Updated authors and translations - #820

I also noticed that scons gettext adds msgid's to rmlint.pot but does not replace them if they are changed, the old ones (no longer existing essentially) are kept.

Thus:

"rmlint was written by Christopher Pahl"

will have a msgid

Changing that string to:

"rmlint was written by Christopher Pahl and somebody else"

will add a new msgid but keeps the previous one.

Deleting the pot first and then scons gettext seems to give better results also with the po files.
Or am I missing something?

I also noticed that scons gettext adds msgid's to rmlint.pot but does not replace them if they are changed, the old ones (no longer existing essentially) are kept.

Yes, this is because --join-existing. It creates smaller commits but drift over time. In 5caea12 I've added a way not to use it:
rm po/rmlint.pot and it gets regenerated. It's a good thing to do just before a release. It deletes old entries as well.

msgmerge on the other hands, will always add the string and marks it fuzzy. It's the translator job to merge with the old one.

For simple translation, you can modify the .po and the .pot directly, without the xgettext/msgmerge chain.

Thus:

"rmlint was written by Christopher Pahl"

will have a msgid

Changing that string to:

"rmlint was written by Christopher Pahl and somebody else"

will add a new msgid but keeps the previous one.

You can replace directly the previous one, it works as well.

Deleting the pot first and then scons gettext seems to give better results also with the po files. Or am I missing something?

Edit: I've just read your comment. Yes, without join-existing is cleaner.

I am checking the translation files so I have updated the PR name as getting the authors in check meant that the po files were also updated. Mostly cleaned - removed all #~ entries and recreated the pot.
Now cross checking if translations are properly retained as the po files are indeed rewritten. Seems fine so far. Will continue tomorrow.

#~

I haven't deleted those because they might be handy to future translators for reference, i.e. to adapt old translations. But there is git for history so I think you can deleted them.

If you do gh pr checkout 769 you get a commit with a full ja.po translation in 日本語 ：）
But I'm not sure if we could use @imoyan's work.

You're welcome to use the Japanese translation. However, the version in #769 was AI-generated. I've since reviewed and revised it myself, and I'll submit the updated ja.po in a separate PR.

You're welcome to use the Japanese translation. However, the version in #769 was AI-generated. I've since reviewed and revised it myself, and I'll submit the updated ja.po in a separate PR.

That's very much appreciated, imoyan.

Please note that we have done some (manual) work akin to your AI-changes:

• i18n: honour TEXTDOMAINDIR on debug #821 to test translations on the rmlint binary
• 5caea12 to make translations working on the GUI
• ec19532 (fixed by bbcdd68) to help testing the schema and the translations on the GUI
• ac2cf80 to fix the log levels

And some modernisation along the way.

@vassilit FreeBSD step fails with an error:

  pkg: No packages available to install matching 'py312-py-cpuinfo' have been found in the repositories
Error: ssh exited with code 1

@vassilit FreeBSD step fails with an error:

#832