This is a rough intro to what Xwayland is and does, with just one
implementation detail so far (Window identification).
I paid no attention to formatting details, those can be polished in
follow-ups. I just want the prose out.
I also just quickly whacked up the diagram, would be happy to see
someone replace it with a nicer one. I just didn't have time to learn
dot for now.
v2:
- typo fix
- rephrase "talking to hardware" as "driving the displays"
- mention circular dependency in intro
- add section to explain rootless and rootful modes
- remove paragraph about Xwayland protocol usage
- move TBD part to the end under a new section header
v3:
- use "advantage" and "disadvantage" instead of "pro" and "con"
- slight rewording on rootful mode and rootless mode paragraphs
- removed the paragraph about the lack of shell and special Wayland
protocol extensions
- removed the commented out list of ideas to write
v4:
- typo fixes pointed out by Yong
Cc: Olivier Fourdan <ofourdan@redhat.com>
Cc: Jonas Ådahl <jadahl@gmail.com>
Cc: Daniel Stone <daniel@fooishbar.org>
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>
Reviewed-by: Jonas Ådahl <jadahl@gmail.com>
In order to support system compositor instances, it is necessary to
allow clients' wl_display_connect() to find the compositor's listening
socket somewhere outside of XDG_RUNTIME_DIR. For a full account, see
the discussion beginning here:
https://lists.freedesktop.org/archives/wayland-devel/2017-November/035664.html
This change adjusts the client-side connection logic so that, if
WAYLAND_DISPLAY is formatted as an absolute pathname, the socket
connection attempt is made to just $WAYLAND_DISPLAY rather than
usual user-private location $XDG_RUNTIME_DIR/$WAYLAND_DISPLAY.
This change is based on Davide Bettio's submission of the same concept
at:
https://lists.freedesktop.org/archives/wayland-devel/2015-August/023838.html.
v4 changes:
* Improved internal comments and some boundary-condition
error checks in test case.
* Refer to compositor as "Wayland server" rather than "Wayland
display" in wl_display_connect() doxygen comments.
* Remove redundant descriptions of parameter-interpretation
mechanics from wl_display_connect() manpage. Reworked things
to make it clear that 'name' and $WAYLAND_DISLAY are each
capable of encoding absolute server socket paths.
* Remove callout to reference implementation behavior in protocol
documented. In its place there is now a simple statement that
implementations can optionally support absolute socket paths.
v3 changes:
* Added test case.
* Clarified documentation to note that 'name' parameter to wl_display_connect()
can also be an absolute path.
v2 changes:
* Added backward incompatibility note to wl_display_connect() manpage.
* Rephased wl_display_connect() manpage changes to precisely match actual
changed behavior.
* Added mention of new absolute path behavior in wl_display_connect()
doxygen comments.
* Mentioned new absolute path interpretation of WAYLAND_DISPLAY in
protocol documentation.
Signed-off-by: Matt Hoosier <matt.hoosier@gmail.com>
Acked-by: Jonas Ådahl <jadahl@gmail.com>
Reviewed-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>
The existing specification was not explicitly clear on when
wl_subcompositor.get_subsurface request actually adds the sub-surface to
the parent in the compositor's scenegraph. The implicit assumption was
that this happens immediately, but it was not written anywhere.
If it happens immediately, the client doing things in a wrong order may
cause a glitch on screen. Particularly, if the wl_surface B that is
going to be a sub-surface for wl_surface A (the parent) already has a
buffer committed, and the parent surface is mapped, then get_subsurface
will (may?) cause wl_surface B to become mapped immediately. That leaves
no time to set up the sub-surface z-order or position before mapping,
hence there can be a visible glitch.
The way to avoid that, given that the parent surface is mapped, is to
not commit a buffer to wl_surface B until all the sub-surface setup is
done.
However, doing the sub-surface setup always requires a wl_surface.commit
on the parent surface unless the defaults happen to be correct.
To make setting up a subsurface slightly easier by removing one
possibility for a glitch, this patch amends the specification to require
a wl_surface.commit on the parent surface for get_subsurface to
complete. The sub-surface cannot become mapped before a parent commit.
This change may break existing clients that relied on the glitchy
sequence to not need a parent surface commit to map the sub-surface.
However, presumably all uses would at least issue a
wl_subsurface.set_position, which requires a parent surface commit to
apply. That would guarantee that there is a parent surface commit after
get_subsurface, and so reduces the chances of breaking anything.
In other cases, this change may simply remove a possibility for the
glitch.
This patch also adds a note about changing wl_surface.commit behaviour
on wl_subcompositor.get_subsurface. (That could be a separate patch.)
The behaviour of wl_subsurface.destroy remains as specified, even though
it is now slightly asymmetrical to get_subsurface. This is emphasized by
adding the word "immediately". The effects of destruction were already
explicitly documented, as is the way to achieve synchronized unmapping,
so changing destruction behaviour would likely be more disruptive, and
also open up more corner cases (what would happen between destroy and
unmapping?).
Bug: https://phabricator.freedesktop.org/T7358
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>
Reviewed-by: Daniel Stone <daniels@collabora.com>
Reviewed-by: Derek Foreman <derekf@osg.samsung.com>
Reviewed-by: Jonas Ådahl <jadahl@gmail.com>
Reviewed-by: Martin Gräßlin <mgraesslin@kde.org>
Add a --strict flag for making wayland-scanner fail if the DTD
verification fails. This is useful for testing, so that a test case can
fail a scan when the protocol doesn't comply with the DTD.
Signed-off-by: Jonas Ådahl <jadahl@gmail.com>
There were two places where we did the same calculation manually.
Signed-off-by: Derek Foreman <derekf@osg.samsung.com>
Reviewed-by: Daniel Stone <daniels@collabora.com>
Both the blocks in this if/else clause do the same thing, so combine
the comparisons into one.
No functional change.
Signed-off-by: Derek Foreman <derekf@osg.samsung.com>
Reviewed-by: Bryce Harrington <bryce@osg.samsung.com>
Reviewed-by: Daniel Stone <daniels@collabora.com>
All current callers close all fds, so this has gone unnoticed, but if
we close less than all fds with close_fds() we leak all the unclosed
ones and ruin further event demarshalling.
A future patch will close less than the full buffer's worth of fds,
so this is now noticed.
Signed-off-by: Derek Foreman <derekf@osg.samsung.com>
Reviewed-by: Daniel Stone <daniels@collabora.com>
There is a lot of files created with .so links to non-installed
files, making most of installed pages useless. The files
referenced in .so links are not suitable for installation nor
do they contain any useful information for them to be worth
fixing.
Signed-off-by: Armin Krezović <krezovic.armin@gmail.com>
Acked-by: Daniel Stone <daniels@collabora.com>
This *technically* changes the semantics of the return value of the source callbacks.
Previously you could return a negative number from a source callback and it would prevent
*other* source callbacks from triggering a subsequent recheck.
Doing that seems like such a bad idea it's not worth supporting.
v2: Log this case if it is hit, so we don't silently change behaviour.
Signed-off-by: Christopher James Halse Rogers <christopher.halse.rogers@canonical.com>
Reviewed-by: Daniel Stone <daniels@collabora.com>
Now that xdg_shell is stable and much better defined than wl_shell we
can finally deprecate wl_shell and guide users towards xdg_shell
instead.
Signed-off-by: Jonas Ådahl <jadahl@gmail.com>
Acked-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>
Reviewed-by: Daniel Stone <daniels@collabora.com>
This makes the header self-contained, since the struct is considered
opaque from waylad-cursor POV.
As we're here move the wl_shm fwd. declaration alongside the others.
Making it easier to read and track.
Signed-off-by: Emil Velikov <emil.velikov@collabora.com>
Reviewed-by: Yong Bakos <ybakos@humanoriented.com>
Without this attribute, these macros were making Weston’s tests fail to
build with LTO enabled.
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=94602
Signed-off-by: Markus Trippelsdorf <markus@trippelsdorf.de>
Signed-off-by: Emmanuel Gil Peyrot <emmanuel.peyrot@collabora.com>
Tested-by: Emmanuel Gil Peyrot <emmanuel.peyrot@collabora.com>
Reviewed-by: Daniel Stone <daniels@collabora.com>
This is a preparatory patch for the next one.
Signed-off-by: Emmanuel Gil Peyrot <emmanuel.peyrot@collabora.com>
Reviewed-by: Daniel Stone <daniels@collabora.com>
It is possible to trigger heap overflows due to an integer overflow
while parsing images.
The integer overflow occurs because the chosen limit 0x10000 for
dimensions is too large for 32 bit systems, because each pixel takes
4 bytes. Properly chosen values allow an overflow which in turn will
lead to less allocated memory than needed for subsequent reads.
See also: https://cgit.freedesktop.org/xorg/lib/libXcursor/commit/?id=4794b5dd34688158fb51a2943032569d3780c4b8
Fixes: https://bugs.freedesktop.org/show_bug.cgi?id=103961
Signed-off-by: Tobias Stoeckmann <tobias@stoeckmann.org>
[Pekka: add link to the corresponding libXcursor commit]
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>
Add some inline information, what the macro is used for, why it came to
be and what we shouldn't do if we consider further deprecation in the future
deprecation.
Cc: Pekka Paalanen <pekka.paalanen@collabora.co.uk>
Signed-off-by: Emil Velikov <emil.velikov@collabora.com>
Reviewed-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>
I've heard some complaints that wl_display.get_registry "leaks"
server memory because wl_registry has no destructor. While this isn't
strictly true - all those resources are freed when the client
disconnects - it's a bit of a gotcha for neophytes.
Since wl_registry's version is not requested in any way through
wl_display.get_registry, we can't add a destructor request without
breaking ABI. So let's be a little more clear about the result
of getting too many wl_registry objects.
Signed-off-by: Derek Foreman <derekf@osg.samsung.com>
Reviewed-by: Daniel Stone <daniels@collabora.com>
Reviewed-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>
All the foo.in files are in the tarball, as long as their foo
counterparts are listed in AC_CONFIG_FILES
For example - *.pc.in, Makefile.in files, etc.
Signed-off-by: Emil Velikov <emil.velikov@collabora.com>
Reviewed-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>
This documents all the public API related to wl_event_loop and
wl_event_source objects.
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>
Reviewed-by: Yong Bakos <ybakos@humanoriented.com>
[Pekka: fixed typos pointed by Yong]
[Pekka: fixed typos pointed by Christopher]
Reviewed-By: Christopher James Halse Rogers <christopher.halse.rogers@canonical.com>
When input/output files are passed as arguments to wayland-scanner,
instead of using stdin/stdout, warning and error messages will contain
the file name, together with line number, of the warning/error. Doing
this helps IDEs jump to the correct line.
Signed-off-by: Jonas Ådahl <jadahl@gmail.com>
Reviewed-by: Emil Velikov <emil.velikov@collabora.com>
[Pekka: dropped the src/scanner.mk hunk, file deleted]
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>
File uses tabs, barring the few instances fixed with this patch.
Signed-off-by: Emil Velikov <emil.velikov@collabora.com>
Reviewed-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>
Nothing in the existing codebase references the file.
Signed-off-by: Emil Velikov <emil.velikov@collabora.com>
Reviewed-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>
With recent versions of graphviz, generation of the diagrams in the documentation
fails with:
/usr/bin/dot -Tpng -oxml/x-architecture.png dot/x-architecture.gv
Warning: flat edge between adjacent nodes one of which has a record shape - replace records with HTML-like labels
Edge xserver -> comp
Error: getsplinepoints: no spline points available for edge (xserver,comp)
Error: lost xserver comp edge
Error: lost xserver comp edge
Error: lost comp xserver edge
Error: lost comp xserver edge
http://www.graphviz.org/content/i-havent-been-able-render-these-files-graphviz-226 indicates
that the error message basically means that the authors of graphviz consider record-style
labels to be deprecated and are no longer fixing errors with them. This patch changes
the labels to be in the HTML style, which seems to require duplicating style between all
the nodes, but it's not like these files are often edited.
The result is not exactly the same but is quite similar.
Reviewed-by: Armin Krezović <krezovic.armin@gmail.com>
Tested-by: Armin Krezović <krezovic.armin@gmail.com>
All fd.o projects are now covered by the Contributor Covenant. Include a
reference to this in the Contributing doc, making it clear that we are
all expected to behave like human beings.
Signed-off-by: Daniel Stone <daniels@collabora.com>
Reviewed-by: Yong Bakos <ybakos@humanoriented.com>
Acked-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>
Acked-by: Jonas Ådahl <jadahl@gmail.com>
GCC 7 now requires an explicit comment noting that case statements
without a break fall through. We already had one of those in the
scanner, but GCC wasn't smart enough to pick it up.
Quiet the warning by making the comment less elaborate.
Signed-off-by: Daniel Stone <daniels@collabora.com>
Reviewed-by: Yong Bakos <ybakos@humanoriented.com>
Used only internally and explicitly marked as such with commit
cf04b0a18f ("Move private definitions and prototypes to new
zwayland-private.h")
Signed-off-by: Emil Velikov <emil.velikov@collabora.com>
Reviewed-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>
Those struct members are no longer used so we can remove them.
Signed-off-by: Sergi Granell <xerpi.g.12@gmail.com>
Reviewed-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>
New IDs are internally dealt with as objects, however this test
expected to deal with 'n' as the uint32_t type that's just seen
through the wire. We should give it an object instead, and
expect an object from it.
https://bugs.freedesktop.org/show_bug.cgi?id=99899
Signed-off-by: Carlos Garnacho <carlosg@gnome.org>
Reviewed-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>
Tested-by: Kalev Lember <kalevlember@gmail.com>
Reviewed-by: Yong Bakos <ybakos@humanoriented.com>
Fix this set of warnings appearing three times during a build:
/home/pq/git/wayland/src/wayland-server.c:1868: warning: class
`wl_priv_signal' for related function `wl_priv_signal_init' is not
documented.
/home/pq/git/wayland/src/wayland-server.c:1884: warning: class
`wl_priv_signal' for related function `wl_priv_signal_add' is not
documented.
/home/pq/git/wayland/src/wayland-server.c:1899: warning: class
`wl_priv_signal' for related function `wl_priv_signal_get' is not
documented.
Our Wayland docbook don't include private things, so make sure these do
not end up there. This removes the mention of wl_priv_signal_emit from
the Server API docbook. I have no idea why the other functions did not
appear there.
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>
Reviewed-by: Yong Bakos <ybakos@humanoriented.com>
connection-test.c did not cover wl_argument_from_va_list, so add one
test that specifically tests this method.
Signed-off-by: Yong Bakos <ybakos@humanoriented.com>
Reviewed-by: Daniel Stone <daniels@collabora.com>
Check that all the objects in an event belong to the same client as
the resource posting it. This prevents a compositor from accidentally
mixing client objects and posting an event that causes a client to
abort with a cryptic message.
Instead the client will now be disconnected as it is when the compositor
tries to send a null for a non-nullable object, and a log message
will be printed by the compositor.
Reviewed-by: Yong Bakos <ybakos@humanoriented.com>
Reviewed-by: Bryce Harrington <bryce@osg.samsung.com>
Signed-off-by: Derek Foreman <derekf@osg.samsung.com>
Reviewed-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>
Until now, we haven't done anything to prevent sending additional
events to clients after posting an error.
Acked-by: Daniel Stone <daniels@collabora.com>
Signed-off-by: Derek Foreman <derekf@osg.samsung.com>
Reviewed-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>
These have grown a little in size but are almost identical, factor
out the common code.
Signed-off-by: Derek Foreman <derekf@osg.samsung.com>
Reviewed-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>
The old wl_signal is kept for backwards compatibility, as that is also
present in the deprecated public wl_resource struct, and that must be
kept working.
Signed-off-by: Giulio Camuffo <giulio.camuffo@kdab.com>
Reviewed-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>
wl_list_for_each_safe, which is used by wl_signal_emit is not really
safe. If a signal has two listeners, and the first one removes and
re-inits the second one, it would enter an infinite loop, which was hit
in weston on resource destruction, which emits a signal.
This commit adds a new version of wl_signal, called wl_priv_signal,
which is private in wayland-server.c and which does not have this problem.
The old wl_signal cannot be improved without breaking backwards compatibility.
Signed-off-by: Giulio Camuffo <giulio.camuffo@kdab.com>
Reviewed-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>
When using a wl_global, a server may need to retrieve the associated
wl_interface and user data.
Add a couple of convenient functions wl_global_get_interface() and
wl_global_get_user_data() for this purpose.
Signed-off-by: Olivier Fourdan <ofourdan@redhat.com>
Reviewed-by: Yong Bakos <ybakos@humanoriented.com>
Reviewed-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>
Add a new API to let compositor decide whether or not a wl_global
should be advertised to the clients via wl_registry_bind() or
display_get_registry()
By using its own filter, the compositor can decide which wl_global would
be listed to clients.
Compositors can use this mechanism to hide their own private interfaces
that regular clients should not use.
- Hiding interfaces that expose compositor implementation details
makes it harder for clients to identify the compositor. Therefore
clients are a little less likely to develop compositor-specific
workarounds instead of reporting problems upstream.
- Hiding can be used to diminish the problems from missing namespacing:
if two compositors happen to use the same named global with
different interfaces for their special-purpose clients, the client
expecting the different interface would probably never see it
advertised.
Signed-off-by: Olivier Fourdan <ofourdan@redhat.com>
Reviewed-by: Jonas Ådahl <jadahl@gmail.com>
Reviewed-by: Yong Bakos <ybakos@humanoriented.com>
Acked-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>
Reviewed-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>
