May 2008 - Xquartz-dev - lists.macosforge.org

Re: 2.2.1 Released
by Jeremy Huddleston 02 May '08

02 May '08

Xfake was released to replace Xvfb a couple years ago... we've been phasing out Xvfb since Xfake does everything it did. On May 1, 2008, at 20:45, Bill Janssen wrote: >>> 1. No Xvfb. >> >> yep. good riddance. > > Not for me. > >> When was the last time Xfake did work for you? > > Never. I never had to use it before, because I had Xvfb! > > Bill > _______________________________________________ > Do not post admin requests to the list. They will be ignored. > X11-users mailing list (X11-users(a)lists.apple.com) > Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/x11-users/jeremyhu%40freedesktop.org > > This email sent to jeremyhu(a)freedesktop.org

1 0

Re: 2.2.1 Released
by Jeremy Huddleston 02 May '08

02 May '08

On May 1, 2008, at 16:57, Bill Janssen wrote: > I just installed this on a new Leopard (10.5.2) machine, and I see a > couple of things: > > 1. No Xvfb. yep. good riddance. > 2. Xfake doesn't run: > > % Xfake :20 > InitConnectionLimits: MaxClients = 255 > Sync Extension 3.0 > Fopen(/usr/X11/lib/xserver/SecurityPolicy), fp = 0xa0c5d374 > Pclose: fp = 0xa0c5d374 > added keyboard (null) with dix id 3 > (dix) initialising device 0 > (dix) initialising device 1 > (dix) initialising device 2 > initialising pointer Generic Pointer ... > (dix) initialising device 3 > initialising keyboard (null) > bad keycode (1,116) in XkbAllocClientMap > > Fatal server error: > Couldn't allocate client map in XkbInitDevice When was the last time Xfake did work for you?

1 0

1.4.0-apple6
by Jeremy Huddleston 02 May '08

02 May '08

I've uploaded a new Xquartz binary for the 1.4 branch. This fixes the crash that sometimes occurred when exiting as well as the multimonitor issue when the primary display was not the upper left monitor. Just drop it in over 2.2.1's /usr/X11/bin/Xquartz to test. http://xquartz.macosforge.org/downloads/Xquartz-1.4.0-apple6.bz2 There are currently no known crashes and no known regressions over 1.3.0-apple20 (which is what comes with 2.2.1). Please test this and let me know how it goes for you. I'm looking around May 15 or so for the 2.3 release which will be mostly the same as 2.2.1 but with the newer server. Please test out the following areas: multi-monitor support: It shouldn't be worse than the 1.3 server ... but I'm going to try improving it more for the next release. tablet support: Ben is handling this, and it looks like it's coming along, but there are still some issues where gtk isn't seeing it as a tablet xkb: Please test out 1.4 if you have any keyboard other than an Apple US English Keyboard Also, if you've been having the _xp_async_dequeue lockup, I've heard that it is gone in 1.4... so all known crashes and lockups are gone. (in other words, make it crash and tell me what you did to break it) --Jeremy

1 0

Pen / tablet support and Xquartz
by Ben Byer 01 May '08

01 May '08

I'm trying to add support for tablet events into Xquartz, the Mac OS DDX part of the X.org Xserver. I've had trouble finding any real documentation about this, so I've mostly been trying to read the gdk source to see what it's looking for, and looking at the events sent by the Wacom driver on Linux. I now have events coming in that look valid, but gsumi does still reports that I have "no input devices", and the Gimp doesn't recognize my taps with the pen on the tablet. Here's what xinput reports: $ xinput list "Virtual core keyboard" id=0 [XKeyboard] Num_keys is 248 Min_keycode is 8 Max_keycode is 255 "Virtual core pointer" id=1 [XPointer] Num_buttons is 32 Num_axes is 2 Mode is Relative Motion_buffer is 256 Axis 0 : Min_value is 0 Max_value is -1 Resolution is 0 Axis 1 : Min_value is 0 Max_value is -1 Resolution is 0 "Quartz Pointing Device" id=2 [XExtensionPointer] Num_buttons is 7 Num_axes is 2 Mode is Relative Motion_buffer is 256 Axis 0 : Min_value is 0 Max_value is -1 Resolution is 0 Axis 1 : Min_value is 0 Max_value is -1 Resolution is 0 "pen" id=3 [XExtensionPointer] Num_buttons is 3 Num_axes is 7 Mode is Relative Motion_buffer is 256 Axis 0 : Min_value is 0 Max_value is -1 Resolution is 0 Axis 1 : Min_value is 0 Max_value is -1 Resolution is 0 Axis 2 : Min_value is 0 Max_value is -1 Resolution is 0 Axis 3 : Min_value is 0 Max_value is -1 Resolution is 0 Axis 4 : Min_value is 0 Max_value is -1 Resolution is 0 Axis 5 : Min_value is 0 Max_value is -1 Resolution is 0 Axis 6 : Min_value is 0 Max_value is -1 Resolution is 0 "Quartz Keyboard" id=4 [XExtensionKeyboard] Num_keys is 248 Min_keycode is 8 Max_keycode is 255 $ xinput test 3 [...] otion a[0]=535 a[1]=541 a[2]=137 a[3]=0 a[4]=0 motion a[0]=534 a[1]=541 a[2]=90 a[3]=0 a[4]=0 motion a[0]=534 a[1]=541 a[2]=74 a[3]=0 a[4]=0 motion a[0]=533 a[1]=542 a[2]=58 a[3]=0 a[4]=0 motion a[0]=524 a[1]=707 a[2]=337 a[3]=0 a[4]=0 [...] motion a[0]=514 a[1]=535 a[2]=447 a[3]=0 a[4]=1 motion a[0]=514 a[1]=535 a[2]=447 a[3]=0 a[4]=5 motion a[0]=514 a[1]=535 a[2]=443 a[3]=0 a[4]=9 motion a[0]=514 a[1]=535 a[2]=443 a[3]=0 a[4]=9 motion a[0]=505 a[1]=535 a[2]=592 a[3]=8 a[4]=0 motion a[0]=505 a[1]=535 a[2]=592 a[3]=9 a[4]=0 [...] The code on the Quartz side is pretty simple: http://cgit.freedesktop.org/xorg/xserver/tree/hw/xquartz/darwinEvents.c#n363 (but edited here for clarity) /* These values were chosen to match the output of xinput under Linux */ #define SCALEFACTOR_TILT 64.0 #define SCALEFACTOR_PRESSURE 1000.0 void DarwinSendPointerEvents(int ev_type, int ev_button, int pointer_x, int pointer_y, float pressure, float tilt_x, float tilt_y) { int i, num_events; DeviceIntPtr dev; int valuators[5] = {pointer_x, pointer_y, pressure * SCALEFACTOR_PRESSURE, tilt_x * SCALEFACTOR_TILT, tilt_y * SCALEFACTOR_TILT}; if (pressure == 0 && tilt_x == 0 && tilt_y == 0) dev = darwinPointer; else dev = darwinTablet; mieqEnqueue_lock(); { num_events = GetPointerEvents(darwinEvents, dev, ev_type, ev_button, POINTER_ABSOLUTE, 0, dev==darwinTablet?5:2, valuators); for(i=0; i<num_events; i++) mieqEnqueue (dev,&darwinEvents[i]); DarwinPokeEQ(); } mieqEnqueue_unlock(); } Elsewhere, the devices are set up like this: InitPointerDeviceStruct((DevicePtr)pPointer, map, 7, GetMotionHistory, (PtrCtrlProcPtr)NoopDDA, GetMotionHistorySize(), 2); pPointer->name = strdup("Quartz Pointing Device"); InitPointerDeviceStruct((DevicePtr)pPointer, map, 3, GetMotionHistory, (PtrCtrlProcPtr)NoopDDA, GetMotionHistorySize(), 7); InitProximityClassDeviceStruct(pPointer); pPointer->name = strdup("pen"); Can anyone give me some advice on what I need to do to get gtk to recognize my input devices? Thanks! Ben

2 1

X11-2.2.1-rc2 and xorg-server-1.4.0-apple5
by Jeremy Huddleston 01 May '08

01 May '08

I've put together an updated X11 package (2.2.1-rc2). There are only a few changes over 2.2.1-rc1: XBell() now works (you can toggle between this doing NSBeep() and the "System Alert" sound effect in preferences Closing an X11 window will no longer cause you to jump to another space xinit now uses CFProcessPath instead of argv[0] trickery (thus Xquartz will show up as /usr/X11/bin/X in ps, top, etc) RECORD extension is disabled http://xquartz.macosforge.org/downloads/X11-2.2.1_rc2.pkg Assuming nobody reports any show stoppers, I'll be releasing this as 2.2.1 on 5/1. --- 1.4.0-apple5 incorporates all the fixes mentioned above plus: further cleanup of the startup code Ben committed some patches today targeting tablet support (although he has my tablet, so I haven't tested it) Some possible threading issues have been corrected, and people have reported that the _xp_async_dequeue server lockup is no longer occurring Additionally, multi-monitor support should be better with 1.4.0- apple5 than 1.4.0-apple4. Please test this out and let me know how it works for you. There is one known crash with this version. It is a threading issue that can only happen on server shutdown. I think I know a fix for it, but I haven't implemented it yet... still, please report anything you find. The next package release of X11 after 2.2.1 will most likely use the 1.4 branch, so it's imperative that this get some good testing exposure. http://xquartz.macosforge.org/downloads/Xquartz-1.4.0-apple5.bz2 (Install it after 2.2.1_rc2). --Jeremy

3 11