Best workflow to track changes in original strings?

Before I launched my forum I took care that all user-facing texts (not moderator/admin) were translated to Slovenian.

But now after several months and regular updates I have hard time to track all the changes to original english text in order to keep having everything translated, as some translations seems to get lost due to change to English original texts.

For example, I have noticed that we “lost” Slovenian transaltion for discourse_narrative_bot.new_user_narrative.hello.message text:

Why (and when) is translation removed from existing key (and moved to History)?
How can I (as translator) better track and fix such regressions? Can I get notified about such changes before next Beta release to fix this before release?
How do other translators handle such changes?


That happens when the English translation changes. You can filter for untranslated strings on Transifex.

Transifex can notify you via email when the English source files are updated.


@gerhard, I know about translated:no filter, but then I get many strings as most of admin interface is not translated yet. I was looking to get only the ones that had translation and it was removed due to source change.

For now, I will try to identify them by using:
source_updated_before:2019-06-01 translated:no

I will put the date of last translation job I did, to find changes/overrides since then…