postinstall(8): Do obsolete _then_ opensslcertsrehash.
Otherwise any certificates removed by an update will remain in
/etc/openssl/certs/ca-certificates.crt, which is very bad!
PR bin/60932: postinstall obsolete + opensslcertsrehash interact badly
PR lib/60885 and PR lib/60886 - add test cases to demonstrate bugs
The changes in this test (amongst others, including adding tests of
the nb_NO.UTF-8 domain for PR bin/60496 which should already be fixed)
should expose the problems in printf(3) and scanf(3) noted in the PRs.
Note that we do not currently have tests for wscanf(3) which has a similar
problem to PR lib/60885, and consequently remains untested.
These new tests are currently expected to fail - but are not going to
be marked as expected fails, I want to see them actually fail.
Fixes for the issues already exist, and will be committed to the tree
soon after test failures have been observed. Similarly, a fix for the
wscanf(3) issue will be made at the same time (the bug is glaringly obvious,
and its fix is considerably simpler than that for scanf(3) - not that that
one was difficult) without that one ever having been observed - unless
the proverbial someone quickly implements a test case for that.
None of this is intended to be pulled up to any already released version
[11 lines not shown]
Add 3 new (bizarre) locales for ATF test use only.
These locales are inspired by issues reported in PR bin/60496 , PR lib/60677 ,
PR lib/60885 and PR lib/60886 (though while 60496 inspired all of this, it
does not need these locales to be reproduced).
Add the x1_X1 x1_X2 and X1_X3 locales (all .UTF-8) to be used to stress test
implementations with bizarre locale features that no real world locale would
ever use. Use of these locales for purposes other than implementation
testing is not intended, and getting weird results from any other use is to
be expected.
In the very near future, a test of the LC_NUMERIC categories of these locales
will appear (for 60496, 60885 & 60886) - that will depend upon the settings
of these locales' LC_NUMERIC parameters, so those values must not be altered
without also altering the tests.
In the not too distant future, a test of the LC_MONETARY categories of these
locales will also appear (for 60677) - and likewise will depend upon the
[37 lines not shown]
Fix a typo (share/locale/Makefile.locale)
While in an actual Makefile variable assigmnent, this value seems
not to be used (the LANGUAGE_XX variables appear to be never used),
so the typo was completely benign. Note that there is no "au" in
the LANGUAGE list.
Fix it for consistency, and so no-one else ever needs to wonder
what is going on here (for which the answer is "nothing").
pchtemp(4): Don't block suspend.
Don't think this keeps any interesting state or operations on it that
need to be quiesced while suspended.
PR port-amd64/60441: pchtemp(4) blocks suspend
Revert previous (x86/ALL: Enable nvmm(4)).
- i386 has no nvmm support (HAVE_NVMM=no) and fails with incomplete type.
- amd64 cannot have both nvmm and KASAN and fails pmap_ept_* functions
as KASAN clears __HAVE_DIRECT_MAP and therefore pmap_ept_* gets excluded.
Both kernels compile again.
Ok: Taylor R Campbell
atto(1): Initial import of tiny text editor, as discussed on
tech-userlevel@.
This is Hugh Barney's atto text editor modified for inclusion in
NetBSD. Since the upstream repository is inactive and the original
author wishes to keep the code base below 2000 lines of C, this is
effectively a fork that lives in usr.bin/atto. It serves as an
alternative to vi(1) for users more accustomed to modeless text
editors.
Notable changes include reworked keyboard input code to handle a wider
range of terminals (via libcurses cursor key translation), more useful
help for unrecognized key presses, and standardized option handling
using getopt (with syntax highlighting disabled by default).
Thanks martin and other folks for feedback and testing.
ATTO(1) General Commands Manual ATTO(1)
[115 lines not shown]
Additionally pull up following revision(s) (requested by riastradh in ticket #1380):
lib/libc/gen/pthread_atfork.c: revision 1.20
Specify MAP_ANON, the default (MAP_FILE) with a fd == -1 makes no sense.
Introduce debug function print_winregs() and clear_winregs().
These debug functions can be used to print out snapshots of the register
window.
The SAVE_WINREGS() macro in locore.s creates the snapshot.
Pull up following revision(s) (requested by yurix in ticket #561):
sys/arch/arm/ti/am3_prcm.c: revision 1.15
am335x: fix timer clock offset
Closes PR 60910
Fix by Elias Howell.
Pull up following revision(s) (requested by joe in ticket #1405):
sys/dev/pci/if_vmx.c: revision 1.19
fix wrong range from computing the MIN of tx and rx queues.
Pull up following revision(s) (requested by joe in ticket #560):
sys/dev/pci/if_vmx.c: revision 1.19
fix wrong range from computing the MIN of tx and rx queues.
