Fixed race conditions when a ServerApp or ServerWindow is created. The
reply to the client that the object has been created successfully was
sent in the thread creating it. Preempted at the wrong time (right after
writing the message to the port) could lead to the object's thread using
the link at the same time, which would screw up all subsequent
communication via that link.
This fixes the problem that mimeset would sometimes fail when building
Haiku in Haiku (can't find the ticket). It probably also fixes#2331,
and the bug that sometimes when a window is opened (Terminal, crash
alert, shutdown window, etc.) it would come up with huge width/height
and tiny other dimension (can't find the ticket).
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@26192 a95241bf-73f2-0310-859d-f6bbb57e9c96
Since ServerWindow removed itself from its ServerApp in _PrepareQuit(), it could
happen quite easily that the ServerApp was deleted before the ServerWindow - and
since deleting WindowLayer as part of that referenced the ServerApp, it crashed.
Now, adding/removing is both done by the ServerWindow in Init() respectively
the destructor.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@17198 a95241bf-73f2-0310-859d-f6bbb57e9c96
even though we might need something similar, we can't use it (since it was
based on BGet++).
* Removed BGet++, it's not used anymore.
* Removed ServerMemIO class, was never used.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@16830 a95241bf-73f2-0310-859d-f6bbb57e9c96
all teams in serveral server areas, and instead of having to eventually clone
them all several times in BBitmap, we now have one or more areas per team,
and BBitmap will only clone areas once if needed. As a side effect, this
method should be magnitudes faster than the previous version.
* This method is also much more secure: instead of putting the allocation
maintenance structures into those everyone-read-write areas, they are now
separated, so that faulty applications cannot crash the app_server this
way anymore. This should fix bug #172.
* Freeing memory is not yet implemented though! (although all memory will
be freed upon app exit)
* There are now 3 different bitmap allocation strategies: per ClientMemoryAllocator
(ie. via ServerApp), per area (for overlays, not yet implemented), and using
malloc()/free() for server-only bitmaps.
* ServerBitmap now deletes its buffers itself.
* Cleaned up BBitmap and BApplication a bit.
* The test environment currently doesn't build anymore, will fix it next.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@16826 a95241bf-73f2-0310-859d-f6bbb57e9c96
when it's not over a view of the application.
* The application cursor is no longer applied when the mouse cursor is over
the border (or tab) of a window.
* Gave up and imitate BeOS behaviour: the mouse cursor now always get the
shape the view below dictates, ie. it will no longer fall back to the
default cursor outside the focus window.
* The window is now set to the default in case there is no window under it
at all (ie. if Tracker isn't running).
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@16685 a95241bf-73f2-0310-859d-f6bbb57e9c96
of the view currently under the mouse, or the application cursor, if the view doesn't
have its own cursor (ie. it will no longer set the wrong cursor in certain situations).
* AS_DELETE_CURSOR has no longer an influence on the application cursor, as this grabs
a ref for its own use - this fixes bug #275.
* AS_SET_CURSOR no longer sets the cursor when the application is not active.
* Minor cleanup.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@16678 a95241bf-73f2-0310-859d-f6bbb57e9c96
work though, as HWInterface can only draw B_RGB32 cursors...
* More build fixes for libbe_test target - it defines __HAIKU__ as well, now
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@16211 a95241bf-73f2-0310-859d-f6bbb57e9c96
though (ideally when launched).
* B_MODAL_APP_WINDOW_FEEL windows are now visible on the initial app workspace
in case it has no other window visible. This fixes the missing BAlerts in
the debug_server or just the "alert" command. Thanks to Stephan for pointing
this out.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@15811 a95241bf-73f2-0310-859d-f6bbb57e9c96
* Desktop::SetWindowFeel() is now working, but doesn't respect the window
position yet (ie. floating windows would have to be moved to front).
* WindowLayer::Workspaces() now always reflects exactly the lists it's in
(at least after Desktop::AddWindow() has been called). WindowList::AddWindow()
and RemoveWindow() now update this flag to be correct at all times.
* Fixed Desktop::_ChangeWindowWorkspaces() to not set the current workspace
for windows that are not on the current workspace, and vice versa.
It also would hide windows that were already hidden, and tried to show
windows that actually were hidden (did no other harm than triggering
a rebuild of the global clipping).
* Floating windows now work as expected.
* Desktop::SetFocusWindow() won't give a window focus that has a modal.
* Renamed OnWorkspace() to InWorkspace().
* ServerApp::InWorkspace() now works correctly, added ServerApp::Workspaces()
as well.
* WindowLayer::SubsetWorkspaces() returns the workspaces mask this modal or
floating window should be in.
* New window flag B_SAME_POSITION_IN_ALL_WORKSPACES should work as well.
* Floating and modal windows now have always set this flag.
* Added a WindowList::HasWindow() method.
* Desktop windows (windows with the desktop feel) can now have focus again
(I accidently broke that before).
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@15434 a95241bf-73f2-0310-859d-f6bbb57e9c96
* introduced ServerApp::OnWorkspace().
* moved AS_CREATE_[OFFSCREEN_]WINDOW into its own method, ServerApp::_CreateWindow().
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@15283 a95241bf-73f2-0310-859d-f6bbb57e9c96
* workspace switch and subset windows functionality temporarily removed
(away with that mess!).
* no more RevealWMState() - we now have methods like ActivateWindow()
and SendWindowBehind() that do all the work - just a little cleaner
and with less overhead.
* Workspace is now a pretty passive class - it only stores configurations
of the windows and screens.
* added an evil work-around for a locking problem (in RootLayer::_SetFocus()).
* I'll plan to move pretty much all of the remaining root layer functionality
to Desktop - so that the all regions lock is only held in case clipping
regions are affected.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@15207 a95241bf-73f2-0310-859d-f6bbb57e9c96
client's token space.
* finding layers by token is now O(1) instead of O(n) (as they are looked
up in the token space).
* removed Layer::FindView() as it's no longer needed (or should be used).
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@15113 a95241bf-73f2-0310-859d-f6bbb57e9c96
task_looper() again.
* removed BMessenger::fPreferred - whenever you had to specify "usePreferred" separately,
you don't have to do that anymore - use B_PREFERRED_TOKEN instead.
* fixed BTokenSpace::GetToken() semantics: it will no longer touch the "object" argument
in case of failure.
* Introduced a BWindow::_DistributeMessage() that will be part of the event dispatcher
counterpart to the app_server (the other will be _DetermineTarget()).
* Made it easier to use Michael's Message4 implementation: just add the following line
to your UserBuildConfig:
AppendToConfigVar DEFINES : HAIKU_TOP src : USING_MESSAGE4 : global ;
* Introduced ServerWindow::HandlerMessenger() and FocusMessenger() - the first will
target the client handler, while the other will target the preferred handler of the
client looper (usually the view having focus).
* Fixed dano message unflattening in the Message4 code.
* Changed BMessage::PrintToStream() to no longer use macros in the Message4 implementation.
* I hope that's all - it's a huge change, but it's all connected.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@15046 a95241bf-73f2-0310-859d-f6bbb57e9c96
distribute any messages to the clients yet.
* removed the working thread from RootLayer - for now, its event handlers are
still called using input filters in the new event dispatcher, though (to
get things started).
* ServerApp is now using a BMessenger to identify its client, and no longer
stores the port/token separately.
* the input_server handshake is a bit simpler now, as it can now just reply
to the app_server message, removed unused code from ServerProtocol.h
* calmed down the MultiLocker (it always printed thread statistics on startup,
because it's compiled in debug mode).
* removed the cursor thread stuff from AppServer.cpp
* the new event dispatcher now uses a cursor thread when supported (only in
native mode, not in the test environment), although it improves cursor
movement under Qemu, the effect is not as good as expected - this might
need some more investigations (might just be a thread priority problem).
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@15012 a95241bf-73f2-0310-859d-f6bbb57e9c96
* Introduced and implemented AS_GET_SCREEN_ID_FROM_WINDOW - it only returns B_MAIN_SCREEN_ID,
though.
* renamed ServerWindow::fHandlerToken to fClientToken.
* The BScreen(BWindow *) constructor now really asks the server for the screen ID.
* ServerApp::fWindowList is now a BObjectList.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@14910 a95241bf-73f2-0310-859d-f6bbb57e9c96
creation/deletion (and management) over to that class.
ServerApp now gets a desktop pointer, and no longer uses gDesktop.
Converted private MessageLooper::_MessagePort() to a public method MessagePort()
so that the looper can be addressed from elsewhere without using PostMessage().
Added a real basic message loop to MessageLooper::_MessageLoop().
BApplication now only asks the app_server to get its desktop object which should
now be used for everything that's not in the realm of the application.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@13824 a95241bf-73f2-0310-859d-f6bbb57e9c96
be improved a bit (Quit() and _MessageLooper() are empty right now).
Removed ServerApp::PingTarget().
Hopefully cleared some confusion about ServerApp::fClientLooperPort and fClientToken
(previously fHandlerToken), even if it's currently unused.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@13807 a95241bf-73f2-0310-859d-f6bbb57e9c96
windows that wouldn't quit on demand, the app_server would have gotten
the kShutdownServer message anyway already (as the last app was quit).
Since that one removed things like gDesktop/gBitmapManager, it liked
crashing.
Now, there is a semaphore that will be send to each app on quit. Only
when this semaphore can be acquired, the shutdown message will be sent.
Removed unused semaphores (decorator, active app).
Replaced fAppListLock with a BLocker (just calling acquire_sem() without
error checking is very unsafe in userland, and should never be done).
BTW the bug was triggered by broken menu code that only sometimes
really quit the window; it leaves a whole lot of zombies around - Stefano,
any quick idea? :-)
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@13384 a95241bf-73f2-0310-859d-f6bbb57e9c96
used for new area.
MemPool::AddToPool() now gracefully deals with NULL pointers (or a size of 0).
BitmapManager was deleting the area it transferred to AreaPool before - it
no longer needs an extra area, though.
Minor other cleanup.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@13260 a95241bf-73f2-0310-859d-f6bbb57e9c96
will update it automatically.
Renamed ServerWindow::fName to fTitle, made it a pointer - it will now just
adopt the title pointer that came from AS_CREATE_WINDOW.
Just another cleanup round: renamed Layer::GetName() to Name(), no
more layer->fName->String() accesses.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@13239 a95241bf-73f2-0310-859d-f6bbb57e9c96
can now quit instantly.
AppServer must no longer call ServerApp's destructor once it's running - it now
has to call Quit() in this case. The ServerApp is now destructed in its own thread.
Some cleanup (like renaming ServerApp::MonitorThreadID() to Thread()).
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@13238 a95241bf-73f2-0310-859d-f6bbb57e9c96
to LinkReceiver, LinkMsgSender to LinkSender, and put everything into the
BPrivate namespace.
Made AppServerLink a cheap object - it will use the applications receiver/sender
and not create its own buffers.
Fixed broken communication stuff here and there (mostly Font.cpp).
Put the newly introduced set|get_system_colors() into the BPrivate namespace -
please don't introduce private functions into the public namespace!!!
Also fixed their broken communication use, as Darkwyrm obviously forgot about
it again: the sequence Flush(); GetNextMessage() without error checking is
purely wrong and can make the app hang and/or crash! :-)
Other minor cleanup.
The input_server used some test mode with the haiku build target which is
probably wrong.
Hopefully I did not forget anything this time.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@13128 a95241bf-73f2-0310-859d-f6bbb57e9c96
Added LinkMsgReader::NeedsReply() method.
Completely redone ServerApp messaging: no more "replyport" from BAppServerLink; instead,
the registered client reply port is used. Fixed some more weak messaging stuff.
ServerApp now recognizes if an unknown message needs a reply, and sends it - for example,
the "Screen" preferences app no longer hangs, but crashes on start :)
Made LinkMsgReader::Read() virtual again, since it's needed by RAMLinkMsgReader.cpp.
Renamed BPortLink::GetNextReply() to GetNextMessage().
Some more cleanup.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@13004 a95241bf-73f2-0310-859d-f6bbb57e9c96
