The `struct timespec` type documented as being passed to the `futex()`
syscall actually needs to be the *kernel’s* timespec type. This will be
a different width from the userspace timespec type if running a 64-bit
userspace on a 32-bit kernel.
That mismatch will cause `g_cond_wait_until()` to return `FALSE`
immediately.
No other uses of `futex()` in GLib use the timeout argument, so they’re
all OK.
Following a detailed suggestion by Rich Felker, pass a different
timespec type into `futex()` if `__NR_futex_time64` is defined. That’s
the 64-bit time version of `futex()` which was added in kernel 5.1, and
which was only added for 32-bit kernels.
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
Fixes: #2634
Inspired by https://gitlab.gnome.org/GNOME/glib/-/issues/2638.
This requires moving the tests to subprocesses, so that
`g_log_set_writer_func()` is only called once per process.
It also adds a test for the new error.
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
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>
As noted by Charlie Barto: "both mingw64 and msvc define _WIN64 in
64-bit mode, and both are LLP64, and both have struct _stat64 defined
the same way."
https://gitlab.gnome.org/GNOME/glib/-/merge_requests/2449#note_1372190
Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
The pipe must be closed, or the child PID watch doesn't get triggered.
We should remove the message callback source on EOF, as EOF during main
loop run will reach a bad assert in the callback.
Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
...for certain Windows locales, since the formats accepted for
g_date_set_parse() will vary depending on the current system locale. For
instance, g_date_set_parse(gdate, "dd/mm/yy") is accepted on locales such
as zh-HK (Chinese (Hong Kong SAR)) but is rejected on zh-TW (Chinese
(Taiwan)).
One can tell from the "date format" settings in the Windows system
control panel whether there is a "dd/MM/YYYY" or "dd/MM/YY" option from the
drop-down list of date formats to display for the locale, which will indicate
whether g_date_set_parse(gdate, "dd/mm/yy") is accepted, which is true for
zh-HK but is not true for zh-TW.
If g_date_set_parse(gdate, "dd/mm/yy") is not accepted, try again with
g_date_set_parse(gdate, "yy/mm/dd") thereafter for the 2-digit-year tests.
It currently fails on FreeBSD with:
```
Bail out! GLib:ERROR:../glib/tests/spawn-test.c:111:test_spawn_basics: assertion failed (erroutput == "sort: cannot read: non-existing-file.txt: No such file or directory\n"): ("sort: No such file or directory\n" == "sort: cannot read: non-existing-file.txt: No such file or directory\n")
```
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
