.
'-h' is the preferred short version of '--help' now and displayed
in '--help' but for backwards compatibility '-?' is still supported.
If existing code uses '-h' for something else, GOptionContext will
not override it.
Fixes bug 556706.
glib/pcre/pcre_ucp_search_funcs.c, glib/pcre/pcre_valid_utf8.c: add
back missing config.h includes, and this time add them to the copies
in glib/update-pcre/ too so they don't get lost again on the next PCRE
update.
glib/garray.c, glib/gbase64.c: fix signed/unsigned pointer casts
gio/xdgmime/xdgmimeglob.c: remove unused variable
gio/tests/live-g-file.c: fix printf args on x86_64
tests/Makefile.am, tests/regex-test.c: remove redundant -DENABLE_REGEX
Bug 572508 – gmarkup speedup ...
* glib/gmarkup.c: Various optimizations: do less allocations by
keeping a pool of GStrings, do in-place unescaping, avoid redundant
utf-8 validation.
