* Remove an unneeded field from LaunchUrisData and add annotations
* Rename local GError* variables to local_error
* Use g_set_object
* Fix indentation
...of the application if many URI's are provided. This is important to note
because the GAppLaunchContext signals may be emitted multiple times during
a single launch operation.
We cannot cancel a spawn operation, but sometimes we have to
spawn the target application mutiple times (e.g. in case the
target app only supports one URI in its command-line, but we
were given multiple URI's), in that case continuously check
the cancellation status before attempting any spawn operation
First, there's no reason not to use the new `epoll_create1` system call,
which quickly obsoleted `epoll_create` which has an obsolete and
unused size argument.
But more specifically, it offers `EPOLL_CLOEXEC` which we want
to use for general hygeine - there's no reason to potentially
leak this file descriptor to forked processes.
(GLib itself carefully closes file descriptors when forking child
processes, but it may be linked with other software that doesn't;
notably in my case for example the Rust standard library does not
do this and hence relies more on the application code using
`O_CLOEXEC` and variants)
This is just a drive-by fix; I saw the system call when I was using
`strace` to debug something else in rpm-ostree.
This utility function will be called by both launch_uris and
launch_uris_async, passing a from_task parameter respectively
as NULL and non-NULL. The from_task parameter will be needed
to know whether GAppLaunchContext signals should be emitted
directly (from_task == NULL) or scheduled for emission on the
main thread (from_task != NULL).
All of these warnings indicate programmer error, so critical is most
appropriate here.
Exceptions: deprecation warnings are just warnings. Also, warnings that
are worded with uncertainty can remain warnings rather than criticals.
Using the Application Activation Manager coclass. Its threading model
is marked as 'both', so it can be instantiated in any apartment type
without marshaling.
gio tool has support for deleting attributes of the file. To delete attribute user
should specify type '--type="unset"'. This is not mentioned in help and therefore not
intuitive. By adding '-d' option, we make this process more obvious.
closes#2588
The prefix for GMarkupParseFlags enumeration members is G_MARKUP; this
means that G_MARKUP_PARSE_FLAGS_NONE gets split into
GLib.MarkupParseFlags.PARSE_FLAGS_NONE by the introspection scanner.
The `/*< nick=none >*/` trigraph attribute is a glib-mkenum thing, and
does not affect the introspection scanner; it would also only affect the
GEnumValue nickname, which is not used by language bindings to resolve
the name of the enumeration member. Plus, GMarkupParseFlags does not
have a corresponding GType anyway.
The prefix is G_TLS_CERTIFICATE, not G_TLS_CERTIFICATE_FLAGS. Having
G_TLS_CERTIFICATE_FLAGS_NONE leads to a FLAGS_NONE nick in the GType,
and a FLAGS_NONE member name in the introspection data.
Enumeration members should either have the name of the type as their
prefix, or they should all have the same prefix.
The "default flags" enumeration member for GApplicationFlags is
unfortunately named G_APPLICATION_FLAGS_NONE, while every other member
of the same type has a G_APPLICATION prefix. The result is that the nick
name of the enumeration member is "flags-none", and that language
bindings will have to use something like
Gio.ApplicationFlags.FLAGS_NONE.
To fix this API wart, we can deprecate the FLAGS_NONE member, and add a
new DEFAULT_FLAGS.
If stdout is the Journal but stderr is not, then we probably only want
to redirect stdout, or vice versa.
Signed-off-by: Simon McVittie <smcv@collabora.com>
This prevents a launched process's output from being mixed up with the
output of the parent process, which can lead to the wrong program being
blamed for warning messages.
Signed-off-by: Simon McVittie <smcv@collabora.com>
This is an internal helper executable, which users shouldn't invoke
directly (see glib#1633).
When building for a single-architecture distribution, we can install
it as ${libexecdir}/gio-launch-desktop.
When building for a multiarch distribution, installing it into an
architecture-specific location and packaging it alongside the GLib
library avoids the problem discussed in glib#1633 where it would either
cause a circular dependency between the GLib library and a common
cross-architecture package (libglib2.0-bin in Debian), or require a
separate package just to contain gio-launch-desktop, or cause different
architectures' copies to overwrite each other.
Signed-off-by: Simon McVittie <smcv@collabora.com>
gio-launch-desktop was removed before checking GIO for potentially
unsafe environment variable references, so reverting its removal brought
this one back. If a setuid program is using GAppInfo then something is
probably already horribly wrong, but let's be careful anyway.
Signed-off-by: Simon McVittie <smcv@collabora.com>
A shell one-liner was enough to set GIO_LAUNCHED_DESKTOP_FILE_PID,
but ideally we also want to do the equivalent of sd_journal_stream_fd()
to set up its standard output and standard error streams.
Ideally we would call sd_journal_stream_fd() in a process that will
exec the real program, otherwise it will report the wrong process ID
in the Journal, but we can't easily do that in a forked child when
using posix_spawn() for subprocesses.
This reverts commit 2b533ca99a.
Signed-off-by: Simon McVittie <smcv@collabora.com>
The dominant implementations of the well-known session and system
message buses are the reference implementation from the dbus project
(dbus-daemon) and the sd-bus-based reimplementation dbus-broker, both
of which have correct implementations for EXTERNAL authentication with
an unspecified authorization identity.
This makes it reasonably safe to assume that the well-known message
buses can cope with the unspecified authorization identity, even if we
cannot make the same assumption for custom servers such as the ones
used in ibus and gvfs (which might have been started with an older
GLib version before upgrading GLib in-place).
Signed-off-by: Simon McVittie <smcv@collabora.com>
When using a GDBus client in a non-trivial user namespace, the result of
geteuid() can differ from the uid in the namespace where the server is
running. This would result in connection attempts being rejected, because
the identity that the client claims to have does not match the identity
that the server derives from its credentials.
RFC 4422 allows us to send an empty authorization identity, which means we
want to authenticate as whatever identity the server can derive from our
out-of-band credentials. In particular, this resolves the authentication
failure when crossing between different Linux user namespaces.
Because D-Bus does not have a way to represent an empty initial response
as distinct from the absence of an initial response, we cannot use the
initial-response optimization (RFC 4422 §4.3.a) in this case, and must
fall back to waiting for the server to send a challenge.
Unfortunately, GDBus versions older than glib!2826 did not implement
the server side of this protocol correctly, and would respond to the
missing initial response in a way that breaks the SASL state machine
(expecting a response without sending a challenge), causing client and
server to deadlock with each waiting for the other to respond. Until
fixed versions of GDBus are widespread, we can't rely on having a server
that can cope with this, so gate it behind a flag, which can be set for
connections that are known to cross non-trivial namespace boundaries.
Originally inspired by
<1ed4723d38>,
and based on earlier work by Giuseppe Scrivano (in which the
cross-namespace behaviour was unconditional, rather than gated by a
flag).
Co-authored-by: Giuseppe Scrivano <giuseppe@scrivano.org>
Signed-off-by: Simon McVittie <smcv@collabora.com>
In Debian-style multiarch (libdir = lib/x86_64-linux-gnu or similar),
Red-Hat-style multilib (libdir = lib64 or lib) and Arch-style multilib
(libdir = lib or lib32), we have to run a separate version of
gio-querymodules to discover 32- or 64-bit modules on x86. Installing
modules in the directory used for each word size needs to trigger
recompilation of the correct modules list.
Debian, Fedora and Arch currently all have patches to facilitate this:
Debian moves gio-querymodules into ${libdir}/glib-2.0 and provides a
compat symlink in ${bindir}, while Fedora and Arch rename one or both
of the gio-querymodules executables to give it a -32 or -64 suffix.
We can avoid the need for these patches by making this a build option.
Doing this upstream has the advantage that the pkg-config metadata for
each architecture points to the correct executable and is in sync with
reality.
I'm using Debian's installation scheme with a separate directory here,
because the word-size suffix used in Fedora and Arch only works for the
common case of 32- and 64-bit multilib, and does not cover scenarios
where there can be more than one ABI with the same word size, such as
multiarch cross-compilation or alternative ABIs like x32.
Now that we have this infrastructure, it's also convenient to use it for
glib-compile-schemas. This works with /usr/share, so it only needs to
be run for one architecture (typically the system's primary
architecture), but using /usr/bin/glib-compile-schemas for the trigger
would result in either primary and secondary architectures trying to
overwrite each other's /usr/bin/glib-compile-schemas binaries, or a
circular dependency (the GLib library would have to depend on a
common package that contains glib-compile-schemas, but
glib-compile-schemas depends on the GLib library). Installing a
glib-compile-schemas binary in an architecture-specific location
alongside each GLib library bypasses this problem.
Signed-off-by: Simon McVittie <smcv@collabora.com>
Instead of using a GDBusConnection, this does the handshake at a lower
level using specific strings in the SASL handshake, to verify that we
will interoperate with various clients including sd-bus, libdbus and
older versions of GDBus.
Signed-off-by: Simon McVittie <smcv@collabora.com>
This is an interoperability fix. The reference implementation of D-Bus
treats "DATA\r\n" as equivalent to "DATA \r\n", but sd-bus does not,
and only accepts the former.
Signed-off-by: Simon McVittie <smcv@collabora.com>
RFC 4422 appendix A defines the empty authorization identity to mean
the identity that the server associated with its authentication
credentials. In this case, this means whatever uid is in the
GCredentials object.
In particular, this means that clients in a different Linux user
namespace can authenticate against our server and will be authorized
as the version of their uid that is visible in the server's namespace,
even if the corresponding numeric uid returned by geteuid() in the
client's namespace was different. systemd's sd-bus has relied on this
since commit
1ed4723d38.
[Originally part of a larger commit; commit message added by smcv]
Signed-off-by: Simon McVittie <smcv@collabora.com>
Sending an "initial response" along with the AUTH command is meant
to be an optional optimization, and clients are allowed to omit it.
We must reply with our initial challenge, which in the case of EXTERNAL
is an empty string: the client responds to that with the authorization
identity.
If we do not reply to the AUTH command, then the client will wait
forever for our reply, while we wait forever for the reply that we
expect the client to send, resulting in deadlock.
D-Bus does not have a way to distinguish between an empty initial
response and the absence of an initial response, so clients that want
to use an empty authorization identity, such as systed's sd-bus,
cannot use the initial-response optimization and will fail to connect
to a GDBusServer that does not have this change.
[Originally part of a larger commit; commit message added by smcv.]
Signed-off-by: Simon McVittie <smcv@collabora.com>
This is an interoperability fix. If the line is exactly "DATA\r\n",
the reference implementation of D-Bus treats this as equivalent to
"DATA \r\n", meaning the data block consists of zero hex-encoded bytes.
In practice, D-Bus clients send empty data blocks as "DATA\r\n", and
in fact sd-bus only accepts that, rejecting "DATA \r\n".
[Originally part of a larger commit; commit message added by smcv]
Signed-off-by: Giuseppe Scrivano <giuseppe@scrivano.org>
Co-authored-by: Simon McVittie <smcv@collabora.com>
Signed-off-by: Simon McVittie <smcv@collabora.com>
This attribute will produce "deprecation" warnings when using it in
code that does not want dependencies on newer GLib versions.
Signed-off-by: Simon McVittie <smcv@collabora.com>
Using GMemorySettingsBackend before any other GSettingsBackend would
cause the following error: "Tried to implement non-registered extension
point gsettings-backend". This is due to a missing call to
_g_io_modules_ensure_extension_points_registered() in the GMemorySettingsBackend
type definition which registers the gsettings-backend extension point.
We don't need a cpp toolchain for building glib so lets just
automatically disable tests requiring one when not available.
Signed-off-by: James Hilliard <james.hilliard1@gmail.com>
The test was flacky because we were only relying on the presence of a
file, while the callback could have not been called yet, while ensure
for both assumptions to be true before stop iterating the loop.
Fix a regression from commit abddb42d14, where it could pass `NULL` to
`g_task_get_cancellable()`, triggering a critical warning. This could
happen because the lifetime of `data->task` is not as long as the
lifetime of the `ConnectionAttempt`, but the code assumed it was.
Fix the problem by keeping a strong ref to that `GCancellable` around
until the `ConnectionAttempt` is finished being destroyed.
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
Fixes: #2687
These headers have all been written manually, by looking through the git
log for each file and noting the copyright of each significant
contribution.
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
Helps: #1415
While `gio_xdgmime` is unlocked, the data which `type` points to in the
xdgmime cache might get invalidated, leaving `type` as a dangling
pointer. That would not bode well for the `g_strdup (type)` call to
insert a new entry into the `type_comment_cache` once `gio_xdgmime` is
re-acquired.
This was spotted using static analysis, and the symptoms have not
knowingly been seen in the wild.
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
Coverity CID: #1474702
This will catch buggy implementations of GProxyResolver before they are
able to return bogus results to higher level code. In particular, if
g_proxy_resolver_lookup() returns NULL, it'd better set an error to
explain why.
It doesn't make sense for a proxy resolver to return NULL without an
error on the first call. Whereas a DNS resolver would do this to
indicate that a query completed successfully but found no results, a
proxy resolver should return "direct://" instead. Therefore, if we are
going to return NULL, we ought to have an error as well. Let's make sure
this actually happens by adding some fallback errors just in case
GProxyResolver feeds us weird results.
Additionally, we should not return any errors except
G_IO_ERROR_CANCELLED after the very first iteration. This is an API
contract of GSocketAddressEnumerator. Let's add some checks to ensure
this.
Note that we have inadequate test coverage for GProxyAddressEnumerator.
It's tested here only via GSocketClient. We could do a bit better by
testing it directly as well. For example, I've added tests to see what
happens when GProxyResolver returns both a valid and an invalid URI, but
it's not so interesting here because GSocketClient always uses the valid
result and ignores the error from GProxyAddressEnumerator.
Fixes#2597
This has no practical impact, since it's only a test, and none of the
test code would have hit this bug, but the GTestProxyResolver's check to
see if the URI scheme is simple:// currently only compares the first
four bytes of the string, so it's actually only checking for the "simp"
and would match anything else after that, e.g. "simpleton://". This is
surely not intended.
This was causing intermittent failures on macOS, depending on whether
the tmpdir ended with a `/` or `/some-dir`. `g_strrstr()` is not the
right function to use to extract a basename from a path, for this
reason.
When it failed, the macOS test was failing with:
```
ok 16 /gsubprocess/env
Bail out! GLib-GIO:ERROR:../gio/tests/gsubprocess.c:1507:test_cwd: assertion failed (basename == tmp_lineend_basename): ("/T\n" == "/\n")
```
The test now passes reliably, which means that it can be removed from
the list of expected failures on macOS.
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
Helps: #1392
Creating a `GAppInfo` from a commandline isn’t currently supported on
macOS, but the implementation was incorrectly returning `NULL` without
setting the `GError`.
This was being caught by the new tests in `gio/tests/file.c`.
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
On macOS the comparison was failing as one of the paths had a trailing
slash while the other didn’t.
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
We used to do get and set atomic operations pair, but these may be
unsafe in some cases as threads may rely on data that is changed in
in between them, however this is not a problem if we do exchange the
pointers.
So just use exchange ops, in this way we can avoid lock/unlock mutex
dances
This makes calls to g_signal_connect_data() and g_signal_connect_object()
with default flags more self-documenting.
Signed-off-by: Simon McVittie <smcv@collabora.com>
Despite the name, the call was still doing blocking operations when
looking for app handlers via GAppInfo. Now it's possible to use fully
async calls.
Together with previous commit, the API is now fully async.
Despite the name, we still used blocking calls to get the default app
for URI, now that we have an async implementation of the API to get the
default implementation for URI scheme, we can remove the blocking calls.
We did not check whether this function worked before, so add a simple
test case for it, providing some test functions to make it possible to
reply the same behavior
In some cases (such as in our CI tests) we may not have any dbus session
set after launching, but we always assumed so.
In case we have not a session bus set, we only have to return early.
In both callbacks of g_file_query_default_handler_async() we were
actually using I/O blocking APIs making it not fully async.
Now that such API is provided, we can use it.
While it's possible to create a directory synchronously via
g_dir_make_tmp(), there's no such API that performs it asynchronously.
So implement it using GFile, using a thread to perform such task.
Avoid re defining cases for GIoErrorEnum when we already handle them
through GFileError, so remove code duplication and just rely on
g_file_error_from_errno() to compute the file error and then use
g_io_error_from_file_error() to get the possible IOError.
In case it's something not handled as GFileError, we can use the same
logic as before.
This is now a safe change to do as we have covered all the supported
cases in tests.
It seems this script has potentially never worked properly under Python
3. It’s supposed to list all the `_get_type()` functions it can find in
the GIO headers, but since the regex string passed to `re.search()` was
not a Python regex, nothing was matching.
Fix that, and do another few small cleanups to the script.
This makes the `defaultvalue` test not skip all the types.
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
When running under a strict TAP parser this was previously producing
problematic non-conforming output.
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
This is unlikely to be a bug in practice, as the certificate pointed to
by `root` should have a ref held on it as the issuer of another
certificate in the chain.
However, we can’t guarantee that’s how the `GTlsCertificate`
implementation behaves, so keep a temporary ref on `root` until it’s no
longer needed.
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
Coverity CID: #1489985
Rather than carrying the copylib around inside GLib, which is a pain to
synchronise and affects our code coverage statistics.
This requires updating the CI images to cache the new subproject,
including updating the `cache-subprojects.sh` script to pull in git
submodules.
It also requires adding `gioenumtypes_dep` to be added to the
dependencies list of `libgio`, since it needs to be build before GVDB as
it’s pulled in by the GIO headers which GVDB includes.
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
Helps: #2603
The interface was ready for this API but it was not provided.
So implement this, using a thread that calls the sync API for now.
Add tests.
Helps with: GNOME/glib#157
`GSocketClient` chains its internal `GCancellable` objects to ones
provided by the caller in two places using `g_cancellable_connect()`.
However, it never calls `g_cancellable_disconnect()`, instead relying
(incorrectly) on the `GCancellable` provided by the caller being
short-lived.
In the (valid) situation where a caller reuses one `GCancellable` for
multiple socket client calls, or for calls across multiple socket
clients, this will cause the internal `GCancellable` objects from those
`GSocketClient`s to accumulate, with one reference left each (which is
the reference from the `g_cancellable_connect()` closure).
These `GCancellable` instances aren’t technically leaked, as they will
all be freed when the caller’s `GCancellable` is disposed, but they are
no longer useful and there is no bound on the number of them which will
hang around.
For a program doing a lot of socket operations, this still-reachable
memory usage can become significant.
Fix the problem by adding paired `g_cancellable_disconnect()` calls.
It’s not possible to add a unit test as we can’t measure still-reachable
memory growth before the end of a unit test when everything has to be
freed.
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
Fixes: #2670
Dynamically, these will only ever be used after they’ve been initialised
due to correct checking of `use_udp` throughout the test. However,
that’s a global variable and the static analyser is assuming it might
change value. So help it out by NULL-initialising the variables so they
can never be used uninitialised.
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
This is a fallback timeout to abort the test if the expected number of
messages aren’t seen in time. However, when running the test under
valgrind it will take longer and sometimes spuriously trigger the
timeout.
There’s no point in having an abort timeout inside the test: the test
runner (Meson) already provides one for us, which we can adjust with a
multiplier when running under valgrind.
So removes the timeout from within the test. This should fix the
gnotification test under valgrind.
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
This re-applies a chunk from commit e63262d49d which was
accidentally lost when upstreaming the commit to xdgmime (as
https://gitlab.freedesktop.org/xdg/xdgmime/-/merge_requests/10).
The upstreamed commit was then re-backported to GLib as a1bfe899ab,
without the missing chunk.
The missing chunk is potentially causing incorrect content type results
for `file://` URIs when used from webkitgtk.
Thanks to Stephen Jung and Michael Catanzaro for investigating.
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
Fixes: #2639
These have all been added manually, as I’ve finished all the files which
I can automatically detect.
All the license headers in this commit are for LGPL-2.1-or-later, and
all have been double-checked against the license paragraph in the file
header.
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
Helps: #1415
This may have been causing an intermittent failure of the pollable test
on BSD, where updating the readable status of a socket takes a bit
longer than on Linux.
```
GLib-GIO-DEBUG: 16:06:41.235: GSocketClient: Starting application layer connection
GLib-GIO-DEBUG: 16:06:41.235: GSocketClient: Connection successful!
Bail out! GLib-GIO:ERROR:../gio/tests/pollable.c:73:check_source_readability_callback: assertion failed (readable == expected): (0 == 1)
```
I have not debugged the test on BSD, though, so this is only a guess.
See https://gitlab.gnome.org/GNOME/glib/-/jobs/2022087
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
This should make each unit test a bit more self-contained and easier to
verify that they’re independent.
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
This test is opportunistic in that it’s not possible to detect whether
the race condition has been hit (other than by hitting a deadlock).
So the only approach we can take for testing is to loop over the code
which has previously been known to cause a deadlock a number of times.
The number of repetitions is chosen from running the test with the
deadlock fix reverted.
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
Helps: #1941
This should prevent unbounded growth of the `event_queue` in the
unlikely case that the `GSource` is removed from its `GMainContext` and
destroyed separately from the `GFileMonitor`.
I’m not sure if that can currently happen, but it could with future
refactoring, so it’s best to address the possibility now while we’re
thinking about this bit of code.
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
Helps: #1941
Taking a reference to the GFileMonitor when handling events may cause
object destruction from th worker thread that calls the function. This
condition happens if the surrounding code drops the otherwise last
reference ot the GFileMonitor. The series of events causes destruction
from an unrelated worker thread and also triggers g_file_monitor_cancel
to be called from g_file_monitor_source_handle_event.
For the inotify backend, this results in a deadlock as cancellation
needs to take a lock that protects data structures from being modified
while events are dispatched.
One alternative to this approach might be to add an RCU (release, copy,
update) approach to the lists contained in the wd_dir_hash and
wd_file_hash hash tables.
Fixes: #1941
An example stack trace of this happening is:
Thread 2 (Thread 0x7fea68b1d640 (LWP 260961) "gmain"):
#0 syscall () at ../sysdeps/unix/sysv/linux/x86_64/syscall.S:38
#1 0x00007fea692215dc in g_mutex_lock_slowpath (mutex=mutex@entry=0x7fea6911e148 <g.inotify_lock_lock>) at ../glib/gthread-posix.c:1493
#2 0x00007fea69222062 in g_mutex_lock (mutex=mutex@entry=0x7fea6911e148 <g.inotify_lock_lock>) at ../glib/gthread-posix.c:1517
#3 0x00007fea6908025a in _ih_sub_cancel (sub=0x1492620) at ../gio/inotify/inotify-helper.c:131
#4 0x00007fea6907f9da in g_inotify_file_monitor_cancel (monitor=0x14a3550) at ../gio/inotify/ginotifyfilemonitor.c:75
#5 0x00007fea68fae959 in g_file_monitor_cancel (monitor=0x14a3550) at ../gio/gfilemonitor.c:241
#6 0x00007fea68fae9dc in g_file_monitor_dispose (object=0x14a3550) at ../gio/gfilemonitor.c:123
#7 0x00007fea69139341 in g_object_unref (_object=<optimized out>) at ../gobject/gobject.c:3636
#8 g_object_unref (_object=0x14a3550) at ../gobject/gobject.c:3553
#9 0x00007fea6907507a in g_file_monitor_source_handle_event (fms=0x14c3560, event_type=<optimized out>, child=0x7fea64001460 "spawned-1", rename_to=rename_to@entry=0x0, other=other@entry=0x0, event_time=<optimized out>) at ../gio/glocalfilemonitor.c:457
#10 0x00007fea6907fe0e in ih_event_callback (event=0x7fea64001420, sub=0x1492620, file_event=<optimized out>) at ../gio/inotify/inotify-helper.c:218
#11 0x00007fea6908075c in ip_event_dispatch (dir_list=dir_list@entry=0x14c14c0, file_list=0x0, event=event@entry=0x7fea64001420) at ../gio/inotify/inotify-path.c:493
#12 0x00007fea6908094e in ip_event_dispatch (event=0x7fea64001420, file_list=<optimized out>, dir_list=0x14c14c0) at ../gio/inotify/inotify-path.c:448
#13 ip_event_callback (event=0x7fea64001420) at ../gio/inotify/inotify-path.c:548
#14 ip_event_callback (event=0x7fea64001420) at ../gio/inotify/inotify-path.c:530
#15 0x00007fea69081391 in ik_source_dispatch (source=0x14a2bf0, func=0x7fea69080890 <ip_event_callback>, user_data=<optimized out>) at ../gio/inotify/inotify-kernel.c:327
#16 0x00007fea691d0824 in g_main_dispatch (context=0x14a2cc0) at ../glib/gmain.c:3417
#17 g_main_context_dispatch (context=0x14a2cc0) at ../glib/gmain.c:4135
#18 0x00007fea691d0b88 in g_main_context_iterate (context=context@entry=0x14a2cc0, block=block@entry=1, dispatch=dispatch@entry=1, self=<optimized out>) at ../glib/gmain.c:4211
#19 0x00007fea691d0c2f in g_main_context_iteration (context=0x14a2cc0, may_block=may_block@entry=1) at ../glib/gmain.c:4276
#20 0x00007fea691d0c81 in glib_worker_main (data=<optimized out>) at ../glib/gmain.c:6176
#21 0x00007fea691f9c2d in g_thread_proxy (data=0x1487cc0) at ../glib/gthread.c:827
#22 0x00007fea68d93b1a in start_thread (arg=<optimized out>) at pthread_create.c:443
#23 0x00007fea68e18650 in clone3 () at ../sysdeps/unix/sysv/linux/x86_64/clone3.S:81
This introduces no functional changes but will make refactoring a bit
easier in the following commit.
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
The user data parameters in callbacks need to be named user_data to
generate correct closure attributes in the introspection data. This
updates parameters missed in GNOME/glib!2633.
Add SPDX license (but not copyright) headers to all files which follow a
certain pattern in their existing non-machine-readable header comment.
This commit was entirely generated using the command:
```
git ls-files gio/tests/*.c | xargs perl -0777 -pi -e 's/\n \*\n \* This library is free software; you can redistribute it and\/or\n \* modify it under the terms of the GNU Lesser General Public/\n \*\n \* SPDX-License-Identifier: LGPL-2.1-or-later\n \*\n \* This library is free software; you can redistribute it and\/or\n \* modify it under the terms of the GNU Lesser General Public/igs'
```
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
Helps: #1415
Add SPDX license (but not copyright) headers to all files which follow a
certain pattern in their existing non-machine-readable header comment.
This commit was entirely generated using the command:
```
git ls-files gio/*.[ch] | xargs perl -0777 -pi -e 's/\n \*\n \* This library is free software; you can redistribute it and\/or\n \* modify it under the terms of the GNU Lesser General Public/\n \*\n \* SPDX-License-Identifier: LGPL-2.1-or-later\n \*\n \* This library is free software; you can redistribute it and\/or\n \* modify it under the terms of the GNU Lesser General Public/igs'
```
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
Helps: #1415
The raw value is escaped according to D-Bus rules. This is probablematic
for Windows backslashed paths. We can use URI unescaping, it seems
that's what gdbusaddress.c is doing too.
Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
The AF_UNIX API is available under all platforms since 2.71.1, and
credentials functions returns NOT_SUPPORTED error appropriately,
we can thus remove the special-casing for !unix.
Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
The "/unix-fd/scm" test is quite Unix-specific, the next patch is going
to add a portable test.
Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Apply --internal to the symbol for the section as well.
Only do this with --external-data, since otherwise gcc
will complain about ignoring the attribute.
This fixes https://gitlab.gnome.org/GNOME/gtk/-/issues/4598
See !2214 for motivation. It doesn't work properly anymore, and the TLS
versions of these functions were already deprecated. Sadly, I missed the
DTLS versions.
Fixes#2646
Use a prefix matching the binary, rename the test "scm", as it involves
SocketControlMessage.
Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
This will help static analysers which think that the type of
`priv->proxy_address` could potentially change between freeing
`dest_hostname` and the `g_return_if_fail()` call below, leading to the
code to continue through to `g_object_new()` and use `dest_hostname`
after freeing it.
This introduces no functional changes.
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
This should fix a scan-build warning about the final `name_unref()` here
being a use-after-free.
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
This should fix a scan-build warning that `resource` is
used-after-freeing in the final `g_resource_unref()` call in
`g_static_resource_fini()`, as `g_resources_unregister_unlocked()` has
already unreffed it.
In reality, each resource has two strong refs on it while active, so the
second unref is correct.
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
It is not only shorter than `not meson.is_cross_build() or
meson.has_exe_wrapper()` but also handle the case of cross compiling to
a compatible arch such as building for i386 on an amd64.
This fixes a scan-build warning:
```
../../../../source/glib/gio/glocalfileinfo.c:1661:28: warning: Although the value stored to 'mydirname' is used in the enclosing expression, the value is never actually read from 'mydirname' [deadcode.DeadStores]
mydirname = g_strdup (dirname),
^ ~~~~~~~~~~~~~~~~~~
```
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
Helps: #1767
`path` was used in building the error message after it had been freed.
Spotted by scan-build.
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
Helps: #1767
This will probably make no functional difference, but will squash two
warnings from scan-build:
```
../../../../source/glib/gio/gsocket.c:503:14: warning: Assigned value is garbage or undefined [core.uninitialized.Assign]
family = address.storage.ss_family;
^ ~~~~~~~~~~~~~~~~~~~~~~~~~
../../../../source/glib/gio/gsocket.c:527:29: warning: Assigned value is garbage or undefined [core.uninitialized.Assign]
socket->priv->family = address.storage.ss_family;
^ ~~~~~~~~~~~~~~~~~~~~~~~~~
```
It seems like a reasonable thing to warn about. Initialising the full
union to zero should avoid any possibility of undefined behaviour like
that.
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
Helps: #1767
This fixes a scan-build warning:
```
../../../../source/glib/gio/tests/gdbus-tests.c:146:3: warning: Value stored to 'watch_id' is never read [deadcode.DeadStores]
watch_id = 0;
^
```
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
Helps: #1767
Tested using:
```sh
touch ~/foo
gio set ~/foo -t bytestring user::test "\x00\x00"
```
(it doesn’t matter that this fails; the bytestring is still decoded)
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
Coverity CID: #1474407
These `memcpy()` calls only happen if `g_inet_address_get_family(group)
== G_SOCKET_FAMILY_IPV4`, so the assertions should never fail.
It’s helpful for understanding the code, and for static analysis, to add
the assertions though.
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
Coverity CID: #1486858
When I enabled unix socketpair test on win32, I left the existing
g_close(fds[1]), but _g_win32_socketpair() returns native sockets
descriptors that must be closed with closesocket() on win32.
Let GSocket handle the socket pair cleanup.
Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
When checking that the connection has the expected number of refs, the
test would block on a `GMainContext` iteration for up to 3s before
waking up and failing (if the refcount was still not as expected).
This check was written in the expectation that changing the refcount of
the connection would only happen due to dispatching a source on
`GMainContext` — hence the `GMainContext` would wake up as the refcount
changed.
That’s probably not actually true though. It might be the case that the
connection’s refcount is changed on from the GDBus worker thread, which
would not cause any wakeups on the main thread’s `GMainContext`.
In this case, the `GMainContext` iteration in
`assert_connection_has_one_ref()` would block for the full 3s, and then
wake up and notice the refcount is correct (then the test would
proceed).
That’s fine, apart from the fact that `test_threaded_singleton()` does
this 1000 times. If the slow case is hit on a significant number of
those test runs, the test will take around 3000s to complete, which is
significantly more than meson’s test timeout of 360s. So the test fails
with something like:
```
220/266 glib:gio+slow / gdbus-threading TIMEOUT 360.07 s
--- command ---
G_TEST_SRCDIR='/builds/GNOME/glib/gio/tests' GIO_MODULE_DIR='' G_TEST_BUILDDIR='/builds/GNOME/glib/_build/gio/tests' /builds/GNOME/glib/_build/gio/tests/gdbus-threading
--- stdout ---
\# random seed: R02S83fe8de22db4d4f376e6d179e2bdd601
1..3
\# Start of gdbus tests
ok 1 /gdbus/delivery-in-thread
ok 2 /gdbus/method-calls-in-thread
\# GLib-GIO-DEBUG: refcount of 0x5602de913660 is not right (3 rather than 1) in test_threaded_singleton(), sleeping
\# GLib-GIO-DEBUG: refcount of 0x5602de913660 is not right (3 rather than 1) in test_threaded_singleton(), sleeping
\# GLib-GIO-DEBUG: refcount of 0x5602de913c60 is not right (3 rather than 1) in test_threaded_singleton(), sleeping
\# GLib-GIO-DEBUG: refcount of 0x5602de913c60 is not right (3 rather than 1) in test_threaded_singleton(), sleeping
\# GLib-GIO-DEBUG: refcount of 0x5602de913260 is not right (3 rather than 1) in test_threaded_singleton(), sleeping
\# GLib-GIO-DEBUG: refcount of 0x5602de913260 is not right (3 rather than 1) in test_threaded_singleton(), sleeping
```
From this log, it can be seen that the sleep is happening on a different
`GMainContext` every other time, so the test *is* making progress.
Assuming this is a correct diagnosis (it’s a lot of guessing), this
commit tries to fix the test by adding a wakeup timeout to the
`GMainContext` in `assert_connection_has_one_ref()`, which will wake it
up every 50ms to re-check the exit condition.
This polling approach has been taken because it doesn’t seem feasible to
make sure that every `g_object_ref()`/`g_object_unref()` call on a
`GDBusConnection` causes the main context to wake up.
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
This might fix a recent test failure:
https://gitlab.gnome.org/GNOME/glib/-/jobs/1929015. Unfortunately
there’s not much debug information in the logs to go on, and I can’t
reproduce it locally. All I have is:
```
192/272 glib:gio / gdbus-peer-object-manager FAIL 0.43 s (killed by signal 11 SIGSEGV)
--- command ---
GIO_MODULE_DIR='' G_TEST_BUILDDIR='/builds/GNOME/glib/_build/gio/tests' G_TEST_SRCDIR='/builds/GNOME/glib/gio/tests' /builds/GNOME/glib/_build/gio/tests/gdbus-peer-object-manager
--- stdout ---
\# random seed: R02Seee9b7325ecd7c19249a3412397aed9b
1..2
\# Start of gdbus tests
\# Start of peer-object-manager tests
```
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
While the assertion always turned out to be true on Linux, it frequently
caused spurious test failures on FreeBSD.
After some remote debugging, I *think* the cause is as written up in the
comment in the code in this commit. However, I cannot be certain, as the
more debugging messages I added, the harder the failure was to
reproduce; and I don’t have access to a FreeBSD machine.
This fixes failures like:
```
Bail out! GLib-GIO:ERROR:../gio/tests/converter-stream.c:1043:test_converter_pollable: assertion failed (error == NULL): Resource temporarily unavailable (g-io-error-quark, 27)
```
It’s succeeded 1000 times in a row on the FreeBSD CI now; previously
it was failing one time in three:
https://gitlab.gnome.org/GNOME/glib/-/jobs/1936395.
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
There’s (deliberately) a bit of race in implementing/handling
`CloseBeforeReturning()` in `gdbus-method-invocation.c`. If the server
closes the D-Bus connection early, the client may exit with `SIGTERM` if
`GDBusConnection:exit-on-close` is set. We don’t want that, as the test
is trying to check that the default handling of a D-Bus method return
after a connection has closed works.
See https://gnome.pages.gitlab.gnome.org/-/glib/-/jobs/1935191/artifacts/_build/meson-logs/testlog.txt
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
G_GNUC_UNUSED does perfectly its job with gcc compiler but the warning
still remains with msvc compiler.
Once the unused variable removed, the finalize vfunc can be removed as
it's doing the same job as the parent function.
We must not open the fd with O_PATH|O_NOFOLLOW,
since the portal rejects that combination. Leaving
out O_NOFOLLOW is fine in this case - we know it
is a directory, we just received EISDIR.
Fixes: #2629
libgamin was last released in 2007 and is dead
[upstream](https://gitlab.gnome.org/Archive/gamin). Distributions may
still ship it (although Fedora no longer does), but we want people to
use inotify on Linux since it’s actively supported.
BSDs use kqueue. Windows uses win32filemonitor.
FAM might still be used on some commercial Unix distributions, but there
are no contributors from those distributions, and certainly no CI for
them to prevent regressions.
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
Fixes: #2614
It used to exist on Solaris, but GLib’s support for it was mostly
removed in 2015 in commit 21ab660cf8.
Remove the final few references.
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
They’re only defined on Unix anyway. `GThreadedResolver` has an entirely
different code path for handling DNS replies on Windows.
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
Success and failure tests. This massively increases test coverage for
parsing DNS records, although it doesn’t get it to 100%.
It should now be useful enough to do more fuzzing on, without
immediately getting trivial failures from the fuzzer.
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
Rather than limiting them to the full length of the answer, which may
include subsequent records.
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
This should catch all kinds of invalid records, and correctly report
them as errors.
Heavily based on work by Patrick Griffis in !2134.
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
Otherwise the code isn’t forwards-compatible, and may be DOSed by
servers returning unknown records, if `G_DEBUG=fatal-warnings` is
enabled for some reason.
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
It is possible for `dn_expand()` to fail; if so, it’ll return `-1`,
which will mess up subsequent parsing.
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
Split out from https://gitlab.gnome.org/GNOME/glib/-/merge_requests/2134
by Philip Withnall so it can be used in advance of HTTPS DNS record
support landing.
Reworked to no longer use test fixtures, as it’s simple enough to build
the response header in each test.
The tests are built on Unix only, as they test the parsing code in
`g_resolver_records_from_res_query()`, which is Unix-specific. The
Windows DNS APIs provide much more structured results which don’t need
parsing.
The SONAME of libutil varies between architectures, so the logic to find
the SONAME of libutil was only correct for native builds (Linux on
Linux), not for cross-builds. The regular expression was also not
sufficiently broad to match the SONAME used on the alpha architecture,
which is apparently libutil.so.1.1.
Instead of screen-scraping the output of ldconfig and using that to
dlopen the library that contains openpty, it seems more reliable to
emit a link-time reference to openpty and let the linker do its job.
It's also less code.
Bug-Debian: https://bugs.debian.org/1007946
Signed-off-by: Simon McVittie <smcv@collabora.com>
These should cover everything to do with returning a value or error from
a `GDBusMethodInvocation` object.
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
`property_info` is only ever set for `Get` and `Set` calls, not for
`GetAll`, as it only represents a single property. So this code was
never reachable.
Move it out so that it is reachable.
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
When doing an early return from `g_dbus_method_invocation_return_*()`
due to passing in the wrong type (or no return value when one was
expected), the parameters were not correctly sunk and were leaked.
Fix that. A unit test will be added in a following commit.
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
Coverity CID: #1474536
The public `g_strv_contains()` API didn’t exist at the time this code
was originally written. Now, happily, it does.
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
The public `g_strv_contains()` API didn’t exist at the time this code
was originally written. Now, happily, it does.
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
When the test has finished writing all the expanded content into the
socket, explicitly close the output stream, which should make the input
stream readable and non-blocking.
The code intended to do this before, but only as a side-effect of
dropping its last reference to `right`. If another reference was being
held to `right` somewhere else, it wouldn’t end up being closed, which
would lead to failures like
https://gitlab.gnome.org/GNOME/glib/-/jobs/1890000:
```
(/var/tmp/gitlab_runner/builds/Ff4WDDRj/0/GNOME/glib/_build/gio/tests/converter-stream:56570): GLib-GIO-DEBUG: 12:56:23.280: GSocketClient: Connection successful!
Bail out! GLib-GIO:ERROR:../gio/tests/converter-stream.c:1042:test_converter_pollable: assertion failed (error == NULL): Resource temporarily unavailable (g-io-error-quark, 27)
stderr:
```
This is a bit of a guess (I’m not sure it’ll fix the intermittent test
error, as I haven’t been able to reproduce that locally), but it’s worth
a try.
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
When multiple tests were run in parallel, this would race on its access
to `~/.dbus-keyrings` to authenticate with the D-Bus server, since the
keyring directory was not appropriately sandboxed to the unit test.
Use `G_TEST_OPTION_ISOLATE_DIRS` to automatically isolate each unit
test’s directory usage.
This should hopefully fix the failure seen in
https://gitlab.gnome.org/GNOME/glib/-/jobs/1849524, where the following
was in the log for a test executed in parallel:
```
GDBus-DBUS_COOKIE_SHA1: Deleted stale lock file '/home/user/.dbus-keyrings/org_gtk_gdbus_general.lock'
```
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
When multiple tests were run in parallel, this would race on its access
to `~/.dbus-keyrings` to authenticate with the D-Bus server, since the
keyring directory was not appropriately sandboxed to the unit test.
Use `G_TEST_OPTION_ISOLATE_DIRS` to automatically isolate each unit
test’s directory usage.
This should hopefully fix the failure seen in
https://gitlab.gnome.org/pwithnall/glib/-/jobs/1879558.
```
228/266 glib:gio / gdbus-non-socket FAIL 8.64 s (killed by signal 6 SIGABRT)
…
--- stderr ---
GDBus-DBUS_COOKIE_SHA1: Deleted stale lock file '/home/user/.dbus-keyrings/org_gtk_gdbus_general.lock'
**
GLib-GIO:ERROR:../gio/tests/gdbus-non-socket.c:253:test_non_socket: assertion failed (error == NULL): Exhausted all available authentication mechanisms (tried: EXTERNAL, DBUS_COOKIE_SHA1) (available: EXTERNAL, DBUS_COOKIE_SHA1) (g-io-error-quark, 0)
```
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
If the whole set of tests takes more than 5 seconds, the failure timeout
from the first test could still trigger, causing an incorrect failure.
Ensure the timeout is removed at the end of each test.i
This will hopefully fix the CI failure seen here:
https://gitlab.gnome.org/pwithnall/glib/-/jobs/1879558.
```
204/266 glib:gio / gdbus-auth FAIL 9.21 s (killed by signal 5 SIGTRAP)
…
ok 1 /gdbus/auth/client/EXTERNAL
Bail out! GLib-GIO-FATAL-ERROR: Timeout waiting for client
```
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
Now that the leaks are fixed, this test can be run under the sanitizer
again.
This is a partial revert of commit
f378352051.
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
Helps: #2312
When destroying `GDBusProxy`s in a custom `GMainContext`, the context
must be iterated enough after finalisation of the proxies that any
pending D-Bus traffic, and the signal subscription data, can be freed.
See the documentation for `g_dbus_connection_signal_unsubscribe()`.
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
Fixes: #2312
This is quite gross, but it looks like the whole content-type code on
Windows is similar. Pass test_subtype.
Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
The `ay` property has type `string` (see the generated code) since it’s
not been annotated to force accepting a `GVariant`.
This means the GObject property machinery expects a string, and calls
`g_strdup()` on the passed-in pointer, rather than sinking the
`GVariant`.
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
Helps: #2312
It periodically hangs due to the `GDBusConnection` having more than 1
ref (and never losing them), so there’s potentially a leaking ref
somewhere:
```
(/builds/alexander.klauer/glib/_build/gio/tests/gdbus-threading:17767): GLib-GIO-DEBUG: 13:18:12.268: refcount of 0x55fe85b1a260 is not right, sleeping
\# GLib-GIO-DEBUG: refcount of 0x55fe85b1a260 is not right, sleeping
```
Add some more debug output to try and track the problem down.
See: https://gitlab.gnome.org/alexander.klauer/glib/-/jobs/1865968
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
It’s periodically failing on FreeBSD and I can’t reproduce the failure
locally nor work out what it is from the logs:
```
Bail out! GLib-GIO:ERROR:../gio/tests/converter-stream.c:1041:test_converter_pollable: assertion failed (res != -1): (-1 != -1)
stderr:
**
GLib-GIO:ERROR:../gio/tests/converter-stream.c:1041:test_converter_pollable: assertion failed (res != -1): (-1 != -1)
```
Add some more debug output to get the value of `error`, in the hope that
will provide some insight.
See: https://gitlab.gnome.org/GNOME/glib/-/jobs/1866486
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
Commit 13c4b9579b seems to have fixed
`network-address` so that it’s reliable everywhere, including on macOS.
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
Helps: #1392
This reverts commit 1ed67a9c44.
It turns out that including options, with their default values, in the
`handle-local-options` signal, which weren’t set on the command line,
breaks some applications.
In particular, it breaks Inkscape, which is the application this commit
was originally meant to fix (a different problem).
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
Fixes: #2611
Breaks: #2329
See: !1953
This allows applications to get their value regardless whether glib is a
subproject or pkgconfig:
gio_dep = dependency('gio-2.0')
giomoduledir = gio_dep.get_variable('giomoduledir')
schemasdir = gio_dep.get_variable('schemasdir')
This reverts commit 5aa03882ca.
It fails to compile on newer Meson versions with
`--fatal-meson-warnings` due to:
```
WARNING: Project targeting '>= 0.52.0' but tried to use feature introduced in '0.54.0': variables arg in declare_dependency.
gio/meson.build:833:0: ERROR: Fatal warnings enabled, aborting
```
That happens regardless of the fact that we’ve correctly limited the use
of the `variables` argument to only when building with Meson ≥ 0.56.
Unfortunately Meson can’t statically detect that the argument is
conditional.
Bumping GLib’s Meson dependency is too much work right now, so this MR
unfortunately has to be reverted.
This allows applications to get their value regardless whether glib is a
subproject or pkgconfig:
gio_dep = dependency('gio-2.0')
giomoduledir = gio_dep.get_variable('giomoduledir')
schemasdir = gio_dep.get_variable('schemasdir')
`mock-resolver.c` is a mock implementation of `GResolver` used in the
`network-address` tests. It returns resolver results, and implements
timeouts, as directed by the test calling it.
In particular, it allows the IPv4 and IPv6 resolver results to be
returned using independent delays. This allows code paths which deal
with IPv4 and IPv6 results being returned at different times to be
tested, as the ‘Happy Eyeballs’ spec mandates various hard-coded
timeouts for returning the best results it can in a reasonable
timeframe.
Previously, `mock-resolver.c` implemented the timeouts by handling
`lookup_by_name()` in a `GTask` worker thread, and calling `g_usleep()`
for the timeout. This seemed to cause occasional CI failures, such as
https://gitlab.gnome.org/GNOME/glib/-/jobs/1843454, where a resolver
error would be returned rather than the expected results:
```
ok 52 /network-address/happy-eyeballs/ipv4-error-ipv6-first
\# GLib-GIO-DEBUG: IPv4 DNS error: IPv4 Broken
(/var/tmp/gitlab_runner/builds/Ff4WDDRj/0/GNOME/glib/_build/gio/tests/network-address:18428): GLib-GIO-DEBUG: 09:03:08.587: IPv4 DNS error: IPv4 Broken
Bail out! GLib-GIO:ERROR:../gio/tests/network-address.c:586:got_addr: assertion failed (error == NULL): IPv4 Broken (g-io-error-quark, 24)
stderr:
**
GLib-GIO:ERROR:../gio/tests/network-address.c:586:got_addr: assertion failed (error == NULL): IPv4 Broken (g-io-error-quark, 24)
```
While I’ve been unable to reproduce these failures locally, I suspect
they might be down to thread spawning occasionally taking long enough on
a CI runner to change the ordering of the timeouts, such that the ‘Happy
Eyeballs’ algorithm returns a different set of results from what the
test expects.
So, this commit rewrites part of `mock-resolver.c` to implement timeouts
in the main thread, rather than in a worker thread. That should
eliminate the delays in spawning threads, and should mean that the
timeout sources in `mock-resolver.c` are attached to the same
`GMainContext` as those from the ‘Happy Eyeballs’ algorithm which are
monitoring them, so a total order over the timeouts can be guaranteed.
Of course, I might be completely wrong since this is just a guess and I
can’t properly test it since I can’t reproduce the failure. Worth a try.
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
The test results weren’t being freed.
This makes the `network-address` test clean under memcheck for me.
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
If these struct members aren't available, we can be more like an
abstraction layer by falling back to yielding 0.
Signed-off-by: Simon McVittie <smcv@collabora.com>
Test failures were previously ignored on macOS because there are 12
tests which consistently fail (and have not yet been fixed, because
there are no regularly active macOS maintainers for GLib; you could help
here!).
However, this means that new test failures can’t be spotted.
So, explicitly mark those 12 tests as `should_fail` on macOS, and then
make other test failures cause failure of the CI run.
We can track the process of fixing those 12 tests on #1392 and #1251.
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
Helps: #1392
As they use `g_test_dbus_*()`, they depend on dbus-daemon, so move them
to the part of the Meson file which lists those tests.
This disables them running on platforms which don’t have `dbus-daemon`
available. Arguably, this should be done by returning an error from
`g_test_dbus_up()` and then calling `g_test_skip()`, so the test is
correctly recorded as having been skipped. But that’s a fix for another
time.
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
Having compared the debug output, it doesn’t show anything unusual
happening that can’t already be seen from other output, for this test.
This is a partial revert of 8fd71dccc5. The debugging output it added to
other tests may still be useful.
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
Helps: #1929
The retry loop for acquiring the lock for the authentication cookie file
currently tries to acquire the lock for 0.5s, then gives up, assumes the
lock file is stale, and deletes it.
That’s great if the lock file *is* stale because it’s been left there by
a crashed process.
It’s not so great if the lock file just happens to have been there every
time this process checked, because the cookie file is highly contested
while (for example) running lots of parallel unit tests.
Check for that situation by comparing the mtime of the lock file and
continuing to retry if it’s changed.
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
Fixes: #1929
Rather than tracking them with a counter. This should close the race in
tracking the finalisation of the tasks by the task worker thread.
There’s no way to synchronise with that thread as it’s internal to
`g_task_run_in_thread()`.
This should hopefully stop the `debugcontroller` test being flaky.
See https://gitlab.gnome.org/GNOME/glib/-/merge_requests/2486#note_1384102
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
We are matching `<parameter>` as well as `<para>`, and we
end up with broken XML in case the (expanded) description
starts with `<parameter>`.
Fixes: #2601
This allows the controller to explicitly be removed from the bus, in a
way that allows the caller to synchronise with it and know that all
other references to the controller should have been dropped (i.e. after
this method returns, there should be no in-flight D-Bus calls still
holding a reference to the object).
This is needed to be able to guarantee finalisation of the controller in
unit tests (and comparable real-world situations).
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
Helps: #1190
The resources data is generated for both GCC and MSVC toolchains, even
though we know beforehand which toolchain we're going to compile it for.
By dropping the data duplication we make the generated resources file
faster to compile, especially when dealing with large embedded data,
instead of relying on the C pre-processor to walk the whole file and
discard the branch we're not using.
There shouldn’t be any issues here with empty argv arrays since an empty
`Exec=` line is already checked for. Encode that explicitly with an
assertion.
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
This won’t really affect anything, but we might as well fix them to not
crash if called with an empty `argv` by someone (ab)using `execve()`.
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
While it’s not usual, it is possible for applications to be called with
an empty `argv` by a process (ab)using `execve()`. Modify the
`GApplication` code example to handle that possibility, so that anyone
copying the example isn’t introducing a bug into their program
immediately.
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
If `GDebugControllerDBus` remains as the only, or default,
implementation of `GDebugController`, `dup_default()` cannot work.
`GDebugControllerDBus` requires a `GDBusConnection` at construction
time, which the `GIOModule` construction code can’t provide it.
Either we use a default D-Bus connection (but which one? and how would
it be changed by the user later if it was the wrong one?), or delegate
singleton handling of the `GDebugController` to the user.
The latter approach seems more flexible.
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
Helps: #1190
On Windows, dbus is launched by spawning the gdbus.exe executable on
demand from the gio module. When linked dynamically, the executable
path is guessed relatively to the gio DLL path. But when linked
statically, the only reference path available is the current
executable path. In this case, gdbus.exe is not necessarily in the same
folder as the current executable.
This patch solves the issue using the same algorithm as the one used
with process spawning in glib core source code two commits above.
This is intended to provide a uniform interface for controlling whether
the debug output from an application (or service) is emitted, typically
to journald, but actually to wherever the application chooses to output
it.
The main implementation of `GDebugController` is `GDebugControllerDBus`,
which is intended to be used on Linux. Other implementations may be
added in future for other platforms, or larger applications may want to
provide their own implementation which integrates with their ecosystem.
The `GDebugControllerDBus` implementation exposes a D-Bus interface at
`/org/gtk/Debugging` with a method to enable or disable debug
output at runtime.
This could be used by external harnesses, such as GNOME Builder or
systemd, to give a uniform way to get debug output from an application.
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
Fixes: #1190
As with the previous commit, this isn’t needed for GLib’s tests to work
correctly, but is probably needed in other projects which might be
tempted to copy and paste the Meson tooling from GLib.
Inspired by https://gitlab.gnome.org/GNOME/gtk/-/merge_requests/4330
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
This isn’t needed to make the tests any better (it doesn’t really affect
them), but is probably needed for anyone who copies this Meson code in
order to add `glib-compile-resources` support to their project. It’s
pretty unlikely that someone would want to compile *and export* a
resource from a shared library.
Inspired by https://gitlab.gnome.org/GNOME/gtk/-/merge_requests/4334
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
Credentials are often used to check peer processes details.
With AF_UNIX sockets on Windows, SIO_AF_UNIX_GETPEERPID can
be used to retrive the peer PID.
We will probably introduce more advanced mechanisms later on, though,
but I am not a Windows API expert.
Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
On !UNIX, return an error for send_fd() & receive_fd().
(the unixfdmessage unit is not compiled on !UNIX)
The header is installed under the common GIO include directory.
Ensure G_TYPE_UNIX_CONNECTION is registered on all platforms.
Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
The header is now also installed under the common GIO include directory.
Sorry if it breaks any build, you had to use the correct header path.
Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Move the header under the common GIO include directory.
Sorry if it breaks any build, you had to use the correct header path.
Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
They are not allowed in the specification, and can lead to infinite
loops when parsing.
That’s a security issue if your application is accepting D-Bus messages
from untrusted peers (perhaps in a peer-to-peer connection). It’s not
exploitable when your application is connected to a bus (such as the
system or session buses), as the bus daemons (dbus-daemon or
dbus-broker) filter out such broken messages and don’t forward them.
Arrays of zero-length elements are disallowed in the D-Bus
specification: https://dbus.freedesktop.org/doc/dbus-specification.html#container-types
oss-fuzz#41428, #41435Fixes: #2557
Glib cannot be built statically on Windows because glib, gobject and gio
modules need to perform specific initialization when DLL are loaded and
cleanup when unloaded. Those initializations and cleanups are performed
using the DllMain function which is not called with static builds.
Issue is known for a while and solutions were already proposed but never
merged (see: https://gitlab.gnome.org/GNOME/glib/-/issues/692). Last
patch is from version 2.36.x and since then the
"constructor/destructor" mechanism has been implemented and used in
other part of the system.
This patch takes back the old idea and updates it to the last version of
glib to allow static compilation on Windows.
WARNING: because DllMain doesn't exist anymore in static compilation
mode, there is no easy way of knowing when a Windows thread finishes.
This patch implements a workaround for glib threads created by calling
g_thread_new(), so all glib threads created through glib API will behave
exactly the same way in static and dynamic compilation modes.
Unfortunately, Windows threads created by using CreateThread() or
_beginthread/ex() will not work with glib TLS functions. If users need
absolutely to use a thread NOT created with glib API under Windows and
in static compilation mode, they should not use glib functions within
their thread or they may encounter memory leaks when the thread finishes.
This should not be an issue as users should use exclusively the glib API
to manipulate threads in order to be cross-platform compatible and this
would be very unlikely and cumbersome that they may mix up Windows native
threads API with glib one.
Closes#692
The gdbus-codegen tool generates documentation from the XML introspection
description of a D-Bus interface. Currently, only DocBook is supported at
the moment, but not every modern documentation generator can handle that
format. The reStructuredText format is a bit more well-supported,
especially in documentation generators for non-C languages.
Unlike DocBook, we get to make our own structure and conventions for how
we structure the documentation when using reStructuredText.
When parsing a comment we're adding <para> elements ourselves, but the
DocBook generator already wraps any block of text that does not start
with a <para> element with one.
xucred does not provide the peer pid id, but this can be fetched
from the socket LOCAL_PEERPID option. Note that we only support
it when creating the credentials from a local socket, if
the credential comes from a message over a socket the peer
pid id will not be set and -1 will be returned when trying
to get the pid for the credential.
D-Bus reference implementation doesn't require more than the claimed
process SID as part of the AUTH initial data for EXTERNAL.
Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
The code is based from #1351 patch and existing similar code in
gdbusprivate.c. The next commit will replace that existing code with
those helpers.
Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
This commit puts an additional underscore before the external symbol
`_g_binary_test1_resource_data` when using gcc for Windows, to match
the compiler's expectation.
Fixes#2571
Signed-off-by: Ralf Habacker <ralf.habacker@freenet.de>
Using ld_flags would work, but that does not propagate ldflags to users
of glib. Meson's dependency() call will propagate apple framework
dependencies to downstream users.
Not everything should be an int. This code is quite dated. We now try to
use `guint8*` to represent arbitrary binary data, rather than `guchar*`
(which makes it sound like some form of string).
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
While it’s not a supported public part of the API, exporting this
symbol from the library means we can easily test the DNS record parser,
which is important to do since it handles untrusted data from the
network.
This tests for #2503. It's fragile, but there is no non-fragile way to
test this. If the test breaks in the future, it will pass without
successfully testing the bug, not fail spuriously, so I think this is
OK.
We should run test_pass_fd twice, once using gspawn's fork/exec codepath
and once attempting to use its posix_spawn() codepath. There's no
guarantee we'll actually get the posix_spawn() codepath, but it works
for now on Linux.
For good measure, run it a third time with no flags at all.
This causes the test to fail if I separately break the fd remapping
implementation. Without this, we fail to test fd remapping on the
posix_spawn() codepath.
Specs say that on Unix id should be desktop file id from the xdg menu
specification, however, currently code just uses basename of .desktop file.
Fix that by finding the .desktop file in all the desktop_file_dirs and use
basename only as a fallback.
See https://specifications.freedesktop.org/menu-spec/latest/go01.html#term-desktop-file-id
and https://specifications.freedesktop.org/desktop-entry-spec/latest/ar01s02.html#desktop-file-id
"To determine the ID of a desktop file, make its full path relative to the
$XDG_DATA_DIRS component in which the desktop file is installed, remove the
"applications/" prefix, and turn '/' into '-'."
Also, add unit test that verifies Desktop Id is being correctly set
Signed-off-by: Ivaylo Dimitrov <ivo.g.dimitrov.75@gmail.com>
This is sad, but GResolver has one member variable for historical
reasons, to implement the reload signal. Since it offers a global
singleton, we should make sure it's really safe to use from multiple
threads at once.
This call was needed once upon a time, when GResolver had subclasses
that presumably relied on this. Nowadays, we have only
GThreadedResolver, which does not need it. res_init() is dangerous
because it modifies global state, so let's get rid of it.
meson in git master now warns about a missing `check:` kwarg, and may
eventually change the default from false to true.
Take the opportunity to require `objcopy --help` to succeed -- it is
unlikely to fail, but if it does something insane happened.
We used to use a pipe for the dbus daemon stdout to read the defined
address, but that was already requiring a workaround to ensure that dbus
daemon children were then able to write to stdout.
However the current implementation is still causing troubles in some
cases in which the daemon is very verbose, leading to hangs when writing
to stdout.
As per this, just don't handle stdout ourself, but use instead a
specific pipe to get the address address. That can now be safely closed
once we've received the data we need.
This reverts commit d80adeaa96.
Fixes: #2537
The code in `g_dbus_message_new_from_blob()` has now been fixed to
correctly error out on all truncated messages, so there’s no need for an
arbitrary programmer error if the input is too short to contain a valid
D-Bus message header.
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
Helps: #2528
Perform strict bounds checking when reading data from the D-Bus message,
and propagate errors to the callers.
Previously, truncated D-Bus messages could cause out-of-bounds reads.
This is a security issue, but one which is only exploitable when
communicating with an untrusted peer (who might send malicious
messages). Almost all D-Bus traffic is with a session or system bus,
where the dbus-daemon or dbus-broker is trusted, and is known to have
already rejected malformed (malicious) messages.
Accordingly, this is only exploitable with peer-to-peer D-Bus
conversations with an untrusted peer.
(Includes some minor cleanups from Philip Withnall.)
oss-fuzz#17408
Fixes: #2528
Since
https://gitlab.gnome.org/GNOME/glib-networking/-/merge_requests/173,
there is now a really surprising implication to using a non-default
GTlsDatabase: your database could do nothing at all other than wrap the
default database, which you would expect to result in no behavior
changes, but in fact it causes fewer security checks to be performed
during certificate verification. This is because certificate
verification moved from GTlsDatabase to GTlsConnection, allowing for
more security checks to be performed. But if using a non-default
GTlsDatabase, we have to fall back to letting GTlsDatabase to the
verification, as before.
This is the best we can do. It's not a regression for applications,
because it means applications get the previous pre-2.72 behavior. But it
does mean that new security checks added in 2.72 are not applied, which
is unfortunate, so we should warn developers about this.
This feature has been reverted for now because I messed up the
implementation and it was doing sync I/O during async API calls. Oops!
Since it's not present in 2.70 nor in 2.72, let's remove the reference
to the exact GLib version that this behavior was introduced in. I'd like
to get it working properly for 2.74, but it's not ready yet and just
changing the version to 2.74 feels optimistic.
Rather than waiting for a fixed period of time, poll in a loop until the
condition the test is expecting is true.
A better solution would be to use a `GSource` and wait until that’s
dispatched. But doing so might affect the behaviour of the
`GInputStream` under test, so busy-wait instead.
Fixes this CI failure: https://gitlab.gnome.org/GNOME/glib/-/jobs/1630758
```
(some socket debug output)
Bail out! GLib-GIO:ERROR:../gio/tests/converter-stream.c:1037:test_converter_pollable: assertion failed (res == -1): (1 == -1)
```
I could not reproduce the failure remotely with a few hundred
invocations of the test, so it might only present itself on BSD, which
presumably has different socket timing behaviour from Linux.
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
gio/tests/gio-du.c: In function 'main':
gio/tests/gio-du.c:74:11: error: parameter 'argc' set but not used
74 | main (int argc, char **argv)
| ~~~~^~~~
gio/ginputstream.c: In function 'g_input_stream_real_skip':
gio/ginputstream.c:433:31: error: comparison of integer expressions of different signedness: 'goffset' {aka 'long long int'} and 'long long unsigned int'
433 | (start + count) > (guint64) end)
| ^
gio/gwin32appinfo.c: In function 'uwp_package_cb':
gio/gwin32appinfo.c:3383:17: warning: comparison of integer expressions of different signedness: 'gint' {aka 'int'} and 'guint' {aka 'unsigned int'}
for (i = 0; i < supported_extgroups->len; i++)
^
gio/gwin32appinfo.c:3389:29: warning: comparison of integer expressions of different signedness: 'gint' {aka 'int'} and 'guint' {aka 'unsigned int'}
for (i_ext = 0; i_ext < grp->extensions->len; i_ext++)
^
gio/gwin32appinfo.c:3430:35: warning: comparison of integer expressions of different signedness: 'gint' {aka 'int'} and 'guint' {aka 'unsigned int'}
for (i_verb = 0; i_verb < grp->verbs->len; i_verb++)
^
gio/gwin32appinfo.c:3463:33: warning: comparison of integer expressions of different signedness: 'gint' {aka 'int'} and 'guint' {aka 'unsigned int'}
for (i_hverb = 0; i_hverb < ext->verbs->len; i_hverb++)
^
gio/gwin32appinfo.c:3478:17: warning: comparison of integer expressions of different signedness: 'gint' {aka 'int'} and 'guint' {aka 'unsigned int'}
for (i = 0; i < supported_protocols->len; i++)
^
gio/gwin32appinfo.c:3541:33: warning: comparison of integer expressions of different signedness: 'gint' {aka 'int'} and 'guint' {aka 'unsigned int'}
for (i_hverb = 0; i_hverb < url->verbs->len; i_hverb++)
^
gio/gwin32appinfo.c: In function 'g_win32_app_info_launch_internal':
gio/gwin32appinfo.c:4799:37: warning: comparison of integer expressions of different signedness: 'gint' {aka 'int'} and 'guint' {aka 'unsigned int'}
for (p_index = 0; p_index <= g_strv_length (envp); p_index++)
^~
gio/gwin32packageparser.c: In function 'WIN32_FROM_HRESULT':
gio/gwin32packageparser.c:99:30: warning: comparison of integer expressions of different signedness: 'long unsigned int' and 'long int'
if ((hresult & 0xFFFF0000) == MAKE_HRESULT (SEVERITY_ERROR, FACILITY_WIN32, 0) ||
^~
This was previously done (by commit 63038d1e4c) in one of the cases
where `kill_test_service()` was called — but not the other.
This meant that one instance of `gdbus-testserver` could still be
around when (as it happens, due to the order of the tests) the
`/gdbus/proxy/no-match-rule` test was run. It would start a second
instance of `gdbus-testserver`, which would exit early due to the test
name still being owned on the bus. The first (killed) instance of
`gdbus-testserver` would then exit, leaving no test servers running, and
hence the new test would fail.
This was being seen as frequent CI failures, particularly on FreeBSD
(must have slightly different timing for process signalling and
termination from Linux).
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
Once upon a time, we tried to return all possible certificate errors,
but it never actually worked reliably and nowadays we have given up.
This needs to be documented because a reasonable developer would not
expect it.
Because mistakes could be security-critical, I decided to copy the same
warning in several different places rather than relying only on
cross-referencese.
These are known leaks, as they were being done in tests which were
checking precondition failures.
However, since we know what happens when the failures occur, we can
still free the input data reliably, so do that.
This improves the valgrind output for `actions` to show zero definite
leaks.
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
The `actions` test previously waited an arbitrary 100ms for various
D-Bus messages to be sent/received, before checking the results of those
messages.
Normally, this would work, but on heavily loaded CI systems, it would
sometimes fail. For example,
https://gitlab.gnome.org/GNOME/glib/-/jobs/1611701.
Fix that by waiting for the condition being checked to evaluate to true,
rather than waiting an arbitrary period of time. On faster machines,
this will speed the tests up too.
Assume that the global default `GMainContext` is in use, so a
`GMainContext*` pointer doesn’t have to be passed around.
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
gio/gcontenttype-win32.c: In function 'get_registry_classes_key':
gio/gcontenttype-win32.c:66:78: warning: comparison of integer expressions of different signedness: 'DWORD' {aka 'long unsigned int'} and 'int'
if (ExpandEnvironmentStringsW (wc_temp, wc_temp_expanded, len) == len)
^~
This allows the flag to allow interactive auth to be set. Previously, it
was unconditionally unset.
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
D-Bus has an upper limit on number of Match rules and it's rather easy to hit
with a big number of proxies with signal subscriptions. This happens with
NetworkManager with hundreds of devices or connection settings. By passing
G_DBUS_SIGNAL_FLAGS_NO_MATCH_RULE to g_dbus_connection_signal_subscribe(), the
user can call AddMatch with a less granular match instead of a match per every
proxy.
Tests subsequently added by Philip Withnall.
Fixes: #1109
- Isolate the first meaningful paragraph, for gi-docgen's summary
- Describe get_object() as a binding API
- Fix reference to get_item() inside get_item_type()
g_check_setuid does more than setuid checks when using AT_SECURE.
Make sure that it is referenced in the error message to help
users debug in case or errors
Closes#2518
Fix another variant of the previous commit, this time specific to the
idle callback of a method call on a subtree object, racing with
unregistration of that subtree.
In this case, the `process_subtree_vtable_message_in_idle_cb()` idle
callback already has a pointer to the right `ExportedSubtree` struct,
but again doesn’t have a strong reference to it.
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
Helps: #2400
If `g_dbus_connection_unregister_object()` (or `unregister_subtree()`)
was called from one thread, while an idle callback for a method call (or
a property get or set) was being invoked in another, it was possible for
the two to race after the idle callback had checked that the
object/subtree was registered, but before it had finished dereferencing
all the data related to that object/subtree.
Unregistering the object/subtree would immediately free the data,
leading the idle callback to cause a use-after-free error.
Fix that by giving the idle callback a strong reference to the data from
inside the locked section where it checks whether the object/subtree is
still registered.
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
Fixes: #2400
This is needed for an upcoming change which decouples their lifecycle
from their presence in the `map_id_to_ei` and `map_id_to_es` hash
tables.
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
Helps: #2400
Move it further up the file, but make no changes to it. This will help
with a subsequent commit.
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
Helps: #2400
This introduces no functional changes; it only simplifies the code.
Instead of maintaining a separate pointer to the backend iff it’s a
`GDelayedSettingsBackend`, just test the `backend` pointer’s type.
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
Helps: #2426
Previously, the delay-apply status of the parent `GSettings` object
would be partially inherited: `settings->priv->backend` in the child
`GSettings` object would point to a `GDelayedSettingsBackend`, but
`settings->priv->delayed` would be `NULL`.
The expectation from https://bugzilla.gnome.org/show_bug.cgi?id=720891
was that `get_child()` would fully inherit delay-apply status.
So, ensure that `settings->priv->delayed` is correctly set to point to
the delayed backend when constructing any `GSettings`. Update the tests
to work again (presumably the inverted test was an oversight in the
original changes).
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
Fixes: #2426
`g_settings_reset()` changes the value of the setting to `NULL`;
`add_to_tree()` was not handling that correctly.
Add a unit test.
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
Fixes: #2426
Add missing (nullable), and use 2-space indentation to avoid markdown
pre-formatted blocks when unwanted.
Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
As this would have undesirable consequence.
Quoting Philip Withnall:
https://gitlab.gnome.org/GNOME/glib/-/merge_requests/2305#note_1294729:
The documentation never said anything about accepting absolute paths,
so any code which is relying on that is relying on undocumented
behaviour. We’re allowed to change that.
Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
`standard::name` must be available for `g_file_enumerator_get_child()`
to work. Emit a critical warning and return if it’s not. This is similar
to the existing behaviour in `g_file_enumerator_iterate()`.
Improve the documentation to mention this.
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
Fixes: #2507
The code cannot function correctly if the `standard::name` attribute is
not present, so upgrade the existing warning to a critical warning and
return if it fails in `g_file_enumerator_iterate()`.
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
Helps: #2507
It was Red Hat specific when it was introduced in 2004, was never
supported by mount(8) upstream, and was removed entirely in 2008.
It’s confusing for GLib to keep references to it around.
Thanks to Karel Zak for digging up the history of it:
https://gitlab.gnome.org/GNOME/glib/-/merge_requests/2298#note_1294519
And thanks to Xidorn Quan for looking into it in the first place (see
!2298).
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
gio/gsocket.c: In function 'g_socket_get_available_bytes':
gio/gsocket.c:3141:17: warning: comparison of integer expressions of different signedness: 'u_long' {aka 'long unsigned int'} and 'int'
if (avail == -1)
^~
gio/gsocket.c: In function 'g_socket_send_messages_with_timeout':
gio/gsocket.c:5283:19: warning: comparison of integer expressions of different signedness: 'gint' {aka 'int'} and 'guint' {aka 'unsigned int'}
for (i = 0; i < num_messages; ++i)
^
gio/gsocket.c:5308:76: warning: operand of ?: changes signedness from 'int' to 'gsize' {aka 'long long unsigned int'} due to unsignedness of other operand
result = pollable_result == G_POLLABLE_RETURN_OK ? bytes_written : -1;
^~
gio/win32/gwinhttpfile.c: In function 'g_winhttp_file_query_info':
gio/win32/gwinhttpfile.c:554:13: warning: comparison of integer expressions of different signedness: 'int' and 'size_t' {aka 'long long unsigned int'}
n == wcslen (content_length))
^~
gio/win32/gwin32fsmonitorutils.c: In function 'g_win32_fs_monitor_handle_event':
gio/win32/gwin32fsmonitorutils.c:107:11: warning: comparison of integer expressions of different signedness: 'GFileMonitorEvent' {aka 'enum <anonymous>'} and 'int'
if (fme != -1)
^~
gio/win32/gwinhttpvfs.c: In function 'g_winhttp_vfs_get_file_for_uri':
gio/win32/gwinhttpvfs.c:172:17: warning: comparison of integer expressions of different signedness: 'int' and 'long long unsigned int'
for (i = 0; i < G_N_ELEMENTS (winhttp_uri_schemes); i++)
^
gio/win32/gwinhttpvfs.c: In function 'g_winhttp_vfs_get_supported_uri_schemes':
gio/win32/gwinhttpvfs.c:210:17: warning: comparison of integer expressions of different signedness: 'int' and 'long long unsigned int'
for (i = 0; i < G_N_ELEMENTS (winhttp_uri_schemes); i++)
^
The "recursive:" kwarg is available in the targeted minimum version of
meson, and is basically required if you want to not emit warnings and
maybe error with --fatal-meson-warnings.
There are two basic solutions to this problem:
- The current default behavior is false, so explicitly opt in to that
value. None of these internal libraries use recursive objects anyway.
- Use link_with to link to the static library directly, rather than the
extracted objects.
Option 2 is what used to be done before commit
62af03bda8, but it only works with meson
>=0.52 and previously had buggy behavior.
Since the minimum version of meson is now 0.52, it is safe to revert
that commit and go back to using link_with, and therefore option 2 is
chosen.
Previously, the code validated that child objects have a path with the
object manager strictly as a prefix. That doesn’t work in the case where
the object manager’s path is `/`. This case is not recommended, but is
supported.
If the object manager’s path is `/`, validate that child objects’ paths
are not equal to it. If they are equal to it, warn the user rather than
emitting a critical warning, since we can’t expect any users who’ve not
been compliant with the spec to instantly rework their D-Bus APIs.
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
Fixes: #2500
Emit this when we're about to spawn or DBus activate a GAppInfo. This
allows lauchers to keep the appinfo associated with a startup id.
We use a GVariant to allow for future exansion of the supplied data.
When using g_desktop_app_info_launch_uris_as_manager the "launched"
signal allows to map a desktop-startup-id to a GAppInfo. Make this
possible for DBus activation too.
Since we don't have a PID there we pass a 0. Update the signal
description accordingly.
These should be implemented by loadable IO module libraries, but are not
callable in GLib itself.
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
Fixes: #2498
This is a partial revert of commit
f378352051, as the previous commits have
silenced the AddressSanitizer warnings for `GContentType`.
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
Fixes: #2310
The function pointer casts silence the compiler and allow the code to
build (and even run in the typical case). However, when building with
control flow integrity checks, the runtime (rightfully) complains about
calling a function via a mismatched function pointer type.
In order to make xdgmime properly relocatable so that unit tests can use
it without it reading and modifying the user’s actual xdgmime files, and
without the need to call setenv() (and get tied up with thread safety
problems), add a xdg_mime_set_dirs() method to allow the dirs to be
overridden. They will still default to the values of $XDG_DATA_HOME and
$XDG_DATA_DIRS.
Signed-off-by: Philip Withnall <withnall@endlessm.com>
Allocate an empty cache object, check cache objects for being empty
before using them.
Otherwise the code will re-read cache every 5 seconds, as NULL cache
does not trigger the code that stores mtime, which makes the cache
file appear modified/unloaded permanently.
https://bugzilla.gnome.org/show_bug.cgi?id=735696
Since returning exactly one match has special significance, don't
give up matching before we've found at least 2 types. Also, make
sure that we don't return the same mime type more than once.
Bug 541236.
If an `InterfacesRemoved` signal is received for an object which doesn’t
exist in the local map of interfaces, don’t emit a warning.
This seems to happen in the real world (see #2401). Without a trace of
the D-Bus traffic it’s not possible to know exactly what situation is
causing this, but it seems possible that the peer could disappear and
its `notify::name-owner` signal could be processed before its
`InterfacesRemoved` signal, or something similar.
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
Fixes: #2401
Relax assertion about opened registry key as it may have been removed
in the meantime between enumeration and when opening, or (more likely)
we may not have the required permissions to open the some enumerated
keys (i.e. RegOpenKeyExW fails and returns ERROR_ACCESS_DENIED).
Fixes https://gitlab.com/inkscape/inbox/-/issues/5669
Instead of calling xterm when it clearly does not exist and causes a silent error,
inform the user that the launch failed so they can take the right action.
Also added (transfer full) or (transfer none) annotations while I was at it.
This is the result of checking each `Returns:` line in these files. I’ve
only considered nullability and transferability, and not other (potentially
missing or incorrect) annotations.
Helps: #2227
If the first power-profile installed test fails (for example, because
xdg-desktop-portal isn’t available), correctly tear down the dbusmock
object, or it will cause setUp() to fail when the next test in the suite
is run.
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
Helps: #2481
When first creating the monitor, correctly set its property value to the
value from the portal, rather than waiting for the portal value to
change to set it.
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
Fixes: #2481
We were lucky that this worked in some cases (the test is racy), but we
should actually run the condition check each loop, rather than when the
function is called.
Spotted by Martin Pitt:
96a8c02d24 (r54773831)
Let's explain the advantages of relying on GTlsConnection to perform
certificate verification.
Also, document that the issuer property is a little tricky, because the
issuer certificate might not be the certificate that actually gets used
in final certification path building. This is very unexpected to anybody
who is not an expert.
Because TLS certificate verification is extremely complex, the lookup
issuer function may be tempting to misuse even by experienced
developers. There is a notion that the issuer certificate will always be
used in the final certification path, but it's just not always true.
Trying to make security decisions based on the results of this function is
a trap, so let's document that.
It turns out that old versions of glib-networking actually reordered the
certificate chain to match the final verification path. This no longer
happens since a long time ago, because it was a buggy mess. Instead, we
rely on the TLS library to build the final verification path. Their path
building is not very good, but at least it's consistent. The point of
these doc updates is to clarify that only the TLS library can make
security decisions.
Document that HTTP requests may be performed to look up missing
certificates.
Finally, let's document that certificate verification using GTlsDatabase
cannot be as smart as certificate verification performed directly by
GTlsConnection.
g_win32_package_parser_enum_packages() reads beyond the end of a buffer
when doing a memcpy. With app verifier enabled on Windows, it causes
the application to crash on startup.
This change limits the memcpy to the size of the source string.
Fixes: #2454
This reverts commit 7aa0580cc5.
As stated in #2316, that commit was a workaround to allow gnome-keyring
and msmtp to continue to get their session bus address from
`DBUS_SESSION_BUS_ADDRESS`, even though they’re `AT_SECURE`. The timeout
on that workaround has expired so that commit is now being reverted.
Fixes: #2316
You need to separate the first entry in the list from the preceding
paragraph, and you should add a space before the enumerating symbol.
GTK-Doc accepts a very lax Markdown syntax, but any other tool parsing
our documentation will likely fail.
The value should be initialized to NULL before calling
g_win32_registry_key_get_value_w(), to ensure that cleanup
can be done unconditionally afterward.
To ensure that the watch is properly re-set every time, call
watch_keys() from the watch callback. Previously the watch was only
renewed after a data update was done in a worker thread, which made
no sense, since the update function was implemented in such a way
that it can (and should) be re-triggered on each key change, until
the changes stop coming, and that can only happen if we renew
the registry watcher right away.
If a key watch is renewed from the key watch callback, it results
in the callback being NULL, since we clear it after we call it.
Rearrange the function to make sure that the changes done by the
callback function are preserved properly.
This function can, in fact, return STATUS_SUCCESS. We shouldn't
assert that it doesn't.
For now interpret it just like STATUS_PENDING (i.e. APC will be called),
see how it goes (it isn't documented how the function behaves in this
case, we have to play it by ear).
Note that while we *can* use a better-documented RegNotifyChangeKeyValue() here,
it communicates back to us via event objects, which means that the registry
watcher would have to interact with the main loop directly and insert its
events (plural; one event per key) there. That would make the API more complicated.
Whereas the internal NT function communicates by calling an APC - we're good
as long as something somewhere puts the thread in alertable state.
When attempting to test Windows support for building libadwaita, since we are
using multiple GResource files, one would hit linker errors where multiple
definitions of the following symbols have been defined, when
glib-compile-resources was invoked without manual register:
resource_constructor_wrapper
resource_destructor_constructor
_arrayresource_constructor
_arrayresource_destructor
In order to avoid that, just prefix the definitions of resource_constructor
and resource_destructor, like what we do when --manual-register is used, with
what we pass in with --c-name so that we ensure that we do not end up in such
name collisions.
Port all existing calls in GLib to the new API so that they can receive
more detailed error information (although none of them actually make use
of it at the moment).
This also serves to test the new API better through use.
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
Helps: #203
The `g_file_trash` function fails with the `Unable to find or create trash
directory` error when the global `.Trash` directory exists. This is because
the commit 7f2af262 introduced the `gboolean success` variable to signalize
the detection of the trash folder, but didn't set it in all code branches.
Since for a time this variable was not initialized the bug wasn't visible
when the trash folder existed. The bug became effective after the `success`
variable was initialized with `FALSE` by the commit c983ded0. Let's explicitly
set the `success` variable in all branches to fix the global trash dir
detection.
Fixes: https://gitlab.gnome.org/GNOME/glib/-/issues/2439
The function which calls `SetFileTime()` works with seconds and
nanosecond, but the functions which call it are doing so with seconds
and microseconds.
Fix them so they convert to nanoseconds first.
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
The code appears to be dealing with time in units of 100ns, not 100µs,
so name the variable accordingly.
The rest of the arithmetic in that function appears consistent and
correct.
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
The modification time test creates a file, gets the modification time in
seconds, then gets the modification time in microseconds and assumes
that the difference between the two has to be above 0.
As rare as this may be, it can happen:
$ stat g-file-info-test-50A450 -c %y
2021-07-06 18:24:56.000000767 +0100
Change the test to simply assert that the difference not negative to
handle this case.
This is necessary when building glib with icecc. Icecc splits the build
process into two parts. The file is locally preprocessed with
-fdirectives-only to resolve any includes. This adds linemarkers to the
intermediate file. Without the new-line at the end of the file this:
#include "gconstructor_as_data.h"
#include "glib/glib-private.h"
Is turned into this:
const char gconstructor_code[] = "...";# 1 "glib/glib-private.h"
...
The result is a compile error:
In file included from ../glib/gio/glib-compile-resources.c:45:
gio/gconstructor_as_data.h:1: error: stray '#' in program
gio/gconstructor_as_data.h:1: error: expected identifier or '(' before numeric constant
In file included from ../glib/glib/glib-private.h:22,
from gio/gconstructor_as_data.h:2,
from ../glib/gio/glib-compile-resources.c:45:
../glib/glib/gwakeup.h:27:1: error: unknown type name 'GWakeup'
../glib/glib/gwakeup.h:28:42: error: unknown type name 'GWakeup'
../glib/glib/gwakeup.h:30:42: error: unknown type name 'GWakeup'
../glib/glib/gwakeup.h:32:42: error: unknown type name 'GWakeup'
../glib/glib/gwakeup.h:33:42: error: unknown type name 'GWakeup'
In file included from gio/gconstructor_as_data.h:2,
from ../glib/gio/glib-compile-resources.c:45:
../glib/glib/glib-private.h:98:3: error: unknown type name 'GWakeup'
../glib/glib/glib-private.h:99:58: error: unknown type name 'GWakeup'
../glib/glib/glib-private.h💯58: error: unknown type name 'GWakeup'
../glib/glib/glib-private.h:102:58: error: unknown type name 'GWakeup'
../glib/glib/glib-private.h:103:58: error: unknown type name 'GWakeup'
In file included from gio/gconstructor_as_data.h:2,
from ../glib/gio/glib-compile-resources.c:45:
../glib/glib/glib-private.h:164:53: warning: file "../glib/gio/glib-compile-resources.c" linemarker ignored due to incorrect nesting
To avoid this, generate gconstructor_as_data.h with a new-line at the end
of the file.
Signed-off-by: Michael Olbrich <m.olbrich@pengutronix.de>
This helps to void deadlocks when two processes call interfaces on each
other one of them being org.freedesktop.DBus.ObjectManager.
Signed-off-by: Guido Günther <agx@sigxcpu.org>
The current code is unsafe to use from multiple threads at once.
GIOStream functions like this are supposed to be semi-threadsafe. It's
allowed for them to be called on both a reader thread and a writer
thread at the same time. Of course, it's still tricky and dangerous,
because it's only *really* threadsafe if the handshake has finished,
and API users have no plausible way to know that because the API
does not require performing an explicit handshake operation. But that's
a glib-networking problem. We can at least avoid the most obvious
threadsafety issue here in the API layer.
Note that we'll need to implement the new vfunc in glib-networking for
this to actually work.
Fixes#2393
The documentation for `g_bus_watch_name()` implies that the
`GDestroyNotify` for the user data will be called in the current thread
default `GMainContext`. Currently, it could be called in any thread, as
`client_unref()` can be called in any thread.
Fix that by deferring it to an idle source if `client_unref()` finalises
the `Client` object in a different thread.
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
These were missing from the test before the previous commit ported from
`GMainLoop` to `GMainContext`.
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
It makes combination exit conditions a lot easier than when using
`g_main_loop_quit()` from different callbacks.
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
The first sentence incorrectly said that it checked the type of the
value, and then the second sentence explicitly said it was a programmer
error to give a value of the wrong type.
According to the code, the second sentence is correct.
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
Fixes: #2425
WebKit wants these private key properties to be readable in order to
implement a deserialization function. Currently they are read-only
because at the time GTlsCertificate was originally designed, the plan
was to support PKCS#11-backed private keys: private keys that are stored
on a smartcard, where the private key is completely unreadable. The
design goal was to support both memory-backed and smartcard-backed
private keys with the same GTlsCertificate API, abstracting away the
implementation differences such that code using GTlsCertificate doesn't
need to know the difference.
The original PKCS#11 implementation was never fully baked and at some
point in the past I deleted it all. It has since been replaced with a
new implementation, including a GTlsCertificate:private-key-pkcs11-uri
property, which is readable. So our current API already exposes the
differences between normal private keys and PKCS#11-backed private keys.
The point of making the private-key and private-key-pem properties
write-only was to avoid exposing this difference.
Do we have to make this API function readable? No, because WebKit could
be just as well served if we were to expose serialize and deserialize
functions instead. But WebKit needs to support serializing and
deserializing the non-private portion of GTlsCertificate with older
versions of GLib anyway, so we can do whatever is nicest for GLib. And I
think making this property readable is nicest, since the original design
reason for it to not be readable is now obsolete. The disadvantage to
this approach is that it's now possible for an application to read the
private-key or private-key-pem property, receive NULL, and think "this
certificate must not have a private key," which would be incorrect if
the private-key-pkcs11-uri property is set. That seems like a minor
risk, but it should be documented.
On Unix platforms, wait() and friends yield an integer that encodes
how the process exited. Confusingly, this is usually not the same as
the integer passed to exit() or returned from main(): conceptually it's
an integer encoding of this tagged union:
enum { EXITED, SIGNALLED, ... } tag;
union {
int exit_status; /* if EXITED */
struct {
int terminating_signal;
bool core_dumped;
} terminating_signal; /* if SIGNALLED */
...
} detail;
Meanwhile, on Windows, wait statuses and exit statuses are
interchangeable.
I find that it's clearer what is going on if we are consistent about
referring to the result of wait() as a "wait status", and the value
passed to exit() as an "exit status".
GSubprocess already gets this right: g_subprocess_get_status() returns
the wait status, while g_subprocess_get_exit_status() genuinely returns
the exit status. However, the GSpawn family of APIs has tended to
conflate the two.
Confusingly, g_spawn_check_exit_status() has always checked a wait
status, and it would not be correct to pass an exit status to it; so
let's deprecate it in favour of g_spawn_check_wait_status(), which
does the same thing that g_spawn_check_exit_status() always did.
Code that needs backwards-compatibility with older GLib can use:
#if !GLIB_CHECK_VERSION(2, 69, 0)
#define g_spawn_check_wait_status(x) (g_spawn_check_exit_status (x))
#endif
Signed-off-by: Simon McVittie <smcv@collabora.com>
Confusingly, g_spawn_check_exit_status() takes a wait status, not an
exit status, so passing g_subprocess_get_exit_status() to it is
incorrect (although both encodings happen to use 0 to encode success
and a nonzero value to encode failure, so in practice this probably
had the desired effect).
Signed-off-by: Simon McVittie <smcv@collabora.com>
Following on from the previous commit, some explicit
`g_main_context_wakeup()` calls were missing from the test code which
only uses `GMainContext`.
Add them, and also add some assertions to check that these functions are
being called in the expected thread (as the code comments say).
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
This is a bit of a compromise. Since the option parsing in
`GApplication` is built on `GOptionContext`, there’s no way to
reliably indicate that a given option was passed by the user, other than
by its value changing. If the default value is zero, but the user
explicitly passed zero, nothing changes, so it’s not obvious that the
option was explicitly provided.
When just `GOptionContext` is being used, this is fine, as that’s
obvious what will happen from the way the API is built. With
`GApplication::handle-local-options`, though, the `GVariantDict`
provided by GLib to the callback claims to only contain the values of
the options provided by the user, and no defaults.
It’s not actually possible for GLib to do that reliably.
Previously, GLib was dropping all numeric values which were zero valued
(i.e. the defaults), as they *could* have been the defaults. It seems
like a slightly better behaviour to instead *not* drop those numeric
values, and err on the side of reporting some defaults as user-provided
(even if they weren’t) rather than dropping some user-provided values
which happen to be the defaults.
This adds a test for the case of parsing a double; the cases for
integers are analogous.
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
Fixes: #2329
The tests in `gdbus-names.c` use a mixture of `GMainLoop` and iterating
a `GMainContext` directly. Some of the helper functions based around the
`OwnNameData` struct use the `loop` `GMainLoop` even when called from
tests like `watch_with_different_context()` which themselves use
`GMainContext` directly.
Thus, it’s possible for the `GMainLoop` to not be running, while the
test is iterating on `g_main_context_iteration()`. In this case,
`g_main_loop_quit()` is a no-op and will not wake up the `GMainContext`.
This causes the test to livelock in around 1 in 1200 test runs.
Fix this by adding an explicit `g_main_context_wakeup()` call after each
`g_main_loop_quit()` call. A more comprehensive fix would be to port all
the tests in this file to iterating `GMainContext` directly, and drop
all the `GMainLoop` usage, but I don’t have time for that right now.
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
If `g_file_monitor_source_dispatch()` drops the last reference to its
`GLocalFileMonitor`, a deadlock will occur, because disposing the
`GLocalFileMonitor` causes synchronous disposal of the
`GFileMonitorSource`, and hence an attempt to re-lock the already-locked
mutex in the `GFileMonitorSource`.
Fix that by dropping the reference to the `GLocalFileMonitor` after
unlocking.
Diagnosed by Ting-Wei Lan. The bug was originally introduced by me in
commit 592a13b483.
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
