What Ain has said neatly fixes everything, but a problem rises for GNU gettext users: The resulting po files are buggy in their headers. The headers are marked fuzzy and the header fields are set to default values. This is OK as far as translate-toolkit and sdf files are concerned, but it causes GNU gettext tools such as msgfmt report fatal errors. Our translators use GNU gettext to test their work since pofilter is extremely incompetent in checking the correctness of the translations. I would love to hear your views and ideas on this. fa
On پنجشنبه, 2006-01-05 at 19:17 +0200, Ain Vagula wrote: > Following steps are needed when migrating to UI strings with no duplicates: > - install updated toolkit > - generate sdf file from your existing translated or half-translated > po-s with new toolkit (to avoid english strings in untranslated > entries) > - backup old po-s somewhere > - generate new set of po-files from just generated sdf-file for your language > - update generated po-files with latest pot-files. > Now every duplicated string in one file has unique identifier, but > string itself is translated as it was in your old po-file. > Note that: > - When you do not regenerate po files, all duplicate entries will be > untranslated. > - When you use sdf file, generated with old toolkit, for regenerating, > untranslated entries will be filled with english strings > > ain > > --------------------------------------------------------------------- > To unsubscribe, e-mail: [EMAIL PROTECTED] > For additional commands, e-mail: [EMAIL PROTECTED] --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
