BSD Commit Log Search

   Document control ranges, from Dane Jensen.

   Set up default and last cell before clearing padding, reported by
   someone on GitHub.

   Do not read off end of buffer if it ends in \ when expanding \c, found
   by ossfuzz.

   Clamp width to terminal width, also change calculation of end of screen
   (it is OK to be outside the screen). Fixes problem reported by Dane
   Jensen in GitHub issue 4969.

   fix previous bound check in dname_expand()

   introduced mistakenly by me while reworking renaud's diff.  spotted the
   hard way by deraadt

   ok jtt@

   geo/qgis: update to 4.0.1

   update to py3-uharfbuzz-0.53.6

   update to tomcat-10.1.54

   update to tomcat-9.0.117

   update to py3-tzdata-2026.1

   update to py3-charset-normalizer-3.4.7

   update to fetchmail-6.6.3

   update to sslscan-2.2.2

   ksh: yank on 'change' in vi-mode

   Until now, only 'delete' operations in ksh's vi-mode yank the deleted
   range to the yank buffer. Make 'change' operations do the same.
   This is consistent with vi(1)'s behavior.

   ok tb@

   match on Intel 800 Series
   ok mlarkin@

   sys/octeon: interrupts and barier cleanup

   octciu_intr_establish() happens on autoconf where only CPU0 exists, and
   on octeon we increase ncpus in cpu hatch, and not attach, which means
   that ncpus is 1 at octciu_intr_establish() time.

   What makes cpuid always 1 for that code and schedules all interrupts on
   CPU0.

   I also moved octciu_intr_barrier() to actual interrupt's cpu.

   OK: visa@, kn@

   regen

   add Intel 800 Series ids

   from:
   Intel 800 Series Chipset Family Platform Controller Hub (PCH)
   Datasheet, Volume 1 of 2, Doc. No.: 833778, Rev.: 003

   ok mlarkin@

   freesynd uses C++ so must use ports-gcc on base-gcc arches

   This doesn't fix the build on sparc64, but it's necessary to
   get there

   Update to avahi-0.9rc4.

   update to scummvm 2026.2.0
   tinygl is enabled again after an upstream change to reduce memory usage

   Update to awscli-1.44.73.

   Update to py3-boto3-1.42.83.

   Update to py3-botocore-1.42.83.

   Update to OpenEXR-3.4.9.

   cdogs-sdl errors out with "multiple definition of `smat4_lerp'"

   Move to ports-gcc on gcc-arches to fix build on sparc64

   Fix order of cpus in CPU_INFO_FOREACH

   The cpus were put into the linked list in mostly inverted order during attach.
   The linked list determines the order used by the CPU_INFO_FOREACH macros.
   Order the linked list by ci_cpuid instead.

   ok jca@ for most archs
   "this should go in" deraadt@

   Update newsraft to 0.36.

   newsboat: update patch to reflect what was committed upstream.

   bluemoon: pass version number through MAKE_FLAGS.

   This removes the need to patch the upstream Makefile.

   Bump REVISION since the binary now reports the correct version; the
   hardcoded version in the patch had not been updated.