This document intends to detail the change log for relatively recent work (roughly since the source code was tracked in Git).
This change log section represents git commits in range v2.8.4..HEAD (commits 541c2ecf0..bf64358b6).
--disable-threading [#3300] Can be useful for
systems with broken libraries, as well as for CI testing of
sequential implementation on build systems WITH threading support
available (and until now unavoidable then). This commit also fixes
problems found when trying such a build.
(lt_dlhandle)1 instead of (void *)1 [#3293]
nut-scanner tool section [#3293]
(lt_dlhandle)1 casting instead of (void *)1 also in comparisons
-isystem fix for glib
-Werror makes those
errors. I have used Gemini for some of these, but quota ran out
after half an hour and \~120 invocations so I completed the rest
manually.
git clean -fdX
ups.load,
input.voltage, output.voltage for e.g. Ugreen US3000 [#3281]
configure.ac, docs/acknowledgements.txt, docs/man/nut-scanner.txt, docs/nut.dict, m4/nut_check_libglib.m4, tools/nut-scanner/Makefile.am, tools/nut-scanner/nut-scan.h, tools/nut-scanner/nut-scanner.c, tools/nut-scanner/nutscan-device.c, tools/nut-scanner/nutscan-device.h, tools/nut-scanner/nutscan-init.c, tools/nut-scanner/nutscan-init.h, tools/nut-scanner/scan_upower.c: Add UPower support to nut-scanner
strcpy() usage [#3253]
@NUT_UPSSTATS_TEMPLATE@
magic token that HTML templates must start with [#3249, #3252]
\\?\c:\... paths to config files. TOTHINK: Maybe move this closer
to GetModuleName() calls?
localhost clients expected.
upsc -j JSON outputs [#3178]
sh without
assuming a static dir. Something like command -v sh
common/common.c: getfullpath(): avoid "DIRNAME../etc" kind of bugs
make check-parallel-builds.
[ '' -ge 0 ] ;
echo $? does not report a non-numeric error, but succeeds instead.
VAR=VAL cmd sets VAR=VAL in
the current shell and does not inherit it into cmd (nor reset it
after the call).
ups.model, causing the CGI to incorrectly report "Not supported".
comparison is always true due to
limited range of data type warning.
-j command-line flag to
upsc to provide machine-readable JSON output for all operating
modes. To support this and share logic with upsstats, the JSON
string escaping logic has been refactored into a new, reusable
module: - A new helper function json_print_esc() now lives in
common/strjson.c - common/Makefile.am builds this as a new
libcommonstrjson.la library. - clients/Makefile.am links both
upsc and upsstats against this new library and adds the
common/ include path. - upsstats.c is updated to use the shared
json_print_esc() function.
configure script would now probe the OS for more
user and group account names [#3173]
nobody or nogroup end up being
used for a build [#3173]
devices instead of upses.
pull_request_target has something to run from the master branch
at all). Even if this turns out to be the end of the road, at least
this way we can let people know where to find those generated
tarballs :)
git
describe fails [#1400]
make dist-sig (via
dist-files) in the CI job [#1400]
dist tarballs of recent iterations [#1400, #2829]
make
dist-docs [#1400]
NOTIFY_STATE_EXTEND_TIMEOUT option and relevant variables [#3003]
AC_REQUIRE() calls have priority over other
code lines in a method, and get executed first - so using extra
methods (also for clearer code base).
shutdown.* commands consistently [#1962]" This reverts
commit 15bc72b0ae7952fe30b013b48eebfd7a6c9bf0a7 prepared problem
log was not a problem after all.
shutdown.* commands consistently [#1962]
PIDPATH is now more
strictly /var/run, unless /run is more preferable for good
reason [#3099]
nutdrv_siemens-sitop(.8) (dash vs. underscore) to match the
driver program name [#1209, #545]
FSD notifications are
always issued at latest when shutdown handling just starts (or
earlier) [#3003, #3110]
SHUTDOWNEXIT setting [#3084, #2133]
SHUTDOWN_HOSTSYNC notification message [#3084]
doshutdown() to handle FSD.
variable 'start' may be uninitialized when used here
in difftime() later on.
upssched -l mode [#3097]