Age | Commit message (Collapse) | Author | Files | Lines |
|
Fedora 12 likes to use a 39x26 animated wait cursor. Avoid spamming
the log with warnings that each frame can't be completely contained
in the 32x32 native cursor
Also reformat log message so it doesn't contain a '\n\t'. I mean,
it's not like we might want to grep the log or something...
Signed-off-by: Jon TURNEY <jon.turney@dronecode.org.uk>
|
|
commit 446fe9eecddd1337f9d5164dd7c301e1ba3dfe32 removes the AC_DEFINE for
SERVER_LOCK and conditional compilation checking it, making it always on
everywhere, except in os/utils.c where code is left under SERVER_LOCK, which
now never gets built, making the '-nolock' option non-functional...
Signed-off-by: Jon TURNEY <jon.turney@dronecode.org.uk>
|
|
This is a temporary solution.
Really we should either improve the way we try to avoid the internal
clipboard client getting killed or perhaps restart it after an interval.
Avoiding getting killed by the XDM is horribly heuristic at the moment,
as we basically try to guess about what the XDM is doing, so this is
always going to be fragile. (Should just check that we can't actually
intercept the close request for the client)
Restarting the clipboard client thread is complicated by the fact that
at the moment the clipboard thread exits on server regeneration and is
restarted by hooks installed by InitInput())
(So we either need to make arrangements not restart when server is restarting,
or only start once and properly re-cache atoms on server regeneration)
Also: http://xkcd.com/292/
Signed-off-by: Jon TURNEY <jon.turney@dronecode.org.uk>
|
|
Reverts the change which makes parented windows non-resizeable
Because this was trying to do something which we should be doing, as an alternative we
try checking WM_NORMAL_HINTS for windows which shouldn't be resizable
If a window has a maximum size specified, no maximize box
If a window has a fixed size (max size == min size, per EWMH 1.3 Implementation Notes), no resizing frame
Signed-off-by: Jon TURNEY <jon.turney@dronecode.org.uk>
|
|
Tidying up other WM detection code when adding SWT/Motif fix
was rather over-agressive and now allows other WMs to think
they can start when the internal WM is running.
Revert to the behaviour of selecting ButtonPressMask events
as well on the root window, so other WMs will be dissuaded
from starting
Signed-off-by: Jon TURNEY <jon.turney@dronecode.org.uk>
|
|
happens at the moment if the locale is C.UTF-8
|
|
Restructure to group similar options and offer some commentary on those groups
Review option descriptions, clarify and bring up-to-date
Signed-off-by: Jon TURNEY <jon.turney@dronecode.org.uk>
|
|
Alphabetize options
Review option descriptions, clarify and bring up to date
Add missing option descriptions
Signed-off-by: Jon TURNEY <jon.turney@dronecode.org.uk>
|
|
Treat XSupportsLocale() returning false as non-critical to internal client
theads startup, and issue a warning, not an error
Signed-off-by: Jon TURNEY <jon.turney@dronecode.org.uk>
|
|
|
|
|
|
All DDXs segfault on Cygwin unless -DFD_SETSIZE=256 is set, so make sure
we do so whether or not we are building XWin.
Signed-off-by: Yaakov Selkowitz <yselkowitz-Rn4VEauK+AKRv+LV9MX5uipxlwaOVQ5f@public.gmane.org>
|
|
X windows with no geometry specified end up being created at the same place,
the top-left, rather than being placed by Windows in it's default arrangement
(this is particularly noticable with a dual-monitor setup with the primary monitor
on the right as the Windows default placement puts new windows on the primary monitor,
but new X windows end up at the top-left which is on the secondary monitor.)
The SetWindowPos(SWP_FRAMECHANGED) we do after applying the WS_POPUP style after creation
causes a WM_STYLECHANGED to be emitted
cygwin-hint-handling-fixes patch adds handling or WM_STYLECHANGED and uses it to adjust
the window size to preserve the client area (and has a side effect of synchronizing
the native Window postion with the X window position)
Previously, however, this inital WM_STYLECHANGE was ineffective, as the hWnd stored in
the window privates was not set until after SetWindowPos() was called
Commit 17e67c40 moved the setting of this hWnd private up to immediately
after CreateWindowEx (not unreasonably, as having a window where it is not set but things
are operating on the window is not good), unfortunately, as shown above, the current code
depends on this window for things to work correctly :-(
Yuck!
To perhaps clean this up a bit, we need to stop being lazy in WM_STYLECHANGED and
write a proper analogue to winPositionWindowMultiWindow() which only sets the window
size (or allow that function, suitably renamed, to accept flags that it should adjust
the X window size and/or position)
|
|
at startup.
Otherwise, this happens lazily after the first keypress, which can lead
to applications which are started from a shell window and inspect the
keyboard state before a character is typed getting the wrong idea about
the desired keymap (e.g. xemacs shows this behaviour)
Signed-off-by: Jon TURNEY <jon.turney@dronecode.org.uk>
|
|
threads exit due to an error
Signed-off-by: Jon TURNEY <jon.turney@dronecode.org.uk>
|
|
Choose display name used to connect internal clients and exported into
environment of processes started by traymenu so that it uses a transport
we know is working
This should mean the server can start correctly with -multiwindow
and/or -clipboard and any two of -nolisten inet6 -nolisten inet
and -nolisten unix (the server will correctly refuse to start if all 3
are used, as it must be listening on at least one socket)
Signed-off-by: Jon TURNEY <jon.turney@dronecode.org.uk>
|
|
clipboard
XConvertSelection() in libX11 always returns 1, so there is no point in
testing it incorrectly against Success. This is probably a bug in
XConvertSelection()
This should fix UTF8String and CompoundText selection via the clipboard.
Signed-off-by: Jon TURNEY <jon.turney@dronecode.org.uk>
|
|
Signed-off-by: Jon TURNEY <jon.turney@dronecode.org.uk>
|
|
It expects all top-level windows to get reparented, and waits until they
do. So workaround that in our internal WM forcing a reparent event to
occur, even though we don't actually need to reparent the window to
frame it (as the frame is a native window, not an X window)
Signed-off-by: Jon TURNEY <jon.turney@dronecode.org.uk>
|
|
Add a timestamp to log messages
Also tidy up some cosmetic issue in log strings:
- Add a missing '\n'
- Fix some strings starting with '\n'
- Remove '\f' from some log strings
|
|
Workaround a bug in iiimxcf (assuming the WM_STATE atom exists),
which can cause many Solaris clients to simply fail with a BadAtom
error
Signed-off-by: Jon TURNEY <jon.turney@dronecode.org.uk>
|
|
Don't conditionalize use of a authorization cookie for internal client
threads on XCSECURITY, always use one (this avoids certain problems
with XDMCP setups where the XDMCP host removes localhost from the access
list etc.)
Conditionalize the use of a XCSECURITY authorization descriptor on XCSECURITY
Consolidate the various places where the authorization cookie is set
for internal threads into a new function, winSetAuthorization()
Use authorization cookie for multiwindow WM X message thread as well
Signed-off-by: Jon TURNEY <jon.turney@dronecode.org.uk>
|
|
|
|
We still accept -clipboard for backwards compatibility. If both are passed,
the last one is accepted (just as other arguments are handled).
|
|
|
|
This wreaks havoc with other programs which manage the keyboard, such
as MS Office Clipboard or Win32 VNC viewers:
http://cygwin.com/ml/cygwin-xfree/2009-02/msg00264.html
|
|
/usr/lib/libXfont.a(fontfile.o):fontfile.c:(.text+0x0): multiple
definition of `_CopyISOLatin1Lowered'
../../../dix/.libs/libdix.a(dixutils.o):dixutils.c:(.text+0xe0): first
defined here
|
|
(which seems to be behind some of the reported failures)
Turn useless #if 0/ErrorF()/#endif into useful winDebug()
|
|
|
|
with a scan code, so look up what scan code corresponds to the virtual key
code if this occurs.
Patch by Paul Loewenstein <paul.loewenstein@gmail.com>
|
|
TODO: install system.XWinrc via hw/xwin/Makefile.am
TODO: Rename sample system.Xwinrc file
|
|
and use separate logs for each display instance (e.g. XWin.0.log).
FIXME: g_pszLogFile should be (const char*) per os/log.c:LogInit().
|
|
|
|
|
|
Provide in dmx, zephyr, fake and sdl DDXs
ddxBeforeReset() if DDXBEFORERESET is defined
OsVendorFatalError() if DDXOSFATALERROR is defined
|
|
with libGL for them instead (as dri_swrast.so is linked with libGL)
Paper over the cracks. Avoid a crash when we have a broken context.
Turn on GLX null context debugging
|
|
Ensure window actually ends up somewhere visible if it tries to create itself offscreen
Signed-off-by: Jon TURNEY <jon.turney@dronecode.org.uk>
|
|
Respect the systems ownership of the clipping region
When the style changes, adjust the window size so the client area remains the same.
Otherwise the window size may change when sizing is reflected from
Windows to X, and some windows are drawn expecting them to be exactly the reqeusted size
(e.g. gmplayer control window)
Signed-off-by: Jon TURNEY <jon.turney@dronecode.org.uk>
|
|
|
|
We handle this with a hack in the .cygport, so the rest of the libraries
can be linked dynamically.
This patch is NOT suitable for pushing upstream.
|
|
Apply the Windows title wide-character in -multiwindow mode. Windows should now
display correct client X Window titles for locales with wide characters.
Copyright (C) Colin Harrison 2005-2008
http://www.straightrunning.com/XmingNotes/
http://sourceforge.net/projects/xming/
|
|
Add -auth option handling for Xming (patch3)
Use an internally generated cookie for authentication of server clients when running the -auth option in -multiwindow mode. (patch11)
XXX: doesn't internal clipboard client need similar authentication?
XXX: does this need some thinking about other security mechanisms we have now...
Copyright (C) Colin Harrison 2005-2008
http://www.straightrunning.com/XmingNotes/
http://sourceforge.net/projects/xming/
|
|
duplicate call to winMultiWindowGetTransientFor())
|
|
|
|
disable it for cygwin
|
|
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
|
|
Introduced in commit 9998105a387e0294054502331a56e1e020cd93e4
The replacement third parameters to PKG_CHECK_MODULES([DMXMODULES]
was not quoted.
Signed-off-by: Gaetan Nadon <memsize@videotron.ca>
Reviewed-by: Dan Nicholson <dbn.lists@gmail.com>
Tested-by: Julien Cristau <jcristau@debian.org>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
(cherry picked from commit 91c1bd78f7240c92702828f8e5a6b6ce944b9e36)
(cherry picked from commit b8623569a3acff31cb18a9f6046e60d0637cc0e0)
|
|
Restores old ABI after 196aff9b18381d700fb28027186cce6e68ad587c.
Signed-off-by: Aaron Plattner <aplattner@nvidia.com>
Tested-by: Julien Cristau <jcristau@debian.org>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
(cherry picked from commit 155e61a9f0429bf28ce493c0fe7a2d076cb7e137)
|
|
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
|
|
That change causes lib(w)fb to make accelerated driver calls after the driver
has entered a software fallback. Most drivers don't expect this, which leads to
corruption or crashes. A change to make this code do the copy in software is
unacceptably slow.
This reverts commit e9aa61e9f0d663d5b34a397b943b4d1df44e873d.
Conflicts:
configure.ac
fb/fbpict.c
Signed-off-by: Aaron Plattner <aplattner@nvidia.com>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
|