uniform comma below characters for romanian translations

There are four conversions

First:
from U+015E LATIN CAPITAL LETTER S WITH CEDILLA
to U+0218 LATIN CAPITAL LETTER S WITH COMMA BELOW

Second:
from U+015F LATIN SMALL LETTER S WITH CEDILLA
to U+0219 LATIN SMALL LETTER S WITH COMMA BELOW

Third:
from U+0162 LATIN CAPITAL LETTER T WITH CEDILLA
to U+021A LATIN CAPITAL LETTER T WITH COMMA BELOW

Fuorth:
from U+0163 LATIN SMALL LETTER T WITH CEDILLA
to U+021B LATIN SMALL LETTER T WITH COMMA BELOW

We have no mechanism to do this right now, and building one would take some care to avoid doing unwanted conversions.

You could work around this by exporting Romanian translations, running a search-and-replace, and re-uploading them if there have been changes.

Jeroen, what cases are there when Romanian uses U+015E and it should *not* be converted to U+0218, etc? If there aren't any, then unwanted conversions would not be an issue.

That we have no mechanism to do this seems like begging the question -- that's not a reason that the function shouldn't be implemented, it's why this bug was reported in the first place.

I've discussed this with Adi (he emailed me directly), and I promised we'd look into it.

My original thoughts have been to prepare a message-filter script which would allow filtering by language and project, and running appropriate filter on it (it could be as simple as running sed, eg. exactly like msgfilter does for PO files).

However, implementing per-language, per-project rules in web interface is something that's going to be much harder, and would require much further discussion and thought.

I had made this request after I found out that Fedora Romanian Transaltion team was using a webpage translation tool implementing this features.

The software is called Narro and is developed by the guy working at Romanian mozilla and open office translation http://code.google.com/p/narro/

I understand why the filtering would need to be per-language, but I don't see that it would need to be per-project. Presumably Windows XP encourages the same errors regardless of what you're translating.

This is a per-language filter as comma below characters are standard in Romanian language in the last 3 years.

A project may decide not to use standards characters but rather go to support legacy characters. But I think this is rare case and new translations will be created for the current standards.

It's simpler to fix this at user leve. This problem would be easily solved by introducing a javascript snippet, can that be done for the Romanian locale only ?