blob: cf91dbc6412ecb9e068f601a3f6dc003419ef735 (
plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
|
https://github.com/mate-desktop/mate-settings-daemon/commit/42e91f2805c508317b4d26166cb3db332dd313e8.patch
From: Colomban Wendling <cwendling@hypra.fr>
Date: Tue, 27 Jun 2023 12:21:16 +0200
Subject: [PATCH] a11y-keyboard-atspi: Fix memory leak
Fix fairly large memory leak when beeping on keys while caps lock is
enabled. The libatspi2 docs and API were quite misleading, so I
overlooked the fact the event parameter should be freed in the
callback.
This changes the constness of the callback argument, which is new in
libatspi2 2.40 -- yet the actual behavior didn't change, only the
qualifier was removed, see [1].
This might however bring up a compiler warning when building against
libatspi2 < 2.40; but on the other hand it fixed build with
clang >= 16, see #399. As it is unlikely to build with clang >= 16
and libatspi2 < 2.40, I think it's a good compromise.
[1] https://gitlab.gnome.org/GNOME/at-spi2-core/-/commit/7dfb0b7fc2d1710ef7fad54f910fa4c6a5e3af17
--- a/plugins/a11y-keyboard/msd-a11y-keyboard-atspi.c
+++ b/plugins/a11y-keyboard/msd-a11y-keyboard-atspi.c
@@ -55,16 +55,16 @@ msd_a11y_keyboard_atspi_class_init (MsdA11yKeyboardAtspiClass *klass)
}
static gboolean
-on_key_press_event (const AtspiDeviceEvent *event,
- void *user_data G_GNUC_UNUSED)
+on_key_press_event (AtspiDeviceEvent *event,
+ void *user_data G_GNUC_UNUSED)
{
/* don't ring on capslock itself, that's taken care of by togglekeys
* if the user want it. */
- if (event->id == GDK_KEY_Caps_Lock)
- return FALSE;
-
- gdk_display_beep (gdk_display_get_default ());
+ if (event->id != GDK_KEY_Caps_Lock)
+ gdk_display_beep (gdk_display_get_default ());
+ /* cast the possible erroneous const away with atspi < 2.40 */
+ g_boxed_free (ATSPI_TYPE_DEVICE_EVENT, (gpointer) event);
return FALSE;
}
|
