BSD Commit Log Search

   Update to lowdown 3.2.1 from maintainer

   also build sleuthkit java bindings, to be used in upcoming autopsy port.

   lots of feedback esp. wrt SUBPACKAGE vs. no SUBPACKAGE from sthen@
   java related feedback and final OK kurt@

   remove TDEPs which aren't referenced at all in the distfile or
   upstream repo (unbreaking sqlports after axolotl removal):

   devel/py-test-rerunfailures
   security/py-axolotl
   security/py-axolotl-curve25519

   Abort enhanced capability loop if reading all ones from the device instead
   of looping infinitely; this value isn't allowed by the spec, but there are
   devices behaving that way, from Intel no less.
   ok deraadt@

   Remove unused bridge_tunneluntag().
   ok bluhm@

   Remove global prototype for bufcache_getcleanbuf(), and only compile it
   #ifdef HIBERNATE, for it is not used anywhere else.
   ok deraadt@

   Remove unused cd9660_mmap() and cd9660_seek().
   ok tb@

   Remove unused rc4_getbytes() and Blowfish_expandstate().
   ok tb@

   Remove unused kernel global utsname.
   tweak&ok dgl@

   Comment out zlibVersion() and zlibCompileFlags(), unused in the kernel.
   ok tb@

   Update to py3-libvirt-12.8.0.

   Update to libvirt-12.8.0.

   Update to meld-3.24.1.

   Drop maintainer.

   Update to gthumb-4.0.

   Update to rclone-1.75.2

   Addresses multiple security vulnerabilities across protocol
   implementations, core dependencies, and service interfaces. Changes:
   https://rclone.org/changelog/#v1-75-2-2026-10-09

   Update to rclone-1.75.2

   Addresses multiple security vulnerabilities across protocol
   implementations, core dependencies, and service interfaces. Changes:
   https://rclone.org/changelog/#v1-75-2-2026-10-09

   vmd(8): add extra mmio region

   this adds a no-op (reads return FFs, writes discarded) mmio region between the
   end of guest RAM and 4GB. specifically recent i686 linux bootloaders/kernels do
   what appear to be errant probes of high PAs in the > 0xc0000000 range, and
   vmd treats all mmio accesses to unmapped regions as fatal.

   this diff adds a region to cover any gap, such that accesses to those addresses
   trap to vmd, which then does what real hw would do. other hypervisors do
   similar things.

   this allieviates the need to assign 4GB ram to these VMs which was previously
   needed to work around the problem.

   ok dv

   wayland/mango: Update to 0.18.0

   I am pretty sure two unveil() and pledge "stdio rpath wpath inet dns"
   will work here.  commiting to collect feedback.

   Use O_NOFOLLOW for the binding dir open

   use the err.h family of functions

   use SOCK_STREAM instead of fcntl O_NONBLOCK

   unveil() and pledge() are pretty easy to add

   use accept4() with SOCK_NONBLOCK rather than fcntl O_NONBLOCK
   ok ratchov

   Update to libical-4.0.6.

   Update to stellarium-26.3.

   devel/codex: update to 0.162.0

   Update to libsoup3-3.8.0.

   Bugfix update to glycin-2.2.2.