Re: msgmerge feature request - 'compendium' flag wanted

[Karl Eichwalder]

Bruno Haible <address@hidden> writes:

> Karl Eichwalder wrote:
>> If you merge (update) PO files using compendia, it sometimes is useful
>> to mark translations originating from the compendia with a special flag
>> ('fuzzy' could be good enough, but a new flag like 'compendium' would be
>> better).  So the user could easily check, whether compendium provided
>> translations are appropriate.
>
> Can you explain a bit more about the use case? If the translator trusts
> the compendium, she uses "msgattrib --set-fuzzy"; otherwise she uses
> "msgattrib --clear-fuzzy".

Thanks for the hints.  I forgot about the possibility to re-use fuzzy
messages from a compendium file directly.  So this could be an option.

But this way we would mix two kinds of fuzzy messages:

1/ the normal fuzzy messages where msgid has changed a little bit and
thus the translator must carefully check and probably update the
translation

2/ messages deriving from the compendium, that are 100% correctly
translated, but should be checked nevertheless, because even a 100%
correctly translated message could be inappropriate in a different
context (I know, you now can use the new context keyword).

So if the translator would see, it is taken from the compendium and a
100% match, she could decide very fast how to deal with such a
translation (e.g., she must verify only short messages such as "label",
"print", or "update configuration"; she can enable long messages
literally without reading them).

I just noted the --set-fuzzy will also set the PO file header "fuzzy".
A option such as --keep-header would be useful.

> Also, these markers in the PO files are made to survive any number of
> msgmerge invocations. Is this important here?

Yes, as long as the msgid did not change the "#, compendium" marker
should stay.  If you re-use such a compendium derived translation for
something different, just turn "#, compendium" info "#, fuzzy".

> If not, wouldn't an "msgdiff" program be more appropriate for the
> task?

I am not sure.  But we all would wlecome such a dedicated diff tool!

-- 
Karl Eichwalder
R&D / Documentation                            SUSE Linux Products GmbH

GPG: 1024D/06EB882E  B2A3 AF2F CFC8 40B1 67EA  475A 5903 A21B 06EB 882E