[Rhythmbox-devel] Followup on an odd bug with oggs.

RB Devs,

I was experiencing crashes with rb with .ogg encoded files, so I decided
to join #rhythmbox and see if I could get some help troubleshooting the

Colin was around and was very helpful in helping me troubleshooting the
problem. I figured I would post a followup here and see what (if any)
kind of bug should be filed.

Initially, my problem was that oggs totally crashed rhythmbox. At first,
I thought the problem was my gstreamer configuration, until I realized
that gst-player played the oggs perfectly. So did every other player on
my debian sid system.

After doing all sorts of hairbrained combinations to find out what was
wrong, I figured the only thing that could possible left was that I was
attempting to play the oggs off an smb share. Turns out, I was dead

After scouring the vorbis lists, I came across a post that mentioned
that MP3 ID3 tags, added to oggs would corrupt them. I ran the command
line tools, and lo and behold, all my oggs were corrupted. How? Grip has
an option to add ID3 tags, even to oggs. Every one of my 400+ oggs I had
encoded where hosed. 

Now, ogg123, xmms, and my windows players could play my oggs. So I
reencoded one of them without the ID3's and used the Ogg descriptions,
and it worked in rhythmbox flawlessly. 

I have since reencoded my .oggs the "right" way and have had no problems
with it.

I would think from all of this that RB itself is rather sensitive to
playing these corrupted oggs, since they seem to play fine in most
players. Either way, since Grip doesn't seem to give any warning that
adding MP3 ID3's to oggs can be bad (and I have confirmed this with some
windows rippers/encoders like EAC), that this might lead to some
problems in the future.

Just wanted to make people aware of this issue. I've got all sorts of
junk about it written down if someone feels that this is a bug. ( I
mean, is it really? Other players are more forgiving, does that make rb
wrong?) Maybe a "You idiot, you made your ogg wrong" dialog is better
than a total crash.

Extra thanks to Colin Walters for helping me get my brain moving on
this, his Debian packaging work has been excellent. :)

jorge o. castro <jorge@whiprush.org>

This is a digitally signed message part

[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]