- May 14, 2020
-
-
Kevin Pouget authored
These buttons are not mapped by GTK (`/usr/include/gtk-3.0/gdk/gdktypes.h`) but still exposed, so I hardcoded their values: #define _GTK_BUTTON_PREV_MASK (1 << 15) /* button 9 */ #define _GTK_BUTTON_NEXT_MASK (1 << 16) /* button 8 */
-
- May 11, 2020
-
-
We can now properly manage mouse pointer lock by using the wayland protocols. Includes fix from Frediano Ziglio to manage switching from mouse in server mode to client mode without releasing the pointer (see fziglio/spice-gtk@c14b047e ) Signed-off-by:
Francesco Giudici <fgiudici@redhat.com> Acked-by:
Frediano Ziglio <fziglio@redhat.com>
-
This will be used later to couple it with relative mouse movement under Wayland: we just get the mouse movement from the Wayland protocols callback. Signed-off-by:
Francesco Giudici <fgiudici@redhat.com> Acked-by:
Frediano Ziglio <fziglio@redhat.com>
-
add utilities to lock the pointer to a window and to get relative mouse movement in Wayland. This is made possible thanks to the wayland protocols introduced in the previous commit. Code freely taken and adapted from Christophe Fergeau branch: https://gitlab.freedesktop.org/teuf/spice-gtk/-/tree/wayland Signed-off-by:
Francesco Giudici <fgiudici@redhat.com> Acked-by:
Frediano Ziglio <fziglio@redhat.com>
-
Generate wayland protocols: these will be used later for locking the mouse pointer and getting relative mouse movements when using the mouse in server mode in Wayland. The meson build steps have been freely taken and adapted from Weston source code: https://gitlab.freedesktop.org/wayland/weston/-/blob/master/protocol/meson.build Signed-off-by:
Francesco Giudici <fgiudici@redhat.com> Acked-by:
Frediano Ziglio <fziglio@redhat.com>
-
- May 06, 2020
-
-
When SASL is in use, its buffer may contain remaining data from previously received messages. The spice_channel_iterate_read should use it, even if c->in socket is not readable. Fixes: spice/spice-gtk#126 Acked-by:
Frediano Ziglio <fziglio@redhat.com>
-
- Apr 30, 2020
-
-
Frediano Ziglio authored
The executable is not intended to be called by users. Many distributions already relocate the executable to another directory. This fixes Gitlab issue #123. Signed-off-by:
Frediano Ziglio <freddy77@gmail.com> Acked-by:
Francesco Giudici <fgiudici@redhat.com>
-
- Apr 26, 2020
-
-
Avoid one ternary operator inside another ternary operator. Also remove no-C syntax. Signed-off-by:
Frediano Ziglio <fziglio@redhat.com> Acked-by:
Francesco Giudici <fgiudici@redhat.com>
-
- Apr 25, 2020
-
-
Fixes: spice/spice-gtk#124 Acked-by:
Frediano Ziglio <fziglio@redhat.com>
-
- Apr 24, 2020
-
-
Frediano Ziglio authored
This allows SPICE common to request the right version. Signed-off-by:
Frediano Ziglio <freddy77@gmail.com>
-
Frediano Ziglio authored
This brings in the following changes: Frediano Ziglio (3): helper-fuzzer-demarshallers: Add an helper for fuzzy testing demarshallers code Add helper code for agent messages agent: Fix incompatibility with Visual Studio compiler Signed-off-by:
Frediano Ziglio <freddy77@gmail.com>
-
Frediano Ziglio authored
This brings in the following changes: Frediano Ziglio (1): build: Allows to request a greater SPICE protocol version Signed-off-by:
Frediano Ziglio <freddy77@gmail.com>
-
- Apr 22, 2020
-
-
Frediano Ziglio authored
usbredir only supports little endian, using on big endian machine won't work causing crashes. Signed-off-by:
Frediano Ziglio <freddy77@gmail.com> Acked-by:
Francesco Giudici <fgiudici@redhat.com>
-
- Apr 16, 2020
-
-
On file systems without capabilities attributes, drop capabilities at runtime from the start. Signed-off-by:
Marc-André Lureau <marcandre.lureau@redhat.com> Acked-by:
Frediano Ziglio <fziglio@redhat.com>
-
A privileged executable should operate with the least amount of privileges possible. For spice-client-glib-usb-acl-helper, this is CAP_FOWNER for acl_set_file() to succeed. Signed-off-by:
Marc-André Lureau <marcandre.lureau@redhat.com> Acked-by:
Frediano Ziglio <fziglio@redhat.com>
-
- Apr 09, 2020
-
-
It was released on 2011-03-03, and is available in old debian stable and all distros we care about. Signed-off-by:
Marc-André Lureau <marcandre.lureau@redhat.com> Acked-by:
Frediano Ziglio <fziglio@redhat.com>
-
emacs mode got removed with: git ls-tree HEAD -r --name-only | xargs sed -i '/Mode: C/d' Signed-off-by:
Marc-André Lureau <marcandre.lureau@redhat.com> Acked-by:
Frediano Ziglio <fziglio@redhat.com>
-
- Apr 07, 2020
-
-
Frediano Ziglio authored
After commit 47e149705a5bb1d4a615facfd9bd0190d9b0647c (cfr "proto: Generate correct constants for smartcard support") msgc_smartcard_header was renamed to a more proper msgc_smartcard_data. Signed-off-by:
Frediano Ziglio <fziglio@redhat.com> Acked-by:
Victor Toso <victortoso@redhat.com>
-
Frediano Ziglio authored
This brings in the following changes: Frediano Ziglio (7): ci: Fix build of spice-protocol lz_common: Remove unused commented out define test-logging: Do not use G_PASTE to join level ci: Use spice-protocol master for make-win job build: Allow to build disabling code generation proto: Allows to specify @deprecated option for flags and enums messages: Remove obsolete structure definition Kevin Pouget (1): spice.proto: add quality_indicator messages Victor Toso (1): sound: remove celt support Signed-off-by:
Frediano Ziglio <fziglio@redhat.com> Acked-by:
Victor Toso <victortoso@redhat.com>
-
Frediano Ziglio authored
VDAgentFileXferStatusMessage message can or cannot contain detailed information attached to it. Detect this correctly checking capabilities and flags. This fixes a small buffer overflow reading in case the details are off the payload. Signed-off-by:
Frediano Ziglio <fziglio@redhat.com> Acked-by:
Victor Toso <victortoso@redhat.com>
-
Frediano Ziglio authored
Signed-off-by:
Frediano Ziglio <fziglio@redhat.com> Acked-by:
Victor Toso <victortoso@redhat.com>
-
- Apr 04, 2020
-
-
Acked-by:
Frediano Ziglio <fziglio@redhat.com>
-
- Mar 20, 2020
-
-
Frediano Ziglio authored
Some mistake in recent patch, thanks to Marc-André's eagle eyes. Signed-off-by:
Frediano Ziglio <fziglio@redhat.com> Acked-by:
Victor Toso <victortoso@redhat.com>
-
- Mar 19, 2020
-
-
Frediano Ziglio authored
In big endian system GUINT32_TO_LE macro uses the parameter multiple time causing serial to be incremented multiple times instead of one. Avoid side effects using a temporary variable. Signed-off-by:
Frediano Ziglio <fziglio@redhat.com> Acked-by:
Victor Toso <victortoso@redhat.com>
-
Frediano Ziglio authored
Signed-off-by:
Frediano Ziglio <fziglio@redhat.com> Acked-by:
Victor Toso <victortoso@redhat.com>
-
- Mar 17, 2020
-
-
Frediano Ziglio authored
Not used anymore Signed-off-by:
Frediano Ziglio <fziglio@redhat.com>
-
Victor Toso authored
GStreamer is a hard requirement since ac0e50f5 or v0.36 wich is the same release that PulseAudio backend was deprecated: 9a4b3bcb "build-sys: deprecate the pulseaudio backend" in 2019-01-14 Signed-off-by:
Victor Toso <victortoso@redhat.com> Acked-by:
Marc-André Lureau <marcandre.lureau@redhat.com>
-
Victor Toso authored
Follow up of spice-protocol's deprecation of celt mode. See: spice/spice-protocol!15 Signed-off-by:
Victor Toso <victortoso@redhat.com> Acked-by:
Marc-André Lureau <marcandre.lureau@redhat.com>
-
- Mar 16, 2020
-
-
Victor Toso authored
Signed-off-by:
Victor Toso <victortoso@redhat.com> Acked-by:
Marc-André Lureau <marcandre.lureau@redhat.com> Acked-by:
Frediano Ziglio <fziglio@redhat.com>
-
Victor Toso authored
Generated with `ninja spice-gtk-update-po` Fixes: spice/spice-gtk#119 Signed-off-by:
Victor Toso <victortoso@redhat.com> Acked-by:
Marc-André Lureau <marcandre.lureau@redhat.com> Acked-by:
Frediano Ziglio <fziglio@redhat.com>
-
Victor Toso authored
So we can have the v0.38 release with a few more tests plus the removal of warning in meson build. Frediano Ziglio (8): test-quic: Avoid namespace conflict with Gdk API test-quic: Cache gdk_pixbuf_get_byte_length value test-quic: Reduce height of test image test-quic: Run 1 random quic test per color mode m4: Make easier to integrate in other projects Add missing dependency in Autoconf integration test-utils: Add a test for some utils.h functions build: Use warning_level option for Meson Kevin Pouget (1): utils: add set_bitmap/test_bitmap functions Signed-off-by:
Victor Toso <victortoso@redhat.com> Acked-by:
Marc-André Lureau <marcandre.lureau@redhat.com> Acked-by:
Frediano Ziglio <fziglio@redhat.com>
-
- Mar 10, 2020
-
-
Victor Toso authored
Thanks to @Illumaria Fixes: #96
-
Marc-André Lureau authored
Signed-off-by:
Marc-André Lureau <marcandre.lureau@redhat.com>
-
Marc-André Lureau authored
The protocol uses a u8 for the selection value. Make sure the given argument value fits there, or throw a critical. The other places seem to use u8 variables already. Signed-off-by:
Marc-André Lureau <marcandre.lureau@redhat.com>
-
Marc-André Lureau authored
Signed-off-by:
Marc-André Lureau <marcandre.lureau@redhat.com>
-
Marc-André Lureau authored
Delay the release events for 0.5 sec. If no further grab comes in, then release the grab. Otherwise, let's skip the release. This avoids some races with clipboard managers. Related to: spice/spice-gtk#82 Signed-off-by:
Marc-André Lureau <marcandre.lureau@redhat.com>
-
Marc-André Lureau authored
On the client side, whenever the grab owner changes (and the clipboard was previously grabbed), spice-gtk sends a clipboard release followed immediately by a new grab. But some clipboard managers on the remote side react to clipboard release events by taking a clipboard grab, presumably to avoid empty clipboards. The two grabs, coming from the client and from the remote sides, will race in both directions, which may confuse the client & remote side, as both believe the other side is the current grab owner, and thus further clipboard data requests are likely to fail. Let's avoid sending a release event when re-grabing. The race described above may still happen in other rare circunstances, and will require a protocol change. To avoid the conflict, a discussed solution could use a clipboard serial number. Tested with current linux & windows vdagent. Looking at earlier version of the code, it doesn't seem like subsequent grabs will be treated as an error. Signed-off-by:
Marc-André Lureau <marcandre.lureau@redhat.com>
-
- Mar 09, 2020
-
-
Frediano Ziglio authored
warning_level 2 setting for Meson add the same "-Wall -Wextra" options to GCC compiler. This removes a warning using Meson. Signed-off-by:
Frediano Ziglio <fziglio@redhat.com>
-
Frediano Ziglio authored
Usually the default options for C are "-O2 -g". So this build type is more conservative. Normal users wanting to build the package won't find themselves with a bad performance executable. This will catch some potentially warning not detected with debug build. Signed-off-by:
Frediano Ziglio <fziglio@redhat.com>
-
Frediano Ziglio authored
In Meson better to use --werror option instead to manually add to options Signed-off-by:
Frediano Ziglio <fziglio@redhat.com>
-