[etoys-dev] Etoys: GetText-Richo.9.mcz

[Bert Freudenberg]

On 28.05.2010, at 06:29, Korakurider wrote:
> 
> On Fri, May 28, 2010 at 1:01 PM, Ricardo Moran <richi.moran at gmail.com> wrote:
>> Yes, I found a couple of strings in english po that have different
>> translations. I guess I should find a different way of telling if a
>> translation wasn't found. Maybe if we make NaturalLanguageTranslator return
>> nil when a translation is not found? I just don't want to break any existing
>> assumptions.
>    Translator might return three kind of result:
>           a. The translator doesn't know such original string (e.g. msgid)
>           b. The translator doesn't know translation for the string,
> so return original string as-is
>           c. The translator return translation for the string
> 
>    We want only case #a  fallback to #translateInAllDomains.
>    In current implementation also case #b fallback, that isn't good.

I thought the mo files did not contain untranslated msgid? In that case, the translator could not tell the difference between a and b.

But why would we need to distinguish those cases anyway?

Another issue that plays into this ...
http://tracker.squeakland.org/browse/SQ-91

- Bert -