Closed TheOneric closed 1 year ago
Found a way to make aegisub.appdata.xml
build with i18n
afterall and also discovered the existing .desktop
integration is borked. The latter also affects the POT update here, so converted to draft. I’ll fix this up and add the desktop and appdata stuff to this PR later.
Fixed desktop file string extraction and added meson integration for desktop and appdata files + updating obsolete links in aegisub.appdata.xml
. Should be ready for review.
Remaining future improvements for someone more familiar with meson i18n details and the build setup:
packages/desktop/aegisub.appdata.in.in
somehow wire up remaining translatable string sources to meson’s aegisub-pot
target, so it can supersede make_pot.sh
. This will require custom logic. See comment in po/meson.build
, commit messages and make_pot.sh
for details.
Force-pushed to amend a commit message and added a one-liner fix for brazilian portugese from wangqr’s fork. Checking against portugesian Portugese and a dictionary it seems correct. Afaict this is the only other translation addition or fix from wangqr/Aegisub not present here yet.
I just remembered there were some make_pot.sh
changes not present here though. If make_pot.sh
isn't replaced by full meson integration soon, it might be a good idea to pick them up.
Thanks a lot! This mostly LGTM except for a few corrections on the added German translations.
I did also update some of the bestsource messages (actually this was long overdue since some of them were carried over from back when it was still a proof-of-concept and never updated) in 0c057ebddb1b28780a1d70bbc9f6d6d48a25f554, though I still kept some of them. I replaced "Exacting" with "Indexing" and "Finding Keyframes and Timecodes..." with "Reading timecodes and frame/sample data" (this isn't strictly accurate since bestsource doesn't read samples here, but it probably still makes the most sense for the user), but I kept "Scanning" as it is since that is a separate step that (for now) needs to run every time the file is opened, as opposed to the Indexing step that only needs to happen once per file.
Thanks for the review! Translations are now fixed as suggested and rebased for the new BS source strings.
To make backporting to wangqr easier I split fixing existing translations from wangqr into a separate commit. As a bonus I found and fixed another mistake in the old 2013 translations (Wahlen
→ Wählen
).
New version now also includes other make_pot.sh
improvements from wangqr and some new additional improvements. Some noteworthy points:
Aegisub group
as a translateable string. I’m unsure if this should be treated as a proper noun and left as is, or localised. (Aegisub Group
does suggest a higher degree of organisation and unity then actually exists in reality.)msgmerge
would reverse the order of commented strings. wangqr’s version of make_pot.sh
always runs msgmerge
twice for each updated language to workaround this issue. However, this workaround was added when there was no released version with a fix yet. At this point we might not really need to add this workaround here as well.Future improvements not in scope of this PR:
packages/desktop/aegisub.appdata.xml.in
and po/aegisub.pot
make_pot.sh
New changes which can be backported to wangqr/Aegisub afterwards:
Thanks again! There was one more missing space in the VapourSynth description text, but I just fixed that myself.
The current meson integration for POT generation is incomplete and misses out a lot of strings. The last commit improves this for strings from C++ source files, but strings from other sources are still missing. For
*.xml
and*.desktop
adding support might be as simple as renaming to.in.in
instead of.template.in
and dropping any extra underscores added forintltool
.ReFor the remaining formats I have no idea what the best way to integrate them is.packages/desktop/aegisub.appdata.xml.template.in
also see [issue creation pending].For this PR’s POT and PO update, I therefore stuck with
make_pot.sh
andmsgmerge
. If you’d prefer keeping the changeset down, I can omit rebasing non-de
PO files to the new POT version.Cosmetic nit: it might be a good idea to unify the (original) message texts for
ffms2
andbestsource
’s keyframe and timecode indexing/scanning dialogue.