[mp3togo] Re: mp3togo 0.5.0

[sim]

Hi Mark,
Comments inline:

On 6/12/06, Mark J Hewitt <mjh at elsabio.demon.co.uk> wrote:
> Sim,
>
> Sorry for delay - just got back from the US.

Nice, I hope you had a good trip.

> I am trying out 0.5.3, and this is the log I see:
>
> [mjh at tardis workspace]$ mp3togo  -o /workspace/resampled -v 9 -t .2go -d
> 3 /music/collection/classical/*/*
> ogg123 binary found at /usr/bin/ogg123
> lame binary found at /usr/bin/lame
> faad binary found at /usr/bin/faad
> oggenc binary found at /usr/bin/oggenc
> normalize-audio binary not found
> mpg321 binary found at /usr/bin/mpg321
> dcop binary found at /usr/bin/dcop
> flac binary found at /usr/bin/flac
> metaflac binary found at /usr/bin/metaflac
> Third party module: xmms is not available. Doing without
> Third party module: ogg.vorbis found
> Third party module: ID3 found
> 'normalize-audio' binary not found. Normalization disabled.
> mp3togo 0.5.3
>
> <space> or 'p' to pause, <esc> or 'q' to quit
>
> (1/668)
> /music/collection/classical/3_Gymnopedies_&_Other_Piano_Works/01-Satie,Erik=Gymnopédie_No._1=Pascal_Rogé.ogg:
> Creating dirs: Done.
> Decoding ogg: Done. 1s elapsed.
> Encoding mp3:  94% ##################Exception
> exceptions.UnicodeEncodeError: <exceptions.UnicodeEncodeError instance
> at 0x2aaaaecb4b90> in <bound method ID3.__del__ of <ID3.ID3 instance at
> 0x2aaaaecb4a28>> ignored
> Encoding mp3: Done. 12s elapsed.
> Tagging: Done.
> Total time this track: 14s
>
---8<--- Rest of tracks snipped

Hmm, I can see that the exception no longer stops the program, I fixed
that at least. Does it produce output files? are they tagged?

This is a bit of a head scratcher, as you are decoding ogg files, the
ID3 object shouldn't even be instantiated until the tagging step. The
tags are read from ogg files using the ogg.vorbis module and stored in
an internal dictionary until it is time to write them out.

This failure seems to be occuring during the decode step.

Are you running an SMP or Dual core system?

> If you think it might help, I could send you one of these ogg files to
> see if you can reproduce the problem.

That would be a big help, I'd be in much better shape if I could
reproduce the problem. You are welcome to send one of the failing
files to sim at puddle.ca

> I wonder if it may be because this is a 64 bit platform?

I thought those elapsed times seemed pretty small!

I can't see that being a problem, but I could be wrong.

The program uses a thread to monitor the subprocess, so it might be a
concurrency issue, I've not written much threaded code, I wouldn't be
surprised if I've screwed it up somehow.

> Or perhaps because my locale is em_GB?

I doubt this would be a problem. The unicode error seems to be a
definite clue though. I had a quick look through the ID3 code when you
first reported this bug, but nothing jumped out at me.

> Mark J Hewitt.

There is certainly something strange going on here. I'll study the
code and see if I can figure out where this exceptional ID3 object is
coming from. Things almost seem to be happening out of order here.

Maybe you could try mp3togo with the new --no-tags option and see if
that also fails. It's no solution, but it might provide more clues.

I'd like to get this and the Debian Stable bug resolved before I
continue work on the PyGTK front end.

Oh, a few versions ago I switched the default --file-tag option to
'.2go' from '-PORTABLE-' as it seemed much prettier, so you needn't
use the -t option.

Thanks again for the bug report Mark.

> sim wrote:
> > I have just posted mp3togo 0.5.1
> >
> > This release adds a command line option '--no-tags' that disables file
> > tagging. The unicode exception thrown by ID3 is now caught, which will
> > prevent mp3togo from rolling back the track, but will not tag the
> > output file in case of error.
> >
> > This isn't exactly a bug fix, but it is more robust behaviour in the
> > face of failure. I'm not sure what I can do to properly fix this bug
> > as I'm not able to recreate it. With any luck, it will disappear in
> > the tagging subsystem overhaul.
> >
> > I hope this solves your problem Mark, or at least gets you listening
> > to your music. If not, please let me know.
> >
> > The new release is available from the website here:
> > http://puddle.ca/mp3togo/download.html
> >
> >
>


-- 
sim