Re: [PULL] Input changes in preparation for multitouch
On 12/08/2011 07:04 PM, Keith Packard wrote: * PGP Signed by an unknown key On Fri, 9 Dec 2011 10:50:33 +1000, Peter Huttererpeter.hutte...@who-t.net wrote: dix: switch the syncEvent queue to a struct list This exposes PlayReleasedEvents as a global function (for testing), which seems odd, but I guess it's mostly harmless? It also causes a server segfault in XTS's XGrabKeyboard-3 test: Program received signal SIGSEGV, Segmentation fault. 0x00522a22 in EnqueueEvent (ev=0x7fffd530, device=0xa0ebe0) at events.c:1166 1166(tail-event-any.type == ET_Motion) (gdb) bt #0 0x00522a22 in EnqueueEvent (ev=0x7fffd530, device=0xa0ebe0) at events.c:1166 #1 0x005647e4 in mieqProcessDeviceEvent (dev=0xab7930, event=0x7e45c0, screen=0x8b8c80) at mieq.c:526 #2 0x005649ff in mieqProcessInputEvents () at mieq.c:586 #3 0x00419824 in ProcessInputEvents () at InitInput.c:56 #4 0x0050fafb in Dispatch () at dispatch.c:398 #5 0x00500ef9 in main (argc=2, argv=0x7fffe1f8, envp=0x7fffe210) at main.c:287 (gdb) p tail-event $1 = (InternalEvent *) 0x0 -- Aaron --- This email message is for the sole use of the intended recipient(s) and may contain confidential information. Any unauthorized review, use, disclosure or distribution is prohibited. If you are not the intended recipient, please contact the sender by reply email and destroy all copies of the original message. --- ___ xorg-devel@lists.x.org: X.Org development Archives: http://lists.x.org/archives/xorg-devel Info: http://lists.x.org/mailman/listinfo/xorg-devel
[PULL] Input changes in preparation for multitouch
A whole bunch of basic changes required for the multitouch support. Does not include any actual MT stuff yet though. Commit cba34d2a7e72ea0c7e5a2d8482607157ceccf8db The following changes since commit fb22a408c69a84f81905147de9e82cf66ffb6eb2: Save major/minor opcodes in ClientRec for RecordAReply (2011-12-01 15:11:20 +) are available in the git repository at: git://people.freedesktop.org/~whot/xserver for-keith Peter Hutterer (22): include: add list_append() include: add list_last_entry to get the tail of a list include: point to the tests in list.c as examples include: a new list element does not need initialization, state so include: update struct list documentation to use one set of structs only include: rename foos to list_of_foos in the struct list documentation dix: switch the syncEvent queue to a struct list dix: add AllocGrab helper function dix: add CopyGrab() function dix: allocate temporary grabs on the heap dix: switch the dev-deviceGrab.activeGrab from GrabRec to GrabPtr Add a new XI2Mask struct and a few helper functions. input: swap the server over to use the XI2mask struct xfree86: bump the input ABI dix: use BUG_WARN for input mask size issues Xi: when removing a device, reset ClientPointers where needed xfixes: don't dereference a NULL cursor Xi: skip superfluous cast Xi: check button mapping value _before_ assigning it Xi: rename state to corestate in ProcessDeviceEvents xfree86: include xorg-config.h from xaalocal.h include: add BUG_WARN_MSG for custom error message on bug condition Xi/exevents.c | 88 +++-- Xi/ungrdevb.c | 32 +++-- Xi/ungrdevk.c | 31 +++-- Xi/xichangehierarchy.c | 15 ++ Xi/xigrabdev.c | 15 ++- Xi/xipassivegrab.c | 50 +--- Xi/xiselectev.c|5 +- dix/devices.c |3 + dix/events.c | 288 dix/grabs.c| 84 +++- dix/inpututils.c | 181 + hw/xfree86/common/xf86Module.h |2 +- hw/xfree86/xaa/xaalocal.h |4 + include/dix.h |1 + include/dixgrabs.h |4 + include/input.h|1 + include/inputstr.h | 22 ++- include/inpututils.h | 12 ++ include/list.h | 81 include/misc.h |9 +- test/input.c | 65 + test/list.c| 37 + test/xi2/Makefile.am |5 +- test/xi2/xi2.c | 129 ++ xfixes/cursor.c|4 +- 25 files changed, 892 insertions(+), 276 deletions(-) create mode 100644 test/xi2/xi2.c pgpJORcnDqmCe.pgp Description: PGP signature ___ xorg-devel@lists.x.org: X.Org development Archives: http://lists.x.org/archives/xorg-devel Info: http://lists.x.org/mailman/listinfo/xorg-devel
Re: [PULL] Input changes in preparation for multitouch
On Fri, 9 Dec 2011 10:50:33 +1000, Peter Hutterer peter.hutte...@who-t.net wrote: dix: switch the syncEvent queue to a struct list This exposes PlayReleasedEvents as a global function (for testing), which seems odd, but I guess it's mostly harmless? dix: add CopyGrab() function This one says that it introduces a memory leak that will be fixed when callers are switched to use FreeGrab, but I don't see any patch that fixes the leak. That doesn't seem like a good plan to me... -- keith.pack...@intel.com pgpI5HxObVqTA.pgp Description: PGP signature ___ xorg-devel@lists.x.org: X.Org development Archives: http://lists.x.org/archives/xorg-devel Info: http://lists.x.org/mailman/listinfo/xorg-devel
Re: [PULL] Input changes in preparation for multitouch
On Thu, Dec 08, 2011 at 07:04:05PM -0800, Keith Packard wrote: On Fri, 9 Dec 2011 10:50:33 +1000, Peter Hutterer peter.hutte...@who-t.net wrote: dix: switch the syncEvent queue to a struct list This exposes PlayReleasedEvents as a global function (for testing), which seems odd, but I guess it's mostly harmless? Yeah, no-one else but the tests need this. I doubt anyone will call this from a module and IIRC if it's not marked _X_EXPORT it won't be usable from modules anyway. If you have a good idea on how to mark functions as static if it wasn't for a test, that'd be great. dix: add CopyGrab() function This one says that it introduces a memory leak that will be fixed when callers are switched to use FreeGrab, but I don't see any patch that fixes the leak. That doesn't seem like a good plan to me... dix: allocate temporary grabs on the heap does that. I ran this through valgrind and it didn't complain, iirc. Cheers, Peter ___ xorg-devel@lists.x.org: X.Org development Archives: http://lists.x.org/archives/xorg-devel Info: http://lists.x.org/mailman/listinfo/xorg-devel
Re: [PULL] Input changes in preparation for multitouch
On Fri, 9 Dec 2011 13:33:33 +1000, Peter Hutterer peter.hutte...@who-t.net wrote: dix: allocate temporary grabs on the heap does that. I ran this through valgrind and it didn't complain, iirc. With that understood, and with the segfault in 'make check' fixed that we discussed on IRC (a bug in the test), this series has been merged. 6369b59..522f8bc master - master -- keith.pack...@intel.com pgpeW43FZ0yPP.pgp Description: PGP signature ___ xorg-devel@lists.x.org: X.Org development Archives: http://lists.x.org/archives/xorg-devel Info: http://lists.x.org/mailman/listinfo/xorg-devel