Replacing all uses of LuaToAssEntry with LuaToTrackedAssEntry
also replaced its use in LuaParseKaraokeData, which would cause a double
free when canceling a script after calling parse_karaoke_data.
Instead, count how many consecutive times the entry has been found to be
unused and delete it once that count exceeds a limit. This will prevent
excessive reallocating of extradata ID's in applications like folding.
This wasn't necessary before since the internal representation of
folds is be checked for consistency after each commit, but after the
switch to extradata fold operations would leave the extradata in an
invalid state. This isn't technically a problem, but it does leave more
extradata entries lying around than necessary, and it can trip up
automation scripts that aren't prepared for inconsistent fold state.
These commands were revamped in 0ef9963 but the default hotkeys were
never updated. The hotkeys were automatically migrated, but resetting
the settings back to defaults would still set invalid settings.
This would cause an assertion failure in functions like lua_for_each
when the given closure throws an error and thus leaves some values on
the stack. This can make Aegisub crash entirely instead of just catching
and reporting the error. Instead, these stack_checks can be done
manually.
While meson already provides the aegisub-update-po target
for this exact purpose, it insists on also running meson’s
aegisub-pot target first, leading to missing translation strings.
This fixes a bunch of misspellings, borked C format strings and
increases translation coverage from 1288/1612 to 1516/1612
where still missing strings mostly don’t exists in wangqr’s build.
Original commit:
98abe53b15
Using make_pot.sh and
msgmerge --no-fuzzy-matching --sort-by-file $i.po aegisub.pot
since meson’s aegisub-pot target are still not covering all
translatable strings yet.
Base POSIX leaves the handling of backslahes in the operand
implementation defined. The XSI extension specifies several
escape sequences, like e.g. \n, which shall be transformed
upon printing.
Current make_pot.sh expects XSI behaviour and indeed e.g.
dash’s echo builtin implements this. echo builtins of other common
shells such as bash however, do not (by default). Avoid this portability
pitfall by just using printf at all relevant places. See:
https://pubs.opengroup.org/onlinepubs/9699919799/utilities/echo.htmlhttps://pubs.opengroup.org/onlinepubs/9699919799/utilities/printf.html
This also allows us to stop substituting each backslash with four
backslashes before piping to maybe_append. This substituion existed
to safely pass backslahes through maybe_append’s two layers of echo.
Also improve quoting. For consistency prefer single-quote quoting over
shell backslash escapes. QUote the unguarded *.lua to ensure it will
continue to work if there happen to be any lua files in the scripts
working directory.
This supersedes commits from wangqr/Aegisub which adjusted quoting and
echo usage to work with different shells. This is more throughout and
doesn't introduce semi-broken intermediate states.
0fbcaea8714aee271d03940181c7bc
(Note the second commit actually didn't manage to achieve the intended
portability, since several echo usages remained and it forgot to adjust
the backslash substitution, which was then fixed in the third commit.)
gettext can deal with AppData (and other XML files)
since version 0.19.6 (2015). We already depend on gettext anyway so
let’s utilise this, to get rid of two build dependencies and simplify
the extraction logic.
This depends on the removal of underscore prefixes
as done in a prior commit.
After this the developer name "Aegisub Group"
is now also treated as a translateable string.
gettext understands desktop files since 0.19 (2014),
so there’s no need for custom grep logic. This change
matches meson’s i18n which also uses gettext for this.
As a side efffect of this some long strings now get
broken into multiple lines (cosmetic only) and some
blank line gaps are normalised by gettext.
Cherry picks: 0e29c8d0e4
With fixups from: e98b7292f1
find may output matching files in any order.
To avoid unnecessary changes between regeneration
sort its output. For C++ sources xgettext’s --sort-by-file
already has an equivalent effect. All other files are already
statically listed in a fixed order.
Adopts part of: 4aee271d03
And document which aprts are still missing.
The keyword list is are copied from make_pot.sh.
Sorting by file helps keeping future PO(T) update diffs small'ish
and is also done in wangqr’s PO-file update procedure. (POT file
generation also sorts, but by preapplying `LC_ALL=C sort` to the
file list instead of using xgettext’s sorting option.)
Without this, meson’s aegisub-pot missed the majority of translatable
strings from C++ sources. While this brings the target much closerto
make_pot.sh, the remaining strings aren't straigthforward to integrate
unfortunately. At the time of writing meson has no support for iss files
and make_pot.sh uses custom logic to select only specific strings
from Lua and JSON files.
Even after this commit, POT updates should
thus still continue to use make_pot.sh.
It used to be genrated and installed by the old custom build system,
but apparently got lost in switching to meson.
Rename to .in.in instead of .template.in, since i18n otherwise gets
utterly confused failing to recognise the file type and both erroring
out with "no ITS rules found" during merge_file and failing to extract
any translatable strings.
Also drop underscores which were previously processed by intltool.
This doesn't yet automatically replace the release version and date
in the appdata file, but neither did the old build.
Note: while this slightly improves the situation, meson’s aegisub-pot
target is still missing many strings. make_pot.sh should be used.
meson’s i18n module get’s confused by .template and fails
to extract translatable strings, thus rename to in.in.
Note: while this slightly improves the situation, meson’s aegisub-pot
target is still missing many strings. make_pot.sh should be used.
a91c4f70c3 switched processing of the
desktop file from intltool to gettext-backed i18n.merge_file and dropped
leading underscores which used to be processed by intltool.
However, this commit omitted dropping the underscore from Keywords,
resulting in this field being stripped from the installed version.
Furthermore, make_pot.sh was broken since it too relied on the
underscore meaning POT regeneration will now remove the desktop files
strings from translation files.
Since it was not added to po/POTFILES meson’s native aegisub-pot target
also doesn't add desktop file strings to the translation pool. Though,
this target still misses a lot of other strings as well and shouldn't be
used yet anyway.
This commit removes leftover underscores and reintegrates desktop file
strings into both make_pot.sh. Integration into aegisub-pot is done in a
follow-up commit.
This consists of a meson option that
- sets a flag to make Aegisub read paths relative to the executable
- makes meson set up the symlinks like AppRun and .DirIcon as following
the AppDir format
and a CI workflow that builds with this option, installs to an AppDir
directory, and bundles it as an AppImage.
Modified from arch1t3cht/Aegisub#12 .
Co-authored-by: Fred Brennan <copypaste@kittens.ph>
