Hi,
using tovid in a European locale, I've been having problems using accented
characters. Accented characters are basically those outside the 7-bit ASCII
specification. In Germany, umlauts are quite commonly used: äÄöÖüÜß
While this works from the command line:
> tovid todisc -files input.mpg -titles "This has Umläüts" -out DVD -dvd -pal
-ffmpeg
tovid titlesets will also accept them in its input fields, but their processing
will fail along the way. On IRC, I had already proposed one or two "unicode()"
workarounds which have found their way into SVN, but they don't cover
everything.
Solution:
- wrap relevant code sections with some kind of to_unicode()/to_utf8(), as
outlined here: http://farmdev.com/talks/unicode/
- take similar code from yum
Problems:
- You can't just "switch the app" to behave properly (except if you use
Python3, according to the docs, it should "just work").
- You have to find affected code sections.
I have muddled my way through "tovid titlesets", using accented characters in
titleset and thumbnail menues. I inserted appropriate conversions at every
failing point. With the resulting attached patch, I got all the way through the
wizard without failure and with the correct result on DVD.
I was using a en_US.UTF-8 locale. I don't actually know whether Python or X11
are interested, and what affect it has on the input fields in X (or tkinter). I
didn't try non-UTF-8 locales (such as latin1 of Japanese), though my guess is
that they should transparently work.
Here are the caveats of this patch I can think of:
- Other locales: See above.
- Test coverage: I have only gone through this scenario with two titlesets of
two videos each.
- File names: I did not try file names with accented characters yet. I have no
idea whether they are transparent to Python (they better be, otherwise it's
asking for trouble!) or need special handling.
- There are probably function calls in other places which I haven't hit, which
require the same wrapping. This needs code review!
- setup_locale() may actually be redundant or a no-op; omission of it is yet to
be tested.
Please do give the attached patch (against tovid SVN r3301) a go and at least
check whether it breaks anything for you. If not, it might be a candidate for
inclusion despite the caveats. :-)
Apart for that, this should be inspiration for proper implementation. I have
tried my best to go along the docs.
Thanks.
P.S.: I'm gibson/gibson_ on IRC's #tovid.
Original issue reported on code.google.com by mbst...@googlemail.com on 29 Sep 2010 at 3:13
Original issue reported on code.google.com by
mbst...@googlemail.comon 29 Sep 2010 at 3:13Attachments: