Ain Vagula wrote:
Tímár András wrote:
Ain Vagula írta:
Latest translate-tools can operate so, that all strings in one po file
are unique (using kde-style identifier in msgid).
This is really good news. A question regarding the efficiency of this
approach: How about the 100% matches and fuzzy matches when an update
comes? Can the tools handle smartly the following scenarios?
1. ID changes and source text remais the same (e.g. relocation of the
text between modules)
2. ID is the same and source text changes (e.g. typo fixes,
clarifications)
3. A new ID is introduced but the text is already translated with a
different ID.
There were only cosmetical changes after I started use it, so lets see
when we are moving back to development version after 2.0 release with
massive changes in strings.
I'm sure, it handles 2. correctly, and almost sure that 3. is not
handled. I think, David Fraser is more competent to answer. As author ;)
Yes, I'm not sure how you could even handle 3. You can use translation
memory tools for this, but the problem is that most PO tools do assume
that you only have one translation per original string text.
Anyway, I would think the whole point of being able to have different
translations for different IDs is that you shouldn't assume that the new
ID should have the same translation! So in this case you should at least
be manually checking that the new ID should have the same translation,
in which case its really a form of translation memory you need...
Cheers
David
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
