BSD Commit Log Search

   update recoil-rts to latest release 2026.06.12. From fabien@ - thanks!

   Includes some other changes:
   - disable the new use of mimalloc (and use otto malloc instead)
   - fix a memory alignment bug related to upstream reliance on mimalloc;
     planned to be upstreamed
   - build type change from RELWITHDEBINFO to RELEASE
   - binary size reduction due to this and removal of headers for bundled
     static libs

   enable 64-bit DMA mbuf support for aq(4)

   tested on "Aquantia AQC107S" rev 0x02: F/W version 3.1.86

   from brad at comstyle.com, ok kettenis@

   dia-gtk2 uses C++ so must have ports-gcc to build on sparc64

   base-gcc errors out with
   wrong number of arguments specified for 'deprecated' attribute

   Move to ports-gcc to fix build on sparc64

   base-gcc errors with "No option -Wincompatible-pointer-types"

   Switch to ports-gcc to fix build on sparc64

   switch ports to gcc-16

   OK: tb@, sthen@

   telephony/libosip2: update to 5.3.2

   7.9-stable: update to got-0.127

   update to p5-NetAddr-MAC-1.01, from Mikolaj Kucharski (maintainer)
   (noop update, just lists Test::Exception as a dependency in the perl
   control files)

   Force the multiprocessing "fork" start method.  Python 3.14 changed the
   POSIX default to "forkserver" (cpython gh-84559), which breaks Salt's
   fork-based process model, setproctitle output and rc.d(8) pgrep matching.
   From Mikolaj Kucharski.

   net/irtt: regen & add modules.inc to fix build

   imsg_init.3: mark imsg_get as intentionally undocumented

   noticed by chohag at jtan.com
   ok claudio

   Use imsg_get_data for IMSG_CTL_LOG_VERBOSE instead of blindly trusting the
   size of the imsg.

   Reported by Shibo, Hugo Systopia Team

   Unify IMSG_CTL_LOG_VERBOSE handlers to look the same.

   redis/main port changes:

   - use cc instead of ld to link modules for tests; when ld is run directly
   this fails on at least i386 and clang+lld sparc64 with "undefined hidden
   symbol: __guard_local". ("ld -o commandfilter.so commandfilter.xo -shared"
   vs "cc -o commandfilter.so commandfilter.xo -shared")

   - add STRICT_ALIGNMENT patch for the internal copy of xxHash

   above ok tb@

   - drop pkg/MESSAGE referring to test problems, I'm not sure what was
   happening before (many tests were failing) but they seem in pretty good
   shape at least on i386/amd64 when I run them now. a few failures but
   don't seem particularly serious..

   use DIST_TUPLE rather than hand-rolling, ok bluhm

   adjust include order when building tests, so that internal headers are
   pulled in before system headers. fixes build of flavoured version if
   unflavoured is installed, as found in bulk builds when pulled in via
   p5-OPCUA-Open62541.  ok bluhm

   Fix inverted logic in bus_dmamap_load() and its variants that checks
   whether virtual address continuity is needed.  This fixes an issue where
   cache flushes during bus_dmamap_sync() would fault.

   ok jca@

   Fix inverted logic in bus_dmamap_load() and its variants that checks
   whether virtual address continuity is needed.  This fixes an issue where
   cache flushes during bus_dmamap_sync() would fault.

   ok jca@

   rename re-rolled distfile under the same name, pointed out by volker

   update to py3-tomlkit-0.15.1

   update to py3-sqlalchemy-2.0.51

   update to re2c-4.5.1

   Update to appstream-1.1.4.

   Update to libfyaml-0.9.6.

   update to dropbear-2026.93

   update to dropbear-2026.93

   Update to gemini-cli-0.52.0.

   update to py3-hatchling-1.31.0

   update to py3-build-1.5.1