We were emitting the extern declarations of all types used in the protocol,
even if not defined in it. This caused warnings to be produced when using
the -Wredundant-decls compiler flag when building an extension that uses
e.g. wl_surface. However we only need the extern declarations if the
protocol defines a factory for those external interfaces. That is a
bad design and can be however done by including the dependent protocol
header first.
So only emit the extern declarations for the types that the protocol
actually defines, this restoring the behavior we were using in 1.7.
Fixes https://bugs.freedesktop.org/show_bug.cgi?id=90677
Reviewed-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>
Tested-by: Arnaud Vrac <rawoul@gmail.com>
This test checks that the protocol and library headers include only what
they are supposed to include. That is, that the core headers do not
include the protocol headers and that the core protocol headers do not
include the non core library headers.
The build process now generates core protocol headers, but they are only
used in the test and don't get installed.
Reviewed-by: Bryce Harrington <bryce@osg.samsung.com>
When using this new option the generated code will include the new
core headers instead of the old ones. The default needs to remain
unchanged for backward compatibility with old code.
With this change the generated headers will now forward declare all
types and interfaces it uses; that is needed when generating headers
for a my-extension.xml with --include-core-only, since it may use
types defined in wayland.xml.
The same is done also without --include-core-only, since it is an
harmless change.
getopt_long() is used for the option handling.
Reviewed-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>
The new core header doesn't include any other header, since it really
is not needed.
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Reviewed-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>
wayland-client.h and wayland-server.h include the protocol headers generated
at build time. This means that a libwayland user cannot generate and use
protocol code created from a wayland.xml newer than the installed libwayland,
because it is not possible to only include the API header.
Another use case is language bindings, which would generate their own protocol
code and which only need to use the library ABI, not the generated C code.
This commit adds wayland-client-core.h and wayland-server-core.h which do not
include the protocol headers or any deprecated code.
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Reviewed-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>
Without this patch, the scanner would generate invalid C which wouldn't
compile anyway, so lets be nice and fail earlier and point out where the
error is.
Signed-off-by: Jonas Ådahl <jadahl@gmail.com>
Reviewed-by: Derek Foreman <derekf@osg.samsung.com>
Reviewed-by: David Fort <contact@hardening-consulting.com>
Reviewed-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>
I wasted a lot of time before I figured out that I needed to add those
square brackets to get this to work. Sigh...
Reviewed-by: Bryce Harrington <bryce@osg.samsung.com>
Add note about what all wl_pointer.release does. Mainly that
it destroys the proxy object, so programmer must not call
wl_pointer_destroy() on the pointer any further.
Signed-off-by: Marek Chalupa <mchqwerty@gmail.com>
Reviewed-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>
Add support for direct file reading and writing in wayland-scanner.
Signed-off-by: Jussi Pakkanen <jpakkane@gmail.com>
Reviewed-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>
Acked-by: David Fort <rdp.effort@gmail.com>
There are two same error messages with different cause.
Let user know what is the cause of the error.
Signed-off-by: Marek Chalupa <mchqwerty@gmail.com>
Reviewed-by: Daniel Stone <daniels@collabora.com>
Reviewed-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>
It's useful to know how long the current cursor frame should be displayed
so we can wait that long to change it.
Signed-off-by: Derek Foreman <derekf@osg.samsung.com>
Reviewed-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>
Change wording to be more consistent with other parts of the subsurface
protocol. Before this change, wl_subsurface.set_position explicitly
stated that the new state was to be applied on the parents
wl_surface.commit and wl_subsurface.place_above/below only said "on
the next commit of the parent surface". What "committed" means is
ambiguous considering that a wl_surface.commit actually defers the
actual commit when in synchronized mode, but the intention has always
been that placement of a subsurface should be considered part of its
content, i.e. placement state should be applied when other state
(buffer, regions). This patch makes that more clear.
Note that prior to this patch, one could correctly have interpreted
the protocol meaning that placements operations takes effect explicitly
on wl_surface.commit of the parent surface no matter whether other state
of the parent surface is applied at that point. This patch clarifies that
that is not the case.
https://bugs.freedesktop.org/show_bug.cgi?id=88857
Signed-off-by: Jonas Ådahl <jadahl@gmail.com>
Acked-by: Jasper St. Pierre <jstpierre@mecheye.net>
Reviewed-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>
This paritally reverts commit a4afd90f9f.
On older expat versions (ie the one on Ubuntu 12.04) there is no pkg-config
file, so fall back to a test for the header and library. In addition the
source for expat does not seem to be in a git repository but in cvs instead
and it seems preferrable to not require cvs to build wayland.
The restored test has been updated to use AC_SEARCH_LIBS. This version
uses empty square brackets for the unused branches, similar to many other
if statements in configure.ac.
Reviewed-by: Bryce Harrington <bryce@osg.samsung.com>
Reviewed-by: Andrew Oakley <aoakley@espial.com>
This fixes a regression in the testsuite since c3653f7f, where four of
the timeout tests fail with "Timeouts suppressed" messages.
The timeouts are being suppressed because the testsuite is erroneously
detecting that a debugger is attached. This detection mechanism
(adopted from libinput) uses ptrace to test if there is a debugger
parent process that can be attached. Unfortunately, this is an
unreliable test: Kernel security policies exist to restrict the scope of
ptrace to prevent processes from snooping on one another.[1] This
security policy is set as the default on Ubuntu, and potentially other
Linux distributions.[2]
The Yama documentation suggests, "For software that has defined
application-specific relationships between a debugging process and its
inferior (crash handlers, etc), prctl(PR_SET_PTRACER, pid, ...) can be
used. An inferior can declare which other process (and its descendents)
are allowed to call PTRACE_ATTACH against it." This prctl call has no
effect if Yama LSM is not loaded.
The child needs to be synchronized to the client to prevent a race
condition where the child might try to operate before the parent has
finished its prctl call. This synchronization is done via pipes.
This patch can be tested by running sanity-test with
/proc/sys/kernel/yama/ptrace_scope set to 0 or 1; the test must pass for
either value.
1: http://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=2d514487faf188938a4ee4fb3464eeecfbdcf8eb
2: https://wiki.ubuntu.com/SecurityTeam/Roadmap/KernelHardening#ptrace_Protection
Signed-off-by: Bryce Harrington <bryce@osg.samsung.com>
Reviewed-by: Marek Chalupa <mchqwerty@gmail.com>
Reviewed-by: Daniel Stone <daniels@collabora.com>
Reviewed-by: Derek Foreman <derekf@osg.samsung.com>
v4: Allow parent to communicate error state to child to prevent leaving
child in zombie state if parent hits an error.
v5: Check errno instead of rc for error. Don't waitpid on ppid.
Signed-off-by: Bryce Harrington <bryce@osg.samsung.com>
After running a test in fork, we were waiting for any child to terminate.
It is OK unless the child forks again. If the child calls fork, the waitid can
catch the child's child termination, stop block and run another test
while the former test is still running. This is racy i. e. when adding socket.
Since we have test compositor which uses fork, this situation can occur
pretty frequently.
Signed-off-by: Marek Chalupa <mchqwerty@gmail.com>
Reviewed-by: Jon A. Cruz <jonc@osg.samsung.com>
Reviewed-by: Daniel Stone <daniels@collabora.com>
Corrects an issue that would cause out-of-tree builds to fail and also
a few items that would cause distcheck to fail.
Signed-off-by: Jon A. Cruz <jonc@osg.samsung.com>
Reviewed-by: Bryce Harrington <bryce@osg.samsung.com>
(This patch has been modified to apply atop current master)
This makes it considerably easier to edit the text and make it different
for each library.
To address previous concerns with this patch, I wrote some more complete
introductory text. This is based on my understanding of these libraries, which
may not be correct, and is pretty rudimentary for libwayland-server!
However this intro text demonstrates how to create links to the
doxygen-generated text. It looks like you cannot link to methods easily as the
link name contains a hash number, but links to objects and classes work.
Reviewed-by: Jon A. Cruz <jonc@osg.samsung.com>
Tested-by: Jon A. Cruz <jonc@osg.samsung.com>
This is now done in the same way as the libffi dependency and still
allows the library to be installed in a non-standard location (with
PKG_CONFIG_PATH).
Otherwise a parallel make invocation could fail due to the directory
not existing.
Signed-off-by: Rui Matos <tiagomatos@gmail.com>
Reviewed-by: Jon A. Cruz <jonc@osg.samsung.com>
To fix a shutdown crash in weston's x11 compositor I want to move the
weston X window close to an idle handler.
Since idle handlers are processed at the start of an event loop, the
handler that deals with window close will run at the start of the
next input_loop dispatch, after which the dispatcher blocks on epoll
forever (since all input events that will ever occur have been consumed).
Dispatching idle callbacks both at the start and end of event-loop
processing will prevent this permanent blocking.
Note that just moving the callback dispatch could theoretically
result in an idle callback being delayed indefinitely while waiting
for epoll_wait() to complete.
Callbacks are removed from the list when they're run, so the second
dispatch won't result in any extra calls.
Signed-off-by: Derek Foreman <derekf@osg.samsung.com>
Reviewed-by: Giulio Camuffo <giuliocamuffo@gmail.com>
Reviewed-by: Bryce Harrington <bryce@osg.samsung.com>
Added xslt processing to give DocBook output diagram image maps/hot-linked
areas consistent with those automatically generated by Doxygen.
Signed-off-by: Jon A. Cruz <jonc@osg.samsung.com>
Switches diagrams from using static PNG images to instead generate them via
simple graphviz DOT markup files.
Signed-off-by: Jon A. Cruz <jonc@osg.samsung.com>
In tests that are using external libraries (i. e. pthread) we
can get failure because of leaks in the external library.
Until we have some better solution (if ever), let these (and only these)
tests to disable leak checks.
Signed-off-by: Marek Chalupa <mchqwerty@gmail.com>
Reviewed-by: Daniel Stone <daniels@collabora.com>
When we release event queue with queued events, we can leak
proxies in some cases.
Signed-off-by: Marek Chalupa <mchqwerty@gmail.com>
Reviewed-by: Daniel Stone <daniels@collabora.com>
When we destroy closure, we are leaking memory sometimes.
Signed-off-by: Marek Chalupa <mchqwerty@gmail.com>
Reviewed-by: Daniel Stone <daniels@collabora.com>
