- Update to 1.8.9
  * Fix regressions introduced in 1.8.8 (!245, !248) - this includes
    reverting for now the previous "Fix XIM input sometimes jumbled
    (#198, !236)"
- supersedes
  * U_0001-xlibi18n-restore-parse_line1-for-WIN32-builds.patch
  * U_0002-Revert-imDefLkup-Commit-first-info-in-XimCommitInfo.patch
  * U_0003-Revert-ximcp-Unmark-to-fabricate-key-events-with-XKe.patch

- U_0001-xlibi18n-restore-parse_line1-for-WIN32-builds.patch
  U_0002-Revert-imDefLkup-Commit-first-info-in-XimCommitInfo.patch
  U_0003-Revert-ximcp-Unmark-to-fabricate-key-events-with-XKe.patch
  * fix regressions in 1.8.8 (issues #204, #205, #206, #207, #208)

• Files Changed
• Browse Source

- update to 1.8.8
 * Fix XIM input sometimes jumbled (#198, !236)
 * Fix _XkbReadGetDeviceInfoReply for nButtons == dev->buttons (!237)
 * Drop ifdefs for platforms that are no longer supported (!242, !243)
 * Assorted memory handling cleanups

• Files Changed
• Browse Source

• Files Changed
• Browse Source

- this update is needed due to jsc#PED-7282; it includes the
  security fix for CVE-2022-3555 (bsc#1204425, bsc#1208881) and
  a fix for a race condition in libX11 that causes various
  applications to crash randomly (boo#1181963)

• Files Changed
• Browse Source

- update to 1.8.7
  This release contains fixes for the issues reported in security
  advisory here: 
    https://lists.x.org/archives/xorg-announce/2023-October/003424.html
  * fixes CVE-2023-43785 libX11: out-of-bounds memory access in 
    _XkbReadKeySyms() (boo#1215683)
  * fixes CVE-2023-43786 libX11: stack exhaustion from infinite recursion
   in PutSubImage() (boo#1215684)
  * fixes CVE-2023-43787 libX11: integer overflow in XCreateImage()
    leading to a heap overflow (boo#1215685)
  along with:
  * Fail XOpenDisplay() if server-provided default visual is invalid (!233)
  * Bring XKB docs in line with actual implementation (!231, !228)
  * Xutil.h: declare XEmptyRegion() and XEqualRegion() as Bool (!225)
  * Assorted updates to en_US.UTF-8 compose keys (!213, !214, !215, !216,
    !217, !219, !220, !222, !223, !226, !227, !229)

• Files Changed
• Browse Source

• Files Changed
• Browse Source

- U_InitExt.c-Add-bounds-checks-for-extension-request-ev.patch
  * Buffer overflows in InitExt.c (boo#1212102, CVE-2023-3138)

• Files Changed
• Browse Source

- Update to version 1.8.5
  * gitlab CI: Add libtool to required packages
  * configure: raise minimum autoconf requirement to 2.70
  * configure: replace deprecated AC_HELP_STRING with AS_HELP_STRING
  * configure: Use LT_INIT from libtool 2 instead of deprecated AC_PROG_LIBTOOL
  * gitlab CI: add workflow rules
  * nls: delete compose sequences that pointlessly mix upper and lower case
  * nls: remove four hundred and sixty untypable Greek compose sequences
  * nls: remove twenty two untypable Greek compose sequences
  * XSetScreenSaver.man: restore the part that was accidentally snipped
  * nls: make the Amharic compose sequences use the dead-vowel symbols
  * nls: sort three sequences alphabetically in their group, like all others
  * nls: delete six compose sequences that cannot be typed
  * nls: use a slash instead of a combining solidus in compose sequences
  * NLS: move long S compositions to respective blocks
  * NLS: implement the expansion of the six Breton N-graph keysyms
  * NLS: move dead-caron subscript compositions to the relevant Unicode block
  * NLS: Remove strange dead_cedilla cedi sign sequences
  * nls: add compose sequence for capital schwa, and delete a deviant one
- Users of the Amharic (am_ET.UTF-8) compose key sequences provided by libX11
  will also want to upgrade to xkeyboard-config 2.39 (releasing soon), in order
  to keep those sequeunces working with this release.

• Files Changed
• Browse Source

- Update to version 1.8.4
  This release fixes the regressions in previous 1.8.x related to the thread-
  -safety-constructor option. (boo#1209176)
- supersedes U_fix-a-memory-leak-in-XRegisterIMInstantiateCallback.patch

• Files Changed
• Browse Source

- Update to version 1.8.1
  This release fixes the --enable-thread-safety-constructor option to the
  configure script to work as intended.  In the previous release, the changes
  for this option may not have been enabled when the option was not specified
  or when the --enable option was specified.
  While we have enabled it by default, believing that doing so will reduce
  the number of bugs users encounter running libX11 clients, in some cases
  it may expose bugs in which clients had previously gotten away with calling
  libX11 functions while a libX11 lock is already held, and thus now deadlock,
  as discussed in https://gitlab.freedesktop.org/xorg/lib/libx11/-/issues/157  
- let's hope this version doesn't suffer yet from the regressions
  reported in boo#1205778, boo#1205818 (reported against 1.8.2);
  we need libX11 thread safe for totem (GNOME 43) :-(

- going back to version 1.7.5 for now to get rid of regressions,
  which were introduced by trying to get thread-safe in libX11
  itself 
- re-introduced U_fix-a-memory-leak-in-XRegisterIMInstantiateCallback.patch
  which was not yet in 1.7.5
- supersedes the following patches
  * U_0001-Add-XFreeThreads-function.patch
  * U_0002-Don-t-use-pragma-inside-a-function-it-breaks-compili.patch
  * U_0003-Fix-797755-Allow-X-IfEvent-to-reenter-libX11.patch
  * U_0004-Indentation-fixes-around-recent-dpy-in_ifevent-chang.patch
  * U_0005-ChkIfEv.c-fix-wrong-handling-of-dpy-in_ifevent.patch

• Files Changed
• Browse Source

- U_0001-Add-XFreeThreads-function.patch
  U_0002-Don-t-use-pragma-inside-a-function-it-breaks-compili.patch
  U_0003-Fix-797755-Allow-X-IfEvent-to-reenter-libX11.patch
  U_0004-Indentation-fixes-around-recent-dpy-in_ifevent-chang.patch
  U_0005-ChkIfEv.c-fix-wrong-handling-of-dpy-in_ifevent.patch
  * adding all patches since 1.8.2 release in order to try fixing
    regressions after introducing thread safety constructor with
    1.8.1 (boo#1205778, boo#1205818)
- supersedes U_Fix-797755-Allow-X-IfEvent-to-reenter-libX11.patch
- re-enabled thread safe constructor

• Files Changed
• Browse Source

- back to "--disable-thread-safety-constructor" for now; we see just
  too many regressions, e.g. firefox freezes and crashes, crashes with
  barrierc, crashes in Godot, assertions with vkquake (boo#1205818,
  boo#1205778)

• Files Changed
• Browse Source

- U_Fix-797755-Allow-X-IfEvent-to-reenter-libX11.patch
  * fixed Firefox freezes (regression since 1.8.2) (boo#1205778)

• Files Changed
• Browse Source

- Update to version 1.8.2
  * This is primarily a bug fix release, including further work on
    improving the thread-safety-constructor and making it work with
    software which had incorrectly called libX11 functions from
    inside X*IfEvent() calls.
- supersedes U_fix-a-memory-leak-in-XRegisterIMInstantiateCallback.patch

• Files Changed
• Browse Source

- U_fix-a-memory-leak-in-XRegisterIMInstantiateCallback.patch
  * security update for CVE-2022-3554 (bsc#1204422)

• Files Changed
• Browse Source

• Files Changed
• Browse Source

osc copypac from project:openSUSE:Factory package:libX11 revision:36, using expand

• Files Changed
• Browse Source

- Update to version 1.8.1
  This release fixes the --enable-thread-safety-constructor option to the
  configure script to work as intended.  In the previous release, the changes
  for this option may not have been enabled when the option was not specified
  or when the --enable option was specified.
  While we have enabled it by default, believing that doing so will reduce
  the number of bugs users encounter running libX11 clients, in some cases
  it may expose bugs in which clients had previously gotten away with calling
  libX11 functions while a libX11 lock is already held, and thus now deadlock,
  as discussed in https://gitlab.freedesktop.org/xorg/lib/libx11/-/issues/157 .

• Files Changed
• Browse Source

- Update to version 1.8
  * The highlight of this release is that we now try to initialize
    thread safety ourselves, rather than hope the application does it.
    This should resolve a number of long-standing bugs with the libxcb
    integration, since the socket handoff mechanism essentially has to
    be thread-safe.

• Files Changed
• Browse Source

- Update to version 1.7.4
  * Don't try to destroy NULL condition variables 

- Update to version 1.7.4
  * bugfix release 
- supersedes p_khmer-compose.diff

• Files Changed
• Browse Source