`SOCKS4_CONN_MSG_LEN` failed to account for the length of the final nul
byte in the connect message, which is an addition in SOCKSv4a vs
SOCKSv4.
This means that the buffer for building and transmitting the connect
message could be overflowed if the username and hostname are both
`SOCKS4_MAX_LEN` (255) bytes long.
Proxy configurations are normally statically configured, so the username
is very unlikely to be near its maximum length, and hence this overflow
is unlikely to be triggered in practice.
(Commit message by Philip Withnall, diagnosis and fix by Michael
Catanzaro.)
Fixes: #3461
libinotify-kqueue is a library that implements inotify interface in terms of
kqueue/kevent API available on Mac OS and *BSD systems. The original kqueue
backend seems to be a predecessor version of the code that is currently present
in libinotify-kqueue. Under the hood the library implements a sophisticated
filesystem changes detection algorithm that is derived from the glib backend
code.
Updating the native glib kqueue backend requires substantial work, because code
bases have diverged greatly. Another approach is taken, instead. libinotify-kqueue
can serve as a drop-in replacement for Linux inotify API, thus allowing to
reuse the inotify backend code. The compatibility, however, comes at cost, since
the library has to emulate the inotify descriptor via an unix domain socket.
This means that delivering an event involves copying the data into the kernel
and then pulling it back.
The recent libinotify-kqueue release adds a new mode of operation called "direct".
In this mode the socket pipe is replaced with another kqueue that is used to
deliver events via a kevent(EVFILT_USER) call.
Employing the direct mode requires minor changes to the client code compared
to using plain inotify API, but in return it allows for reusing libinotify's
algorithms without a performance penalty. Luckily, all required changes are
consolidated in one file called inotify-kernel.c
This puts us in the best of possible worlds. On one hand we share a lot of code
with glib inotify backend, which is far more thoroughly tested and widely used.
On the other we support a range of non-Linux systems and consolidate the business
logic in one library. I plan to do the same trick for QFileSystemWatcher which
will give us the same behaviour between Gtk and Qt applications.
The glib test suite passes for both old kqueue backend and new libinotify-kqueue
one. However, the AppStream FileMonitor tests are failing with the old backend,
but pass with the new one, so this is still an observable improvement.
Relevant libinotify-kqueue PR: https://github.com/libinotify-kqueue/libinotify-kqueue/pull/19
This was intended to make it feasible to debug connection failures where
a modern GDBus client connects to a GDBusServer that is missing
glib!2826, but those GDBusServers should be increasingly rare: we have
had the fixed version of GDBusServer for 5 GNOME stable releases
(2 years), long enough to get into stable releases of Debian and Ubuntu
and a service-pack release of SUSE.
This debug message can be very noisy, especially when unit-testing
GLib code (which is frequently done with G_MESSAGES_DEBUG=all) or when
a third-party GLogFunc ignores the log_level parameter and logs all
messages as though they indicated a serious problem. I think it has
served its purpose now.
Signed-off-by: Simon McVittie <smcv@collabora.com>
Ideally, we would eventually move to always using the
cross-namespace-friendly authentication handshake and never sending an
initial response with the Unix uid or Windows SID, but that's blocked
on waiting for various LTS OS distributions to become unsupported or
otherwise irrelevant. Make a note of some major LTS distributions that
still have an old version that is missing glib!2826.
In the RHEL/CentOS family, CentOS Stream 10 has a fixed version, so
presumably RHEL 10 will have a fixed version when it is released.
In the Debian family, Debian 12 and Ubuntu 24.04 are stable releases
that contain a fixed version.
In the SUSE family, openSUSE 15.6 contains a fixed version, so
presumably so does SLES 15.6.
Faster-moving distributions like Fedora are not a concern here and
can be ignored: it's always the slow-moving and/or LTS distributions
that are going to be holding back interoperability breaks.
Signed-off-by: Simon McVittie <smcv@collabora.com>
The option defaults to 'auto' which keeps the current selection behavior, but
also allows user to override the choice.
Also make the chosen backend is reported to the code via CPP defines.
This introduces no functional changes, but does squash a false positive
warning from `-Wfloat-conversion`:
```
../gio/gapplication.c:462:15: error: implicit conversion turns floating-point number into integer: 'gdouble' (aka 'double') to '_Bool' [-Werror,-Wfloat-conversion]
if (*(gdouble *) entry->arg_data)
~~ ^~~~~~~~~~~~~~~~~~~~~~~~~~~~
1 error generated.
```
Signed-off-by: Philip Withnall <pwithnall@gnome.org>
Remove cmake as we no longer need to build ninja. We can use the
official wheel now since the runner's Python is 3.9 (before: 3.8).
Use the same comment regarding '--wrap-mode' as in the other jobs.
Download and use official ccache binary.
Add myself to the 'only' section in .gitlab-ci.yml so I can have
CI in my fork.
Disable a few deprecation warnings due to the much newer SDK of
the Apple Silicon machine.
Meson 1.5.1 is available in the fd.o SDK and in Debian testing, so the
glib Meson policy says we can update. Update the minimum only as far as
1.4.0 because we don't yet have a need for 1.5.0.
This allows us to:
- Use file.full_path() to avoid deprecation warnings on str.format(file).
- Set c_std=gnu99,c99 to avoid deprecation warnings with gnu99 on MSVC.
Update all the CI builds to use the latest 1.4.x patch release, 1.4.2.
The FreeBSD runner cannot be updated via `gitlab-ci.yml`, so will be
broken for now.
Similarly, the macOS build will not work unless `-Dc_std=gnu99` is
specified at configure time, due to
https://github.com/mesonbuild/meson/issues/13639.
The WSAEnumNetworkEvents API is called every time the socket
needs to be checked for status changes. Doing this in an application
with several sockets could generate a high cpu usage since
this call is done for each socket at each iteration of the main loop.
Since there is also a WSAEvent that gets signaled when there is
a change in the status of the socket, checking its status and
calling the WSAEnumNetworkEvents API only if the event is signaled,
can reduce the overall cpu usage.
This was one of the code paths not currently covered by unit tests, so
let’s add a test for it.
This tests what happens when a structurally valid GResource file, but
with a corrupt entry for a compressed file, is loaded.
Signed-off-by: Philip Withnall <pwithnall@gnome.org>
Helps: #3465
This is slightly more involved than the previous couple of commits, as
`g_resource_lookup_data()` can return two errors: one if the resource is
not found, and another if decompression fails.
We want to avoid allocating the `G_RESOURCE_ERROR_NOT_FOUND` error, as
`g_resources_lookup_data()` will be looping through multiple
`GResource`s trying to find the given path, and all but one of them will
return `G_RESOURCE_ERROR_NOT_FOUND`. For a large application, this can
amount to a lot of `GError`s allocated and then immediately freed on
startup.
Use the split from the previous commit to replace the call to
`g_resource_lookup_data()` with its two constituent parts. We can then
handle errors from them separately, ignoring the `NOT_FOUND` error from
`do_lookup()`, while paying attention to any errors from
`resource_to_bytes()`.
This should result in no functional difference to
`g_resources_lookup_data()`, but fewer allocations overall.
Spotted by Christian Hergert.
Signed-off-by: Philip Withnall <pwithnall@gnome.org>
Helps: #3465
It’s now a call to `do_lookup()` followed by a call to
`resource_to_bytes()`. This makes no functional changes, but will be
useful in the following commit.
Signed-off-by: Philip Withnall <pwithnall@gnome.org>
Helps: #3465
As with the previous commit:
The error here can only ever be `G_RESOURCE_ERROR_NOT_FOUND`, which
`g_resources_get_info()` immediately frees. So let’s avoid allocating
the error in the first place.
Spotted by Christian Hergert.
Signed-off-by: Philip Withnall <pwithnall@gnome.org>
Helps: #3465
The error here can only ever be `G_RESOURCE_ERROR_NOT_FOUND`, which
`g_resources_open_stream()` immediately frees. So let’s avoid allocating
the error in the first place.
Spotted by Christian Hergert.
Signed-off-by: Philip Withnall <pwithnall@gnome.org>
Helps: #3465
This makes it a bit easier to make sure all the translatable strings are
kept in sync. It introduces no functional changes.
Signed-off-by: Philip Withnall <pwithnall@gnome.org>
Helps: #3465
The test in `unix-mounts` to see whether `g_unix_mounts_get_from_file()`
can parse an example file was working fine when GLib is built with
libmount, but not when built without it (and hence typically using
`getmntent()`).
This is because libmount supports mountinfo files (like
`/proc/self/mountinfo`), but `getmntent()` only supports mount files
(like `/proc/mounts`). The test was written only with the former.
So, change the test to use mount files when GLib is built without
libmount support.
Signed-off-by: Philip Withnall <pwithnall@gnome.org>
Fixes: #3456
This kind of deeply nested menu is definitely no longer good UI practice
for most apps (deeply nested menus make things hard to find, and require
good mouse control to navigate). However, it does serve as a good
demonstration of the concepts in `GMenuModel`, so keep it, with a
sentence to acknowledge that it’s not a good UI.
Signed-off-by: Philip Withnall <pwithnall@gnome.org>
Fixes: #3451
These tests are expected to cause a thread to deadlock. That seems to be
fine with glibc on Linux, but the glibc version on FreeBSD can detect
the deadlock, and aborts the whole test process with:
```
GLib (gthread-posix.c): Unexpected error from C library during 'pthread_mutex_lock': Resource deadlock avoided. Aborting.
```
This is fair enough.
To avoid this causing the test suite to fail, run those two tests in
subprocesses. This also means we’re not carrying a deadlocked thread
around for the rest of the test suite.
Improves on commit 62192925b6.
Signed-off-by: Philip Withnall <pwithnall@gnome.org>
Provide examples of what they all represent, and expand on the
descriptions of them in a few places.
Move references to their equivalents from `GnomeVFS` to lower down in
the documentation, since `GnomeVFS` has been deprecated for many years
now, and is unlikely to be pertinent to most readers.
Signed-off-by: Philip Withnall <pwithnall@gnome.org>
When listing schemas from a specified directory, explicitly
create the GSettings object from the schema, don't allow g_settings_new
to do the usual lookup. That lookup fails if no other schemas are
installed in the default directories.
Fixes#3429.
This can never have been tested, it was returning `GUnixMountEntry`
structs from functions which are typed to return `GUnixMountPoint`s.
Signed-off-by: Philip Withnall <pwithnall@gnome.org>
While GLib doesn’t parse these files, it does provide API to access the
fields from them, and does implement some logic based on options fields
in them. It would be nice to be able to test that, and get coverage of
the methods for `GUnixMountPoint` and `GUnixMountEntry`.
We don’t expect users to start querying the fstab or mtab by explicitly
loading data from those file paths. These functions are mainly intended
to prove a controllable entry point into the `gunixmounts.c` code for
unit testing.
It means we can provide a file with controllable contents in order to
test the mount entry/point code on.
See: https://gitlab.gnome.org/GNOME/glib/-/merge_requests/4155
Signed-off-by: Philip Withnall <pwithnall@gnome.org>
