I'm using de_DE.UTF-8 and MPlayer doesn't display german umlauts correctly. With de_DE@euro there's no problem. I introduced in other ebuild the USE-flag "unicode" and included this code: # if specified, convert all relevant files from latin1 to UTF-8 if useq unicode; then for i in file1 file2 file3 etc; do einfo "Convert ${i} to UTF-8" iconv -f latin1 -t utf8 -o "${i}~" "${i}" && mv -f "${i}~" "${i}" || rm -f "${i}~" done fi perhaps we can do this also for MPlayer? Would be great! Just run 'file' over all relevant files and all with "iso-8869-1" or "latin1" should be included in the for-list, if they have are used for output or are documents/help-files. Reproducible: Always Steps to Reproduce:
btw: the iconv loop would be a good candidate for the eutils class! econv -f latin1 -t utf8 <files...> ;-)
This is something I'd rather have upstream take care of, as if I do German, I'd have to do the other languages too. However, I don't know the codepage used for each and every file, so it would be better to ask them to do it, knowing what locales they have.
Created attachment 51270 [details, diff] Adds help text conversion to utf-8 for some languages (applies to mplayer-1.0_pre5-r5.ebuild) When the use flag UNICODE is set, this patch converts the help texts for some languages from their original encodings to utf-8. This way a unicode console/xterm can display the texts correctly. Several languages are missing in this patch, but may be added by native speakers. This patch should be only provisional until the original MPlayer sources are converted to Unicode/UTF-8.
yo, it looks ok to me. Chris? please! :-D
I hope to include a patch in MPlayer soon so that you only need to specify --charset=utf8 to the configure script to achieve this... For details have a look here: http://mplayerhq.hu/pipermail/mplayer-dev-eng/2005-February/033167.html
great! But why not having all stuff in UTF-8 already and convert it at runtime into LOCALE? Unicode is *the* charset "Esperanto" and UTF-8 is the best encoding for it. One should stop using other charsets (at least internally).
I agree that utf8 is best, but I prefer to let the translators choose their prefered charset. Also converting to LOCALE is a bit too much "auto-detection" for my taste, I somehow doubt it will work reliably - so anything beyond what this patch does I leave to someone else. Anyway, that patch is now applied but it works only for languages where the right charset is known (de, en, pl and ru intially, but I expect the rest to follow in a few days).