kempelen
Some updates to my antix-faq-po-maintainer
– Added support for adding Translator Comments
https://gitlab.com/kempelen/antix-faq-po-maintainer/-/blob/master/po/pot-comments.json?ref_type=heads
– Added a summary of the approach where a translator may want to rewrite the content largely (however that will need way more and precise documentation)
https://gitlab.com/kempelen/antix-faq-po-maintainer/-/blob/master/transliteration_insteead_of_translation.md?ref_type=heads (AI written, as a summary of a very long discussion)
Since the Translator Comments cannot be in the HTML File – like they are in the C++ files normally with GNU gettext -, I needed to create a separated JSON config file that gets merged into the POT template and then into the language files.
My “pot” file (some days old) is uploaded to the TESTING area on Transifex. It’s one file instead of 33, it gets split to HTMLs by the scripts. I think text extraction is clearly cleaner than original Transifex HTML extraction. I’ll re-upload that test file after adding translator comments.
– It would be wonderful to have some info on the original source files
— Git commit of the newest update, be able to fork for testing or updating
— Ability to submit English updates (patches, PRs) (based on this forum and en_GB package)
— Ability to include translations in the package and the webpage- including a “language selector” in the main template
— But I don’t dare to @ mention the one man who could answer these.. 🙂 (Did it above and got no reply, so then repeating would be arrogant.)
Thanks for all feedback and help on this guys!
Once again, the summary of my tool set, with all current features:
It uses a system dedicated to converting HTML to PO files and then back (po4a). This results in cleaner strings, better extraction of translation units. It uses one big file (2800 strings currently), so that repeated items are merged by “GNU gettext” automatically. The workflow is compatible with Robin’s uploaded files, assuming one team does not translate on both solutions at same time. Translation work can be migrated to my program or migrated back. Or if we upload this POT file to the Transifex, then translation teams could choose to translate either this 1 file or the 33 files (but fully avoid working on both). My solution does not allow HTML extensions of the web page within a string! That breaks po4a’s ability to merge translations back and forth baaed on HTML structure. And finally, it is fully capable of receiving updates to English text, while keeping translations and translator comments. Translations can be edited in Lokalize or Transifex, as a team wishes. (And result can be copied between them, overwriting the other end.) Additionally there is an alternative approach designed and partially documented, which is based on adoc TXT files, for teams who want to largely rewrite the text while translating.
Now, I really go and just translate. 🙂
- This reply was modified 4 months ago by kempelen.