BSD Commit Log Search

   jdk-ca-tools: set HOMEPAGE so that the GH_ magic does not

   reported by naddy

   lang/gcc/16: enable algol68

   tweak ordering of make vars

   update to acpica-20260930

   update to uefitool a76

   backport upstream commit reducing a lot of noise (which is expected to
   become an error in future python versions) during doc generation, found
   while investigating build failure with new sphinx

   py3-sphinx: backout upstream commit "LaTeX: fix PDF build crash with
   June 2026 LaTeX release + (related to tables with 'colorrows')"
   (https://github.com/sphinx-doc/sphinx/commit/748a569f474f6295a95914d8729b6dd1aa02126b)
   which breaks pdflatex runs with our current version, seen in misc/open62541

   update to py3-sphinx_rtd_theme-3.1.0

   net/unifi: append to RUN_DEPENDS in Makefile.inc

   Using '=' instead of '+=' for RUN_DEPENDS in Makefile.inc overwrote previously
   defined dependencies, i.e., specifically databases/mongodb. Bump REVISION for
   main and v9 ports.

   OK sthen@

   update to tdlib 1.8.69

   update to openh264 2.7.0

   drop QTEST_DISABLE_KEYPAD_NAVIGATION from tests

   This is removed upstream and breaks build on Qt >= 6.12.0

   https://github.com/qgis/QGIS/commit/47ab0643e3c158476dd01dee13e7dff8eab59274

   Spotted by naddy, because of my Qt6 fallout

   add a BDEP on the new toluapp port, rather than a bit of an ugly workaround,
   which has issues with dpb junking of lua. ok martijn

   +toluapp

   Initial revision

   jdk-ca-certs: bump build tool. Only some minor polish.

   Backport 3f5b2882017f4e12c9b021de0ea24e6dd2eb26dd to unbreak Qt6 >= 6.12

   Spotted by naddy

   Correctly check whether CAs were previously confirmed

   The intention was to check using the state from the previous run, but
   this got lost in a refactor.

   Reported by Ties de Kock

   OK tb@

   fix build with Qt 6.12

   uic in Qt 6.12 rejects template class names like "AddressableItemList<>"
   in .ui files. Add a type alias and replace the template name
   in the affected .ui files.

   Spotted by naddy@

   Update cantata to 3.5.0 (Unbreak Qt6 >=6.12)

   jdk-ca-cert: switch to a custom conversion utility, replacing OpenSSL

   The format of jdk-cacerts produced by kirill's eopenssl40 contraption is
   a very simple PKCS#12 file (as far as anything involving PKCS#12 can be
   called simple) with an oracle-specific attribute in each of its certBags.

   Build a custom tool I wrote for this because serializing correct ASN.1
   by hand is way easier than using and extending the OpenSSL PKCS12 API,
   which is what we would need to do if we extended the openssl pkcs12 cli.
   It's also far less scary.

   ok kurt

   Prevent fq_codel panic from drop counter wrap

   Sustained traffic through a congested fq_codel queue can wrap the
   per-flow drop counter to zero, triggering the control_law() assertion
   and panicking the kernel. Saturate drops and decrement a nonzero ldrops
   to preserve the grace-period delta and prevent the wrap.

   Issue reported by Acts1631.

   OK sashan@, claudio@

   hidden dep on linux/input.h; found by naddy@

   update to yc-dyndns 0.5.1

   No m88k support

   Dropped upstream before gcc 4:
   https://github.com/gcc-mirror/gcc/commit/7c942e1747c951401736c0cfc144bc307d728b6c
   Pointed out by Brad

   Backport hppa fixes from lang/gcc/8

   - make __builtin_return_address() not try to read from text sections.
   (backport from base)
   - put switch() jump tables in .rodata in order not to read from text sections

   Fixes from miod@, reminded by Brad, ok kirill@

   bind locally and provide relayd.conf tls reverse proxy example

   bind locally and provide relayd.conf tls reverse proxy example

   update to patchelf 0.19.2

   Fix build with qt6 >=6.12.