#7589 closed bug (fixed)
Two chinese languages
Reported by: | axeld | Owned by: | pulkomandy |
---|---|---|---|
Priority: | normal | Milestone: | R1/alpha3 |
Component: | Kits/Locale Kit | Version: | R1/Development |
Keywords: | Cc: | ||
Blocked By: | Blocking: | ||
Platform: | All |
Description
When launching ReadOnlyBootPrompt, I get two different chinese languages IIRC zn-hans and zn_hans. I would assume one of them is invalid.
Also, are there really only two translations of the welcome text (Russian, and German)?
Attachments (1)
Change History (10)
comment:1 by , 13 years ago
comment:2 by , 13 years ago
Replying to axeld:
Also, are there really only two translations of the welcome text (Russian, and German)?
Axel, see the first paragraph of http://www.freelists.org/post/haiku-i18n/Update-ReadOnlyBootPrompt-and-Welcome-documentation
comment:3 by , 13 years ago
These Chinese catalogs are a PITA. It's insanity!
The zh-hans pack is the correct one, but contains zh_hans.catalog files. On the next export, the script detects the zh_hans.catalog files and continues to download the zh_hans pack, which is the wrong one! At least I think that's what's going on. I'll delete all zh-hans.catalog files and hope they don't crop up again...
I'm doing an export of HTA right now, but that takes verrry long. We'll see how many languages are then supported by ReadOnlyBootPrompt then. Theoretically all with completed and validated catalogs...
comment:4 by , 13 years ago
Resolution: | → fixed |
---|---|
Status: | new → closed |
Fixed with hrev41779. WRT to Korean, that should be a font problem. Do a " installoptionalpackage wqy-microhei" and set the font as system font and the squares are gone.
comment:5 by , 13 years ago
Thanks, but for the solution for the Korean squares is not acceptable for ReadOnlyBootPrompt at all. I'll open a new ticket for this issue.
comment:6 by , 13 years ago
Hm, you changed them all to zn_hans but mentioned that zn-hans would be correct. Looking at the ISO 639-1 language codes, it's "zh" though.
comment:7 by , 13 years ago
Hey, I understand it either... :)
It looks like you can create any kind of catalog at HTA by changing the URL to e.g. http://hta.polytect.org/catalogs/view/1200/yy which would result in a yy.catkeys of that catalog. Apparently there are some created under "zh-hans" and others under "zh_hans". The "zh-hans" pages are the correct/more-complete ones. When downloading all the catalogs, the ZIP contains "zh_hans.catalog" files however. I didn't bother renaming them all, as everyone will have to do that time and time again each export. "zh_hans" seems to be a special "simplified Chinese"...
I do hope we can switch to pootle soon and fix this issue on that occasion.
comment:8 by , 13 years ago
Sorry about the -/_ mess. It's a bit complicated...
- The ICU locale name should be zh-hans, as "hans" is not a country code (like in en_US) but a language variant. This seems to be the standard way to designat chinese, to avoid linking it to a country - for whatever reason it is the way ICU has chosen.
- However, the locale system expects to crawl up the catalog chain by splitting the string on underscores. So the files are named zh_hans, which allows falling back to zh (which currently doesn't map to anything, but that's another problem).
- It ùeans if we want to have country variants of zh_hans, they will have to be specified as zh_hans_contry, instead of the usual zh_country-variant, to allow for a more logical fallback.
Not everything maps perfectly...
comment:9 by , 13 years ago
Thanks for shedding some light on this, pulkomandy. Would the attached diff to hta_committer.sh solve the issue when automatically updating all catalogs? I think it should...
by , 13 years ago
Attachment: | hta_committer.diff added |
---|
diff to always get "zh-hans" catalog-package
Also, the Korean translation only shows squares here - this should really be resolved before the release.