Add sysutils/wimlib
wimlib is a C library for creating, modifying, extracting, and mounting
files in the Windows Imaging Format (WIM files). wimlib and its
command-line frontend 'wimlib-imagex' provide a free and cross-platform
alternative to Microsoft's WIMGAPI, ImageX, and DISM.
jj: update to 0.33.0.
## [0.33.0] - 2025-09-03
### Release highlights
* `jj undo` is now *sequential*: invoking it multiple times in sequence
repeatedly undoes actions in the operation log. Previously, `jj undo` would
only undo *the most recent* operation in the operation log. As a result, a new
`jj redo` command has been added.
* Experimental support for improving query performance over filesets and file
queries (like `jj log path/to/file.txt`) has been added. This is not enabled
by default. To enable this, you must use the `jj debug index-changed-paths`
command.
### Breaking changes
* `jj evolog` templates now accept `CommitEvolutionEntry` as context type. To
[116 lines not shown]
expat: update to 2.7.2.
Release 2.7.2 Tue September 16 2025
Security fixes:
#1018 #1034 CVE-2025-59375 -- Disallow use of disproportional amounts of
dynamic memory from within an Expat parser (e.g. previously
a ~250 KiB sized document was able to cause allocation of
~800 MiB from the heap, i.e. an "amplification" of factor
~3,300); once a threshold (that defaults to 64 MiB) is
reached, a maximum amplification factor (that defaults to
100.0) is enforced, and violating documents are rejected
with an out-of-memory error.
There are two new API functions to fine-tune this new
behavior:
- XML_SetAllocTrackerActivationThreshold
- XML_SetAllocTrackerMaximumAmplification .
If you ever need to increase these defaults for non-attack
XML payload, please file a bug report with libexpat.
There is also a new environment variable
[47 lines not shown]
arm-none-eabi-gdb: restore accidentally dropped options.mk ref
(I've no idea how this happened, as my tree was supposed to be up to
date, but was missing r. 1.10 of Makefile but not the other changes
from that commit.)
sysutils/ups-nut: Recover from wip distfile confusion and fix asciidoc woes
I have been creating distfiles from upstream git for testing via
wip/ups-but, and I believe I got them crossed with real distfiles
before setting a DIST_SUBDIR for wip. There's also a chance upstream
changed it, but so far I don't have believable evidence for that --
but I've raised the issue. I did diff what I had and upstream's, and
the only difference is believable CI stuff that does appear in the
v2.8.4 git tag. This commit adjusts distinfo so it works with the
distfile from upstream.
Upstream's logic for dealing with docs is too complicated. Man pages
are pregenerated in distfiles, but not html. It's too hard and too
confusing to turn off upstream's detection of asciidoc and building of
html, so this commit adds TOOL_DEPENDS on asciidoc to get consistent
doc output. asciidoc isn't that big and it's only TOOL, so that seems
better than fighting upstream.
Probably this resolves various reports I've gotten that didn't make
[2 lines not shown]
unicorn py-unicorn: updated to 2.1.4
2.1.4
bindings/zig: Fix sample_riscv_zig partial writes and logging
Fix physical address truncation on 32-bit systems with addressing extensions
refactor(lib): mark pointers as const where possible
bindings: ruby: fix unexpected uc_query result pointer type
bindings: ruby: fix version identifier to 2.1.3
feat(arm): add an ESR register
fix(rust): watch all source and header files
feat(rust): improve ARM CP register ergonomics
fix(m68k): correct SR register read
fix(python): catch BaseException in wrappers instead of Exception
Remove the size limit for memory read and write (revamped)
Loongarch port
S390x registers
Add from_handle_with_data method
Make Unicorn struct clone-able
[20 lines not shown]