Re: Making it easier on translators

[Karl Eichwalder <keichwa gmx net>]

Keld J=f8rn Simonsen <keld@dkuug.dk> writes:

> Messages in .po sometimes have a high granularity, such
> as eg the function descriptions in gnumeric.
> You may have to read thru 20 lines of detailed description
> just to find out that they put a # before an ENUM value,
> or removed it. A tool that could identify just the few words
> characters changed would help.

Yes, that's another problem I talked about to Fran=e7ois Pinard; he
considered to enlarge Emacs/po-mode with ediff (other tool may want to
rely on wdiff?).

The real problem is the .po file lacks the "old" msgid.  'update.pl'
creates LL.old; this file could be used to find the old msgid.  Maybe,
it would be better to save the old msgid inthe .po file to?

msgid_old ""
"%s not found"
msgid ""
"%s not found.\n"
msgstr ""
"%s nicht gefunden"

This way it wouldn't be that difficult to make use of ediff.  Another
possibility would be to invent (temporary) markers to chain the entries
together:

de.old:
-------
#@ i00001
msgid ""
"%s not found"
msgstr ""
"%s nicht gefunden"

de.po:
------
#@ i00001
msgid ""
"%s not found.\n"
msgstr ""
"%s nicht gefunden"

But as long as every msgid isn't longer than approx. 10 lines it isn't
that hard to compare the new msgid and the msgtr (the old translation)
sentence by sentence.

We've just to make sure hackers will split the strings in reasonable
chunks; happily that's already done for some --help tables :)

-- 
work : ke@suse.de                          |                   ,__o
     : http://www.suse.de/~ke/             |                 _-\_<,
home : keichwa@gmx.net                     |                (*)/'(*)