[ANNOUNCE] xkbcomp 1.3.0
In what is almost an olympic release schedule, here's a new version of xkbcomp. Over the last two-and-a-bit years a number of patches have accumulated, the most interesting one is support for multiple keysyms per level (#25025). I say "parsing support" for a reason, the symbol becomes a NoSymbol, which is arguably still more useful than a parsing error. Plus, multi-sym per key won't work in X anyway. Other than that, misc fixes and changes all over the place. UNIXOS2 support was dropped. Sorry guys, no multi-keysym parsing for you. I'll get a bucket ready for the tears. Adam Jackson (1): configure: Drop AM_MAINTAINER_MODE Alan Coopersmith (10): unifdef -U__UNIXOS2__ config: Add missing AC_CONFIG_SRCDIR Remove unused function entry/exit tracking framework Remove unused uASSERT macro Convert remaining sprintf calls to snprintf Fix many const char * warnings from gcc Remove useless checks for NULL before free in OverlayKeyCreate() Don't dereference the pointer whose allocation failed Remove useless assignment to 'outline' variable Initialize nMatch even if WIN32 is defined Benno Schulenberg (1): Making sure that a copied string is always null-terminated (#66345). Colin Walters (1): autogen.sh: Honor NOCONFIGURE=1 Daniel Stone (2): Add parsing support for multiple keysyms per level Reset scan state when opening a new file Laura (1): Add #include to xkbscan.c Peter Hutterer (7): Use DEBUG, not DEBUG_ON to determine whether debugging is enabled man: document -help/-?, -em1, -emp, -eml Always terminate the scanBuf string (#66345) Parse -w1 flag correctly (#66344) compat: don't warn about redefinition when nothing is defined yet man: replace default include directory with the one from configure xkbcomp 1.3.0 Ryan Pavlik (1): Include Xwindows.h rather than windows.h Thomas Klausner (1): Protect config.h like usual. Vincent Lefevre (1): xkbcomp: Improved -w option parsing git tag: xkbcomp-1.3.0 http://xorg.freedesktop.org/archive/individual/app/xkbcomp-1.3.0.tar.bz2 MD5: 0012a8e3092cddf7f87b250f96bb38c5 xkbcomp-1.3.0.tar.bz2 SHA1: 113c93679c9245141b5899240f59fcc8227d8dc1 xkbcomp-1.3.0.tar.bz2 SHA256: cfac973778fabf5216121ad60b7af8ab74ce7513af0f9260cf8c5309e1622b2a xkbcomp-1.3.0.tar.bz2 PGP: http://xorg.freedesktop.org/archive/individual/app/xkbcomp-1.3.0.tar.bz2.sig http://xorg.freedesktop.org/archive/individual/app/xkbcomp-1.3.0.tar.gz MD5: f8da094e266e2f99316696fab4922d70 xkbcomp-1.3.0.tar.gz SHA1: 32fffd47086a6d204ac842668a598af74d9c172a xkbcomp-1.3.0.tar.gz SHA256: 91d052c895a47ab2930aa1e150bfe7559fdaeb959d035634d5ba259300a3353f xkbcomp-1.3.0.tar.gz PGP: http://xorg.freedesktop.org/archive/individual/app/xkbcomp-1.3.0.tar.gz.sig pgpwk1inJpOv7.pgp Description: PGP signature ___ xorg@lists.x.org: X.Org support Archives: http://lists.freedesktop.org/archives/xorg Info: http://lists.x.org/mailman/listinfo/xorg Your subscription address: %(user_address)s
Re: Graphics speed depends on user profile
Juan Pablo de la Cruz wrote: > We have an library which directly uses the Xlib API to do graphics. > > We are observing the following effect: when two different users log > in the same machine, use the same window manager, and run the very same > program, the graphics are slower (much) for one of the users. > > Concretely, the user for which the drawing is slower, logs himself in, > and the other user executes su $USER on the terminal. > > Both use the exact same libraries (same PATH and LD_LIBRARY_PATH). > Both use the same window manager. > > What other settings/factors could explain this behaviour? What are their $DISPLAY settings? If $DISPLAY is ":0" then the client will connect to the X server via the Unix-domain socket at /tmp/.X11-unix/X0. Whereas if $DISPLAY is "localhost:0", it will connect to the X server via TCP port 6000. The latter will result in the data stream being chopped up into packets and passed through the kernel's TCP/IP stack: routing, firewall rules, etc. It may also prevent the use of features such as direct rendering, as the client and server are likely to be deemed to reside on different systems. -- Glynn Clements ___ xorg@lists.x.org: X.Org support Archives: http://lists.freedesktop.org/archives/xorg Info: http://lists.x.org/mailman/listinfo/xorg Your subscription address: %(user_address)s
Graphics speed depends on user profile
Hello everyone, We have an library which directly uses the Xlib API to do graphics. We are observing the following effect: when two different users log in the same machine, use the same window manager, and run the very same program, the graphics are slower (much) for one of the users. Concretely, the user for which the drawing is slower, logs himself in, and the other user executes su $USER on the terminal. Both use the exact same libraries (same PATH and LD_LIBRARY_PATH). Both use the same window manager. What other settings/factors could explain this behaviour? Could you point me to some source/guide providing some helpful information? Currently I am out of ideas. Thanks and greetings, Juan Pablo -- Juan Pablo de la Cruz Gutierrez c...@mvtec.com MVTec Software GmbH | Neherstr. 1 | 81675 München | Germany www.mvtec.com | Tel: +49 89 457695-0 | Fax: +49 89 457695-55 Geschäftsführer: Dr. Wolfgang Eckstein, Dr. Olaf Munkelt Amtsgericht München HRB 114695 ___ xorg@lists.x.org: X.Org support Archives: http://lists.freedesktop.org/archives/xorg Info: http://lists.x.org/mailman/listinfo/xorg Your subscription address: %(user_address)s
[PATCH] Xi/exevents : fix a issue that window can't be closed by touch the close window button
Fixes: https://bugs.freedesktop.org/show_bug.cgi?id=86459 When matchbox-window-manger(Version 1.2) create a window, there is a close window button on the top-right of this new window which seems like "X". If we use a mouse, and click this "X" close window button, the window can be closed. But when we use multi-touch screen, and touch this "X" close window button, this window can't be closed, and nothing changed on the screen. By the way, multi-touch can work normal when we touch the window's other place except the "X" close button. And both single touch and mouse can work normal, they can close the window by touch/click the "X" close window button. In Xorg(Version 1.16.1), mouse click event, single touch event, and multi-touch event all call the below function to deliver the coming event to the target window. int DeliverDeviceEvents(WindowPtr pWin, InternalEvent *event, GrabPtr grab, WindowPtr stopAt, DeviceIntPtr dev) If we use mouse or single touch to click/touch the "X" close window button, function ProcessDeviceEvent() call the above function. When the event->any.type=ET_ButtonPress, the pWin point to the "X" close window button. The pWin is got by function GetSpriteWindow(). But if we use multi-touch to touch the "X" close window button, function DeliverTouchEmulatedEvent() call DeliverDeviceEvents(). When the event->any.type=ET_ButtonPress, the pWin does not point to the "X" close window button, it actually point to the "X" close window button's parent window. So in function DeliverDeviceEvents(), when call function DeliverOneEvent(), the value child is zero, rather than the target window. This patch use function GetSpriteWindow() to get the pWin in function DeliverTouchEmulatedEvent() when call DeliverDeviceEvents(). With this patch, mouse, single touch, multi-touch all can work well, including click/touch the "X" close window button. Signed-off-by: Haibo Chen --- Xi/exevents.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/Xi/exevents.c b/Xi/exevents.c index b0bc47e..e1c523e 100644 --- a/Xi/exevents.c +++ b/Xi/exevents.c @@ -1430,7 +1430,7 @@ DeliverTouchEmulatedEvent(DeviceIntPtr dev, TouchPointInfoPtr ti, else { GrabPtr devgrab = dev->deviceGrab.grab; -DeliverDeviceEvents(win, ptrev, grab, win, dev); +DeliverDeviceEvents(GetSpriteWindow(dev), ptrev, grab, win, dev); /* FIXME: bad hack * Implicit passive grab activated in response to this event. Store * the event. -- 1.9.1 ___ xorg@lists.x.org: X.Org support Archives: http://lists.freedesktop.org/archives/xorg Info: http://lists.x.org/mailman/listinfo/xorg Your subscription address: %(user_address)s