Arabic code `arb` is being marked as "ltr"

Question

Arabic code `arb` is being marked as "ltr"

Closed this issue 2 years ago · 4 comments

In building the langnames.json, the direction for arb is listed as left-to-right. The other language code for Arabic, ar, is correctly listed as right-to-left.

This seems to be coming from the pipeline somewhere.

Answer 1 · 2023-06-22T15:30:27.000Z

langnames.json isn't generated from this pipeline. Are you referrring to https://td.unfoldingword.org/exports/langnames.json? That comes out of translationDatabase or is this another langnames.json?

Answer 2 · 2023-06-22T15:34:11.000Z

I think it's somewhere in the catalog pipeline. translationDatabase marks arb as rtl, and that's what it is in the target language database, but in the sources, it comes through as ltr. This issue is to try to find out how that information is getting into the pipeline so that we can update the source.

Answer 3 · 2023-06-22T15:39:13.000Z

Got more details from John Wood about this indeed it is a bug thank you.

Answer 4 · 2023-06-22T19:09:12.000Z

@mondele This is now fixed and pushed into production and all CDN content has been purged, you should be able to refresh the languages in writer and see the changes.