John Kåre Alsaker [Fri, 12 Oct 2012 09:28:26 +0000 (11:28 +0200)]
connection: Removed crashing code.
John Kåre Alsaker [Fri, 12 Oct 2012 09:28:25 +0000 (11:28 +0200)]
scanner: Make sure arguments have names.
John Kåre Alsaker [Fri, 12 Oct 2012 09:28:24 +0000 (11:28 +0200)]
connection: Don't call memcpy with null pointer.
Pekka Paalanen [Fri, 12 Oct 2012 11:29:01 +0000 (14:29 +0300)]
protocol: try to clarify wl_buffer doc
Fix few typos in wl_buffer description.
Mention backing storage in wl_buffer.destroy.
Try to clarify the wl_buffer.release semantics by not explaining what
*might* happen. It is important to not suggest, that if release does not
come before frame callback, it will not come before attaching a new
buffer to the surface. We want to allow the following scenario:
The compositor is able to texture from wl_buffers directly, but it also
keeps a copy of the surface contents. The copy is updated when the
compositor is idle, to avoid the performance hit on
wl_surface.attach/commit. When the copy completes some time later, the
server sends the release event. If the client has not yet allocated a
second buffer (e.g. it updates rarely), it can reuse the old buffer.
Reported-by: John Kåre Alsaker <john.kare.alsaker@gmail.com>
Signed-off-by: Pekka Paalanen <ppaalanen@gmail.com>
Pekka Paalanen [Fri, 12 Oct 2012 11:29:00 +0000 (14:29 +0300)]
protocol: say update instead of repaint
wl_surface.commit itself does not force any repainting unless there is
damage, so change the wording to not imply repainting.
Reported-by: John Kåre Alsaker <john.kare.alsaker@gmail.com>
Signed-off-by: Pekka Paalanen <ppaalanen@gmail.com>
Ander Conselvan de Oliveira [Tue, 16 Oct 2012 14:29:09 +0000 (17:29 +0300)]
doc: Add doxygen documentation to wl_display_get_error()
Ander Conselvan de Oliveira [Tue, 16 Oct 2012 14:29:08 +0000 (17:29 +0300)]
doc: Document change of return value of dispatch functions
Ander Conselvan de Oliveira [Tue, 16 Oct 2012 14:29:07 +0000 (17:29 +0300)]
doc: Document the restriction of destroying queues before the display
Ander Conselvan de Oliveira [Tue, 16 Oct 2012 14:29:06 +0000 (17:29 +0300)]
doc: Put wl_display_flush() documentation in the right place
It seems a rebase error caused it to end up in the wrong place.
Jonas Ådahl [Wed, 3 Oct 2012 20:50:55 +0000 (22:50 +0200)]
protocol: Clarify pointer axis event
Pointer axis events are in the same coordinate space as motion events.
Signed-off-by: Jonas Ådahl <jadahl@gmail.com>
Kristian Høgsberg [Tue, 16 Oct 2012 00:49:41 +0000 (20:49 -0400)]
configure.ac: Bump version to 0.99.0
Kristian Høgsberg [Tue, 16 Oct 2012 00:48:12 +0000 (20:48 -0400)]
docs: Make distcheck happy
publican doesn't seem to be a very well-behaved tool and needs a bit
of hacking to work well for both in-tree and out-of-tree builds.
Matt Roper [Wed, 10 Oct 2012 23:56:52 +0000 (16:56 -0700)]
wayland-server: Add touch grab support
Touch grabs allow the compositor to be placed into a mode where touch events
temporarily bypass their default behavior and perform other operations.
Wayland already supports keyboard and pointer grabs, but was lacking
corresponding touch support. The default touch grab handlers here contain the
client event delivery code that was previously called directly in weston.
Signed-off-by: Matt Roper <matthew.d.roper@intel.com>
David Herrmann [Thu, 11 Oct 2012 21:37:53 +0000 (23:37 +0200)]
wayland-client: forward fatal errors to caller
If any callback or helper function fails with a fatal error, we now
set the last_error flag and prevent all further I/O on the wl_display. We
wake up all sleeping event-queues and notify the caller that they
should shutdown wl_display.
Signed-off-by: David Herrmann <dh.herrmann@googlemail.com>
Kristian Høgsberg [Mon, 15 Oct 2012 21:50:36 +0000 (17:50 -0400)]
wayland: Take ownership of fd in wl_display_connect_to_fd()
This means we're free to close it when we want, which we'll use to wake
up the main thread if we hit an error in a different thread.
David Herrmann [Thu, 11 Oct 2012 21:37:51 +0000 (23:37 +0200)]
wayland-client: link all event-queues of each display into a list
We need access to all event-queues of a single wl_display object. For
instance during connection-errors, we need to be able to wake up all event
queues. Otherwise, they will be stuck waiting for incoming events.
The API user is responsible to keep a wl_display object around until all
event-queues that were created on it are destroyed.
Signed-off-by: David Herrmann <dh.herrmann@googlemail.com>
Kristian Høgsberg [Mon, 15 Oct 2012 21:19:38 +0000 (17:19 -0400)]
connection: return error on buffer-overflow during read
wl_connection_read() assumes that the caller dispatched all messages
before calling it. wl_buffer_put_iov() does only provide enough room so we
fill the buffer. So the only case when the buffer overflows, is when a
previous read filled up the buffer but we couldn't parse a single message
from it. In this case, the client sent a message bigger than our buffer
and we should return an error and close the connection.
krh: Edited from Davids original patch to just check that the buffer
isn't full before we try reading into it.
Signed-off-by: David Herrmann <dh.herrmann@googlemail.com>
Kristian Høgsberg [Mon, 15 Oct 2012 21:16:30 +0000 (17:16 -0400)]
connection: Use uin32_t for circular buffer indexes
We rely on well-defined unsigned overflow behaviour so let's make the
index fields actually unsigned. Signed ints aren't guaranteed to have the
behavior we want (could be either ones or twos complement).
David Herrmann [Thu, 11 Oct 2012 21:37:49 +0000 (23:37 +0200)]
connection: fix leaking FDs on buffer-overflow during read
If we read more FDs than we have room for, we currently leak FDs because
we overwrite previous still pending FDs. Instead, we do now close incoming
FDs if the buffer is full and return EOVERFLOW.
Signed-off-by: David Herrmann <dh.herrmann@googlemail.com>
David Herrmann [Thu, 11 Oct 2012 21:37:48 +0000 (23:37 +0200)]
connection: fix buffer-overflow in build_cmsg()
Same problem as we had with close_fds(). We cannot rely on the fds_out
buffer being filled with less than MAX_FDS_OUT file descriptors.
Therefore, write at most MAX_FDS_OUT file-descriptors to the outgoing
buffer.
Signed-off-by: David Herrmann <dh.herrmann@googlemail.com>
David Herrmann [Thu, 11 Oct 2012 21:37:47 +0000 (23:37 +0200)]
connection: close pending incoming FDs on shutdown
Same problem as with outgoing FDs. We need to close these on shutdown,
otherwise we leak open file descriptors.
Signed-off-by: David Herrmann <dh.herrmann@googlemail.com>
David Herrmann [Thu, 11 Oct 2012 21:37:46 +0000 (23:37 +0200)]
connection: fix buffer-overflow in close_fds()
If we push two messages via wl_connection_write() and both messages
contain more than MAX_FDS_OUT file-descriptors combined, then
wl_connection_flush() will write only MAX_FDS_OUT of them, but close all
pending ones, too.
Furthermore, close_fds() will copy more FDs out of the buffer than it can
hold and cause a buffer overflow. Therefore, we simply pass a maximum
limit to close_fds().
During shutdown, we simply close all available FDs.
Signed-off-by: David Herrmann <dh.herrmann@googlemail.com>
David Herrmann [Thu, 11 Oct 2012 21:37:45 +0000 (23:37 +0200)]
connection: close pending outgoing fds on shutdown
When destroying a wl_connection object, there might still be data in the
queue. We would leak open file-descriptors so we must close them.
Signed-off-by: David Herrmann <dh.herrmann@googlemail.com>
David Herrmann [Thu, 11 Oct 2012 21:37:44 +0000 (23:37 +0200)]
wayland-util: return 0 on OOM in wl_map_insert_new()
If we cannot increase the array for new entries, we now return 0 instead
of accessing invalid memory.
krh: Edited to return 0 on failure instead. In the initialization path,
we call wl_map_insert_new() to insert NULL at index 0, which also returns
0 but not as an error. Since we do that up front, every other case of
returning 0 is an unambiguous error.
Signed-off-by: David Herrmann <dh.herrmann@googlemail.com>
David Herrmann [Thu, 11 Oct 2012 21:37:43 +0000 (23:37 +0200)]
wayland-util: return -1 if wl_array_copy() fails
We might have to perform memory allocations in wl_array_copy(), so catch
out-of-memory errors in wl_array_add() and return -1 before changing any
state.
Signed-off-by: David Herrmann <dh.herrmann@googlemail.com>
David Herrmann [Thu, 11 Oct 2012 21:37:42 +0000 (23:37 +0200)]
wayland-client: add wl_display_get_error()
A server may asynchronously send errors via wl_display.error() events.
Instead of aborting we now the a "last_error" flag inside of wl_display
objects. The user can retrieve these via wl_display_get_error().
Signed-off-by: David Herrmann <dh.herrmann@googlemail.com>
Jonas Ådahl [Thu, 27 Sep 2012 16:40:26 +0000 (18:40 +0200)]
protocol: Clarify pointer axis event
Pointer axis events are in the same coordinate space as motion events.
Signed-off-by: Jonas Ådahl <jadahl@gmail.com>
Tiago Vignatti [Fri, 12 Oct 2012 17:54:45 +0000 (20:54 +0300)]
configure: Make documentation option work in fact
Also, now doxygen is mandatory for building the documentation (looks
reasonable because both man-pages and publican will need it).
Signed-off-by: Tiago Vignatti <tiago.vignatti@intel.com>
Tiago Vignatti [Fri, 12 Oct 2012 17:03:44 +0000 (20:03 +0300)]
doc: doxygen: Add .gitignore
Signed-off-by: Tiago Vignatti <tiago.vignatti@intel.com>
Ander Conselvan de Oliveira [Mon, 15 Oct 2012 14:53:23 +0000 (17:53 +0300)]
doc: Improve libwayland-client doxygen documentation
Document wl_proxy, wl_display and wl_event_queue classes and add a
description to all public entry points. Also fix some typos.
Ander Conselvan de Oliveira [Fri, 12 Oct 2012 14:28:57 +0000 (17:28 +0300)]
doc: Add some doxygen documentation to wayland-client entry points
Add some brief documentation for the public libwayland-client entry
points. This is by no means complete, some functions are still
undocumented and some might need extra information.
Signed-off-by: Ander Conselvan de Oliveira <ander.conselvan.de.oliveira@intel.com>
Signed-off-by: Tiago Vignatti <tiago.vignatti@intel.com>
Ander Conselvan de Oliveira [Fri, 12 Oct 2012 14:28:56 +0000 (17:28 +0300)]
doc: Split protocol description paragraphs properly
The xsl translation from the protocol xml to publican would create only
one paragraph for all the text in a description. Make it generate one
paragraph for each block of text separated by two consecutive line
breaks instead.
Signed-off-by: Ander Conselvan de Oliveira <ander.conselvan.de.oliveira@intel.com>
Signed-off-by: Tiago Vignatti <tiago.vignatti@intel.com>
Olivier Blin [Sun, 14 Oct 2012 22:49:13 +0000 (00:49 +0200)]
scanner: use printf format attributes for desc_dump
Olivier Blin [Sun, 14 Oct 2012 22:49:11 +0000 (00:49 +0200)]
scanner: fix writing i586 descriptions
This moves desc as first argument of desc_dump().
Description writing was broken on i586 because desc_dump() used
va_arg() after a vsnprintf() call to find the last argument.
But after calling a function with a va_arg argument, this arguments is
undefined.
Olivier Blin [Sun, 14 Oct 2012 22:49:11 +0000 (00:49 +0200)]
scanner: remove useless desc_dump arguments
This are remnant from the desc_dump generalization in commit
375cb418.
Tiago Vignatti [Thu, 11 Oct 2012 14:42:09 +0000 (17:42 +0300)]
doc: Add auto-generated Wayland Library chapter
For now only Wayland Client API is described on that chapter, which is
extracted via doxygen on ./src/wayland-client.h. We apply a stylesheet
(doxygen-to-publican) on doxygen output so it becomes docbook valid.
Now all we need to do is populate that header while developing in order to
grow a decent documentation. So please use it!
Signed-off-by: Tiago Vignatti <tiago.vignatti@intel.com>
Tiago Vignatti [Thu, 11 Oct 2012 14:42:08 +0000 (17:42 +0300)]
doc: Remove superfluous 'index'
We're not setting any sort of index. Remove for now.
Signed-off-by: Tiago Vignatti <tiago.vignatti@intel.com>
Tiago Vignatti [Thu, 11 Oct 2012 14:42:07 +0000 (17:42 +0300)]
doc: publican: Automate version generation
It seems reasonable to use protocol's version for the documentation as well.
Signed-off-by: Tiago Vignatti <tiago.vignatti@intel.com>
Tiago Vignatti [Thu, 11 Oct 2012 14:42:06 +0000 (17:42 +0300)]
doc: publican: Set table of contents depth to 1
This way looks more pretty, in particular for the Appendix which spawns a big
subsections chain.
Signed-off-by: Tiago Vignatti <tiago.vignatti@intel.com>
Kristian Høgsberg [Mon, 15 Oct 2012 15:38:20 +0000 (11:38 -0400)]
client: Return number of events dispatched from dispatch functions
To let clients determine whether any events were dispatched, we return
the number of dispatched events. An event source with an event queue
(such as wl_display or an X connection) may queue up event as a result of
processing a different event source (data on a network socket, timerfd etc).
After dispatching data from fd (or just before blocking) we have to check
such event sources, which is what wl_event_source_check() is used for.
A checked event source will have its handler called with mask=0 just
before blocking. If any work is done in any of these handlers, we have
to check all the checked sources again, since the work could have queued up
events in a different source. This is why the event handlers must return
a positive number if events were handled. Which in turn is why we need
the wl_display dispatch functions to return that as well.
Kristian Høgsberg [Thu, 11 Oct 2012 21:15:08 +0000 (17:15 -0400)]
client: Add wl_display_dispatch_pending() for dispatching without reading
If the main thread ends up dispatching a non-main queue, and not in
a wl_display_dispatch() callback, we may queue up main queue events and read
all data from the socket fd. When we get back to the main loop, the
socket fd is no longer readable and nothing will trigger dispatching of
the queued up events.
The new function wl_display_dispatch_pending() will dispatch any pending
events, but not attempt to read from the socket. Clients that integrate
the wayland socket fd into a main loop should call
wl_display_dispatch_pending() and then wl_display_flush()
before going back to blocking in poll(2) or similar mechanism.
Kristian Høgsberg [Thu, 11 Oct 2012 21:12:50 +0000 (17:12 -0400)]
client: Discard proxies with no implementation at dispatch time
We need to queue up events even if a proxy doesn't have an implementation
(listener). In case of server created new objects, the client haven't
had a chance to set the listener when the first events to the new object
come in. So now we always queue up events and discard them at
dispatch time if they don't have a listener at that point.
Kristian Høgsberg [Thu, 11 Oct 2012 21:11:54 +0000 (17:11 -0400)]
client: Don't forget to init and destroy mutex
These chunks were dropped at some point, thanks to David Herrmann for
spotting the omission.
Kristian Høgsberg [Thu, 11 Oct 2012 21:08:29 +0000 (17:08 -0400)]
connection: Print object id for new-id arguments in deubug output
We can't use the same behaviour in both the client and the server. In the
client this is a wl_proxy pointer in the server it's a pointer to the
uint32_t object id. This doesn't fix the problem, but it's a slightly
more useful default, since we typically use WAYLAND_DEBUG on the client.
Ander Conselvan de Oliveira [Thu, 11 Oct 2012 11:55:59 +0000 (14:55 +0300)]
client: Fix double locking bug
The function wl_proxy_create_for_id() would try to acquire the display
lock, but the only call path leading to it would call it with the lock
already acquired.
This patch removes the attempt to acquire the lock and makes the
function static. It was exported before because client had to create
proxy's manually when the server sent a new object id, but since commit
9de9e39f [1] this is no longer necessary.
[1] commit
9de9e39f87adfce1ea9755e394928756254c0ba2
Author: Kristian Høgsberg <krh@bitplanet.net>
Date: Thu Jun 28 22:01:58 2012 -0400
Allocate client proxy automatically for new objects
v2: Change the right function. Previous patch changed wl_proxy_create()
instead of wl_proxy_create_for_id().
Pekka Paalanen [Thu, 11 Oct 2012 11:31:11 +0000 (14:31 +0300)]
protocol: clarify multiple wl_surface.attach
Explicitly say what happens with the wl_buffer.release event, if you
attach several wl_buffers without a commit in between.
Reported-by: David Herrmann <dh.herrmann@googlemail.com>
Signed-off-by: Pekka Paalanen <ppaalanen@gmail.com>
Ander Conselvan de Oliveira [Thu, 11 Oct 2012 11:05:37 +0000 (14:05 +0300)]
data-device: Don't fake an attach event on drag icon surface
Emit a new drag icon signal instead and let the compositor handle the
unmapping of the icon surface.
Pekka Paalanen [Thu, 11 Oct 2012 10:58:45 +0000 (13:58 +0300)]
protocol: fix clarification of input region on drags and pointers
The previous clarification did not follow the current implementation in
Weston, where when a surface stops being a cursor or an icon, it becomes
a plain unmapped surface again.
Rewrite the related paragraphs, and fix some typos while at it.
For start drag, make it explicit of which surface argument we are
talking about.
v2:
Make the input region undefined when the use ends. Most likely no-one
will re-use these surfaces for anything else than the same use case, so
leave some slack for the implementations to avoid useless work on
resetting the regions.
Reported-by: Ander Conselvan de Oliveira <conselvan2@gmail.com>
Signed-off-by: Pekka Paalanen <ppaalanen@gmail.com>
Pekka Paalanen [Wed, 10 Oct 2012 09:47:50 +0000 (12:47 +0300)]
protocol: elaborate on wl_buffer
Spell out exactly when a client may re-use a wl_buffer or its backing
storage. Mention the optimization for GL-compositor with wl_shm-clients.
Signed-off-by: Pekka Paalanen <ppaalanen@gmail.com>
Pekka Paalanen [Wed, 10 Oct 2012 09:47:49 +0000 (12:47 +0300)]
protocol: wl_surface.frame needs wl_surface.commit
Clarify, when frame request takes effect.
Explain when to send/receive the callback.
Signed-off-by: Pekka Paalanen <ppaalanen@gmail.com>
Pekka Paalanen [Wed, 10 Oct 2012 09:47:48 +0000 (12:47 +0300)]
protocol: clarify input region on drags and pointers
Drag icon and cursor surfaces must never receive input, so their input
region is always empty.
Signed-off-by: Pekka Paalanen <ppaalanen@gmail.com>
Pekka Paalanen [Wed, 10 Oct 2012 09:47:47 +0000 (12:47 +0300)]
protocol: double-buffered state for wl_surface
This change breaks the protocol.
The current protocol is racy in that updates to surface content and
surface state (e.g. damage, input and opaque regions) are not guaranteed
to happen at the same time. Due to protocol buffering and handling
practices, the issues are very hard to trigger.
Committing damage to a surface at arbitrary times makes it hard to
track when the wl_buffer is being read by the server, and when it is
safe to overwrite (the case of wl_shm with a single buffer reused
constantly).
This protocol change introduces the concept of double-buffered state.
Such state is accumulated and cached in the server, unused, until the
final commit request. The surface will receive its new content and apply
its new state atomically.
A wl_surface.commit request is added to the protocol. This is thought to
be more clear, than having wl_surface.attach committing implicitly, and
then having another request to commit without attaching, as would be
required for a GL app that wants to change e.g. input region without
redrawing.
When these changes are implemented, clients do not have to worry about
ordering damage vs. input region vs. attach vs. ... anymore. Clients set
the state in any order they want, and kick it all in with a commit.
The interactions between wl_surface.attach, (wl_surface.commit,)
wl_buffer.release, and wl_buffer.destroy have been undocumented. Only
careful inspection of the compositor code has told when a wl_buffer is
free for re-use, especially for wl_shm and wrt. wl_surface.damage.
Try to clarify how it all should work, and what happens if the wl_buffer
gets destroyed.
An additional minor fix: allow NULL argument to
wl_surface.set_opaque_region. The wording in the documentation already
implied that a nil region is allowed.
Signed-off-by: Pekka Paalanen <ppaalanen@gmail.com>
Kristian Høgsberg [Wed, 10 Oct 2012 23:25:48 +0000 (19:25 -0400)]
Fix typecheck in case of multiple instances of type meta data
In most cases the pointer equality test is sufficient. However, in
some cases, depending on how things are split across shared objects,
we can end up with multiple instances of the interface metadata
constants. So if the pointers match, the interfaces are equal, if
they don't match we have to compare the interface names.
Ander Conselvan de Oliveira [Wed, 10 Oct 2012 20:29:13 +0000 (23:29 +0300)]
doc: Update drag and drop section and add info about selections
Replace the outdated section about drag and drop support with a
rewritten section covering the data source/offer mechanism and
wl_data_device, explaining how selection and drag ang drop works.
Kristian Høgsberg [Tue, 9 Oct 2012 16:14:34 +0000 (12:14 -0400)]
connection: Move object lookup out of wl_connection_demarshal()
On the client side where we queue up multiple events before dispatching, we
need to look up the receiving proxy and argument proxies immediately before
calling the handler. Between queueing up multiple events and eventually
invoking the handler, previous handlers may have destroyed some of the
proxies.
Kristian Høgsberg [Mon, 8 Oct 2012 17:53:47 +0000 (13:53 -0400)]
Split the global registry into its own wl_registry object
The only way to make the global object listener interface thread safe is to
make it its own interface and make different listeners different wl_proxies.
The core of the problem is the callback we do when a global show up or
disappears, which we can't do with a lock held. On the other hand we can't
iterate the global list or the listener list without a lock held as new
globals or listeners may come and go during the iteration.
Making a copy of the list under the lock and then iterating after dropping
the lock wont work either. In case of the listener list, once we drop the
lock another thread may unregister a listener and destroy the callbackk
data, which means that when we eventually call that listener we'll pass it
free memory and break everything.
We did already solve the thread-safe callback problem, however. It's what
we do for all protocol events. So we can just make the global registry
functionality its own new interface and give each thread its own proxy.
That way, the thread will do its own callbacks (with no locks held) and
destroy the proxy when it's no longer interested in wl_registry events.
Kristian Høgsberg [Mon, 8 Oct 2012 17:25:52 +0000 (13:25 -0400)]
scanner: Generate client stubs for wl_display requests
We used to special case this because of the untyped new-id argument in
the bind request. Now that the scanner can handle that, we can
remove the special case.
Switching to the generated stubs does bring an API change since we now
also take the interface version that the client expects as an argument.
Previously we would take this from the interface struct, but the
application may implement a lower version than what the interface struct
provides. To make sure we don't try to dispatch event the client
doesn't implement handlers for, we have to use a client supplied version
number.
Kristian Høgsberg [Mon, 8 Oct 2012 17:21:55 +0000 (13:21 -0400)]
scanner: Send interface name and version for types new_id args
This makes the scanner generate the code and meta data to send the
interface name and version when we pass a typeless new_id. This way, the
generic factory mechanism provided by wl_display.bind can be provided by
any interface.
Kristian Høgsberg [Fri, 5 Oct 2012 17:49:48 +0000 (13:49 -0400)]
client: Add wl_event_queue for multi-thread dispatching
This introduces wl_event_queue, which is what will make multi-threaded
wayland clients possible and useful. The driving use case is that of a
GL rendering thread that renders and calls eglSwapBuffer independently of
a "main thread" that owns the wl_display and handles input events and
everything else. In general, the EGL and GL APIs have a threading model
that requires the wayland client library to be usable from several threads.
Finally, the current callback model gets into trouble even in a single
threaded scenario: if we have to block in eglSwapBuffers, we may end up
doing unrelated callbacks from within EGL.
The wl_event_queue mechanism lets the application (or middleware such as
EGL or toolkits) assign a proxy to an event queue. Only events from objects
associated with the queue will be put in the queue, and conversely,
events from objects associated with the queue will not be queue up anywhere
else. The wl_display struct has a built-in event queue, which is considered
the main and default event queue. New proxies are associated with the
same queue as the object that created them (either the object that a
request with a new-id argument was sent to or the object that sent an
event with a new-id argument). A proxy can be moved to a different event
queue by calling wl_proxy_set_queue().
A subsystem, such as EGL, will then create its own event queue and associate
the objects it expects to receive events from with that queue. If EGL
needs to block and wait for a certain event, it can keep dispatching event
from its queue until that events comes in. This wont call out to unrelated
code with an EGL lock held. Similarly, we don't risk the main thread
handling an event from an EGL object and then calling into EGL from a
different thread without the lock held.
Kristian Høgsberg [Thu, 4 Oct 2012 21:42:49 +0000 (17:42 -0400)]
client: Make wl_display thread safe
Not all entry points are thread safe: global listeners and global lookup
is still only main thread.
Kristian Høgsberg [Thu, 4 Oct 2012 21:34:18 +0000 (17:34 -0400)]
client: Split event handling into demarshal and dispatch steps
This lets us demarshal with a mutex held and then do dispatching after
releasing the mutex.
Kristian Høgsberg [Thu, 4 Oct 2012 20:54:22 +0000 (16:54 -0400)]
Change filedescriptor API to be thread safe
The update callback for the file descriptors was always a bit awkward and
un-intuitive. The idea was that whenever the protocol code needed to
write data to the fd it would call the 'update' function. This function
would adjust the mainloop so that it polls for POLLOUT on the fd so we
can eventually flush the data to the socket.
The problem is that in multi-threaded applications, any thread can issue
a request, which writes data to the output buffer and thus triggers the
update callback. Thus, we'll be calling out with the display mutex
held and may call from any thread.
The solution is to eliminate the udpate callback and just require that
the application or server flushes all connection buffers before blocking.
This turns out to be a simpler API, although we now require clients to
deal with EAGAIN and non-blocking writes. It also saves a few syscalls,
since the socket will be writable most of the time and most writes will
complete, so we avoid changing epoll to poll for POLLOUT, then write and
then change it back for each write.
Matt Roper [Thu, 4 Oct 2012 21:27:39 +0000 (14:27 -0700)]
Ensure cursor_data.c is included in distribution tarballs
Signed-off-by: Matt Roper <matthew.d.roper@intel.com>
Rob Bradford [Tue, 9 Oct 2012 17:46:22 +0000 (18:46 +0100)]
connection: Add missing free from error path
On the error codepath that errors out on ENOMEM we should free the allocated
closure.
Signed-off-by: Rob Bradford <rob@linux.intel.com>
Rob Bradford [Tue, 9 Oct 2012 17:46:21 +0000 (18:46 +0100)]
xcursor: Fix allocation based on string length
strlen() doesn't include the terminating NUL. Therefore when allocating a
block of memory to hold something equivalent to the length of the string we
must increment to take the NUL byte into consideration.
Signed-off-by: Rob Bradford <rob@linux.intel.com>
Tiago Vignatti [Mon, 8 Oct 2012 15:40:00 +0000 (18:40 +0300)]
wayland: Fix typos
My vim spell checker is able to find typos of xml files after adding "syn spell
toplevel" to ~/.vim/after/syntax/xml.vim
aah, and Wayland is capital letter :)
Signed-off-by: Tiago Vignatti <tiago.vignatti@intel.com>
Tiago Vignatti [Mon, 8 Oct 2012 15:39:59 +0000 (18:39 +0300)]
doc: Remove Shared Object Cache section
We don't support anything like that yet.
Signed-off-by: Tiago Vignatti <tiago.vignatti@intel.com>
Tiago Vignatti [Mon, 8 Oct 2012 15:39:58 +0000 (18:39 +0300)]
doc: Auto-generate Protocol/Interfaces section instead
Signed-off-by: Tiago Vignatti <tiago.vignatti@intel.com>
Tiago Vignatti [Mon, 8 Oct 2012 15:39:57 +0000 (18:39 +0300)]
doc: Improve Wire Format section
Fixed the wayland socket name and added documentation for fixed format.
Signed-off-by: Tiago Vignatti <tiago.vignatti@intel.com>
Kristian Høgsberg [Mon, 1 Oct 2012 17:21:29 +0000 (13:21 -0400)]
connection: Drop unused static closures
David Herrmann [Wed, 26 Sep 2012 14:41:50 +0000 (16:41 +0200)]
man: fix compilation without xsltproc
We really shouldn't add the man-pages when HAVE_XSLTPROC is not true so
move it into the if-clause.
But declare the automake-variables outside of the if-clause to avoid
automake complaints.
Signed-off-by: David Herrmann <dh.herrmann@googlemail.com>
David Herrmann [Mon, 24 Sep 2012 11:18:38 +0000 (13:18 +0200)]
man: add man-page infrastructure
This adds a man-page infrastructure based on Docbook XML files. This
allows us to integrate the man-pages into the publican books later. An
example page for wl_display_connect() (with an alias
wl_display_connect_to_fd()) is also added.
Feel free to add more man-pages. Function calls are put in man3 and
overview pages into man7. All pages (including aliases) have to be added
to the Makefile.
Docbook does generate aliases automatically from the additional names that
were put in the XML file. However, a small SED script is needed to fixup
the include-paths in the generated troff files. If someone knows how to
avoid that (or even install them gzip'ped), please fix it up.
Signed-off-by: David Herrmann <dh.herrmann@googlemail.com>
Tiago Vignatti [Tue, 25 Sep 2012 09:19:57 +0000 (12:19 +0300)]
protocol: Fix typo
Signed-off-by: Tiago Vignatti <tiago.vignatti@intel.com>
Diego Viola [Thu, 13 Sep 2012 10:45:06 +0000 (06:45 -0400)]
Fix typo (ratio, not ration)
Signed-off-by: Diego Viola <diego.viola@gmail.com>
David Herrmann [Tue, 11 Sep 2012 14:09:34 +0000 (16:09 +0200)]
event-loop: export wl_event_loop_dispatch_idle()
When integrating the wayland event-loop into another event-loop, we
currently have no chance of checking whether there are pending idle
sources that have to be called. This patch exports the
"dispatch_idle_sources()" call so other event loops can call this before
going to sleep. This is what wl_event_loop_dispatch() currently does so we
simply allow external event-loops to do the same now.
To avoid breaking existing applications, we keep the call to
dispatch_idle_sources() in wl_event_loop_dispatch() for now. However, if
we want we can remove this later and require every application to call
this manually. This needs to be discussed, but the overhead is negligible
so we will probably leave it as it is.
This finally allows to fully integrate the wayland-server API into
existing event-loops without any nasty workarounds.
Signed-off-by: David Herrmann <dh.herrmann@googlemail.com>
David Herrmann [Mon, 10 Sep 2012 09:20:35 +0000 (11:20 +0200)]
wayland-server: return new ID in wl_client_add_resource()
wl_client_add_resource() used to return no error even though the new
resource wasn't added to the client. This currently makes it very easy to
DOS weston by simply posting thousands of "create_surface" requests with
an invalid ID. Weston simply assumes the wl_client_add_resource() request
succeeds but will never destroy the surface again as the "destroy" signal
is never called (because the surface isn't linked into the wl_map).
This change makes wl_client_add_resource() return the new ID of the added
object and 0 on failure. Servers (like weston) can now correctly
immediately destroy the surface when this call fails instead of leaving
the surface around and producing memory-leaks.
Instead of returning -1 on failure and 0 on success, I made it return the
new ID as this seems more appropriate. We can directly use it when calling
it with new_id==0.
Signed-off-by: David Herrmann <dh.herrmann@googlemail.com>
David Herrmann [Sat, 8 Sep 2012 15:47:15 +0000 (17:47 +0200)]
event-loop: remove dead code
There is really no need to increment "n" if we never read the value. The
do-while() loop overwrites the value before it is read the first time.
Signed-off-by: David Herrmann <dh.herrmann@googlemail.com>
Philipp Brüschweiler [Thu, 6 Sep 2012 16:54:02 +0000 (18:54 +0200)]
cursor: add cursor.pcf and extraction program
Philipp Brüschweiler [Thu, 6 Sep 2012 16:54:01 +0000 (18:54 +0200)]
cursor: Add a default cursor theme
This theme is loaded when the specified cursor theme can not be found.
These cursors are extracted from the xorg sources and transformed into
raw ARGB data by a small helper program (commited separately).
Kristian Høgsberg [Wed, 29 Aug 2012 18:12:11 +0000 (14:12 -0400)]
tests: Quiet warning
U. Artie Eoff [Fri, 17 Aug 2012 01:12:05 +0000 (18:12 -0700)]
tests: ensure sanity leak check tests pass when leak checks are disabled.
This finalizes Robert Bradfords patch to allow NO_ASSERT_LEAK_CHECK
environment variable to disable leak checks in unit tests.
Signed-off-by: U. Artie Eoff <ullysses.a.eoff@intel.com>
Rob Bradford [Fri, 17 Aug 2012 01:12:04 +0000 (18:12 -0700)]
tests: Allow disabling leak checking assertions by env
Some code coverage tools trigger these assertions when run against the test
suite since they don't free all their memory.
Kristian Høgsberg [Thu, 16 Aug 2012 03:03:21 +0000 (23:03 -0400)]
Add wl_shm_buffer_create()
Kristian Høgsberg [Tue, 14 Aug 2012 17:16:10 +0000 (13:16 -0400)]
client: Add wl_display_connect_to_fd() function
This lets us connect a display to an already existing socket fd.
Scott Moreau [Thu, 16 Aug 2012 06:04:27 +0000 (00:04 -0600)]
protocol: Fix typo.
Philipp Brüschweiler [Tue, 14 Aug 2012 09:53:08 +0000 (11:53 +0200)]
socket-test: don't try to be clever, fail if no XDG_RUNTIME_DIR is set
Not only setenv(), also putenv() allocates memory on my system
(glibc 2.16.0). Just fail with a clear message if XDG_RUNTIME_DIR is
not set.
https://bugs.freedesktop.org/show_bug.cgi?id=52618
Christopher Michael [Sat, 11 Aug 2012 14:09:02 +0000 (15:09 +0100)]
Add support for X cursor themes.
This patch adds a few more directories to search for xcursor themes.
Along with the weston patch, this adds the ability to configure weston
to use an X11 cursor theme. Previously, wayland cursor would just look
in the icons and pixmaps directories for cursor images to load. This
adds the ability to also search in the x cursors directory.
Christopher Michael [Sat, 11 Aug 2012 14:35:47 +0000 (15:35 +0100)]
Fix grammar in the rendering section.
Upon reading some docs, I found a small grammar mistake in the
rendering section. This patch fixes that.
Kristian Høgsberg [Sat, 4 Aug 2012 02:39:51 +0000 (22:39 -0400)]
connection: zero out string padding
We don't want to send random data to the client and this also keeps
valgrind happy.
Andre Heider [Tue, 31 Jul 2012 16:59:09 +0000 (18:59 +0200)]
protocol: Remove "repeat" from "key_state"
This is unused and apparently was never implemented.
Dima Ryazanov [Thu, 26 Jul 2012 08:39:48 +0000 (01:39 -0700)]
Don't skip the first directory entry when reading cursors.
I end up missing the "bottom_right_corner" cursor.
Kristian Høgsberg [Tue, 24 Jul 2012 19:45:00 +0000 (15:45 -0400)]
Bump version to 0.95.0
Kristian Høgsberg [Tue, 24 Jul 2012 19:10:12 +0000 (15:10 -0400)]
Make distcheck work from builddir configured with --disable-documentation
Daniel Stone [Mon, 23 Jul 2012 18:54:42 +0000 (19:54 +0100)]
More consistent ID printing
Use unsigned rather than signed for IDs, so they match up with what we
see in other prints.
Signed-off-by: Daniel Stone <daniel@fooishbar.org>
Daniel Stone [Mon, 23 Jul 2012 18:54:41 +0000 (19:54 +0100)]
Make NEW_IDs nullable
The connection-handling code already allows this, so make it legal in
the protocol definition too.
Signed-off-by: Daniel Stone <daniel@fooishbar.org>
Daniel Stone [Mon, 23 Jul 2012 18:54:40 +0000 (19:54 +0100)]
Unstatic arg_count_for_signature and get_next_argument
Expose these to other files using wayland-private.h, so wayland-client.c
can walk NULLables properly.
Signed-off-by: Daniel Stone <daniel@fooishbar.org>
Kristian Høgsberg [Tue, 24 Jul 2012 00:14:33 +0000 (20:14 -0400)]
tests: Wrap calloc by just returning NULL if we're called too early
Since glibc dlsym() calls calloc, we get a call to our calloc wrapper as
we try to look up the real calloc implementation. dlsym() will fall back
to a static buffer in case calloc returns NULL, so that's what we'll do.
This is all highly glibc dependent, of course, but the entire malloc
weak symbol wrapper mechanism is, so there's no loss of generality here.
Daniel Stone [Mon, 23 Jul 2012 18:54:39 +0000 (19:54 +0100)]
test-runner: Wrap realloc() too
So all our tests don't start failing just because we had the temerity to
use realloc() rather than malloc().
Signed-off-by: Daniel Stone <daniel@fooishbar.org>
Daniel Stone [Mon, 23 Jul 2012 18:54:38 +0000 (19:54 +0100)]
.gitignore: Add ctags and cscope files
Signed-off-by: Daniel Stone <daniel@fooishbar.org>
Kristian Høgsberg [Sun, 22 Jul 2012 19:46:16 +0000 (15:46 -0400)]
protocol: Add transform argument to wl_output.geometry event