BSD Commit Log Search

   Forcefully disable RVV code in highway to let lang/node build on riscv64

   Bump REVISION just in case, I strongly doubt this built with clang 19.

   ok volker@ (maintainer)

   Drop maintainer. I am barely using these nowadays.
   Any taker, feel free.

   pyside6/*: switch to MODCLANG_VERSION = 22

   ok jtt

   pyside6/tools: remove MODCLANG_LIB_DEPENDS leftover

   Since the port no longer uses the clang module, this LIB_DEPENDS does
   nothing and is confusing.

   ok jtt

   devel/kdevelop: switch to MODCLANG_VERSION = 22

   ok rsadowski (who had and tested the same diff)

   lang/clazy: switch to MODCLANG_VERSION=22

   ok rsadowski

   qt6/qttools: switch to MODCLANG_VERSION=22

   Adjust qdoc supported clang version and pull in upstream commit
   https://github.com/qt/qttools/commit/f75f4090ed8e5c9dc898a313a5dd9b6878c2f034

   ok rsadowski

   qt5/qttools: switch to MODCLANG_VERSION=22

   ok rsadowski

   Update to audacious-4.6.1

   Changes: https://audacious-media-player.org/news/65-audacious-4-6-1-released

   x11/yad: use proper object type-casting macro to appease llvm22

   Fails to build on riscv64

   sysutils/grafana: update to 13.0.2

   drop devel/liboil; no longer actively developed or used by other ports,
   and fails on i386 llvm22. upstream recommends that any remaining users
   switch to Orc:
   https://cgit.freedesktop.org/liboil/commit/?id=04b154aa118c0fdf244932dadc3d085f6290db7a

   print/lilypond: fix build with llvm 22.

   update python to 3.14.5

   ok kmos@ (MAINTAINER), tb@

   update matplotlib-inline to 0.2.2

   downgrade errors to warnings so clonekeen can build with llvm 22

   highway: fix build on sparc64

   from jca, based on an equivalent, more complicated, diff i sent

   sysutils/moor: Update to 2.14.0

   From Maintainer Lydia Sobot, thanks

   Ignore type mismatch in riscv-specific code, fixes the build on riscv64

   clang-22 considers pointer mismatches between unsigned long * and
   unsigned long long * as fatal errors. Make the errors non-fatal until
   I nuderstand what it takes to appease the compiler. Said errors
   shouldn't matter anyway given that said types are always the same size
   on riscv64 anyway.  make test agrees.

   Some performance increases are visible for eg chacha or AES on this
   Spacemit k1, but more efforts are needed to unlock performance
   improvements gated behind riscv_hwprobe checks.

   net/jnettop: delete function call with type error and unused result

   Fixes the build with llvm22.

   wayland/mango: update to 0.14.2

   update to freeradius-3.2.10

   update to freeradius-3.2.10

   update to rspamd-4.1.0

   update to py3-uvicorn-0.49.0

   update to py3-sh-2.2.4

   update to dnsmasq-2.93, from Brad (maintainer)

   update to dnsmasq-2.93, from Brad (maintainer)

   x11/Xdialog: update gtk+2 patch from AUR to fix build with llvm22