cad/openscad-devel: Update and enable experimental features
Update to 2026.10.23 and enable experimental features
PR: 299642
Reported by: Arthur Chance
devel/py-evdev: Use the evdev headers from base
evdev 1.9 expects the UI_FF_* constants from uinput.h, but the port
generated its ecodes module only from devel/evdev-proto's
linux/input.h and linux/input-event-codes.h, so "import evdev" fails
with KeyError: 'UI_FF'.
files/patch-v4l_compat, added with the port in 2015 when base had no
dev/evdev headers, removes upstream's #ifdef __FreeBSD__ branches for
<dev/evdev/...> and makes the port build against evdev-proto instead.
Base now installs dev/evdev/input.h, input-event-codes.h and uinput.h,
which evdev's setup.py uses on FreeBSD by default, and build_ext
generates the ecodes module itself. Drop the patch, the evdev-proto
dependency, USES=localbase and the build_ecodes, --evdev-headers and
--include-dirs arguments, and bump PORTREVISION.
The 14.5 and 15.1 base headers define FF_MAX_EFFECTS; older ones
(13.5, 14.2, 14.3, 14.4 and 15.0) lack it, and evdev falls back for it.
[3 lines not shown]
*/wordpress*: deprecate localized variants
Localized WordPress tarballs are regenerated in place by upstream
whenever translations are updated, so the distfile checksums keep
breaking without a version change. The language packs have the same
problem, and other distributions only ship the English core and let
WordPress download language packs on its own.
Mark the localized ports deprecated, expiring on 2026-12-31. Users
should switch to www/wordpress and choose the site language in
WordPress.
{french,german,russian}/wordpress: chase upstream distfile re-roll
Upstream re-rolled the 7.1.3 de_DE, fr_FR and ru_RU tarballs again
with refreshed translations under wp-content/languages. The
WordPress core files are unchanged.
Update distinfo and bump PORTREVISION as package contents changed.
arm64 linux: Allow FUTEX_REQUEUE on arm64 by setting the brand flag
Better change is to provide the dedicated musl brandinfo, as it is done
on amd64/i386, which should be done later.
PR: 299836
MFC after: 1 week
linux: correct errors from mremap(2)
Linux returns EFAULT when the range to be remapped is not mapped at all,
and ENOMEM only when it is mapped but cannot be grown in place. Callers
depend on that distinction: musl's pthread_getattr_np() measures the
main thread stack by growing a one page mapping and retries while errno
is ENOMEM. Returning ENOMEM for every growth request makes that loop
spin forever.
PR: 299835
MFC after: 1 week
net/glib-networking: fix build with LibreSSL
LibreSSL does not have SSL_R_SHUTDOWN_WHILE_IN_INIT and
SSL_CTX_set1_sigalgs_list() yet. Upstream recommends local patches
for this as they strictly support OpenSSL 1.1 and later.
https://gitlab.gnome.org/GNOME/glib-networking/-/merge_requests/287
