Note: This is a public test instance of Red Hat Bugzilla. The data contained within is a snapshot of the live data so any changes you make will not be reflected in the production Bugzilla. Email is disabled so feel free to test any aspect of the site that you want. File any problems you find or give feedback at bugzilla.redhat.com.
Bug 132888 - sound-juicer totally broken
Summary: sound-juicer totally broken
Keywords:
Status: CLOSED RAWHIDE
Alias: None
Product: Fedora
Classification: Fedora
Component: sound-juicer
Version: 3
Hardware: All
OS: Linux
medium
medium
Target Milestone: ---
Assignee: Colin Walters
QA Contact:
URL:
Whiteboard:
Depends On: 133312
Blocks: 131589
TreeView+ depends on / blocked
 
Reported: 2004-09-18 19:36 UTC by Havoc Pennington
Modified: 2007-11-30 22:10 UTC (History)
4 users (show)

Fixed In Version: 030-27
Doc Type: Bug Fix
Doc Text:
Clone Of:
Environment:
Last Closed: 2004-10-14 20:07:54 UTC
Type: ---
Embargoed:

Attachments (Terms of Use)

Description Havoc Pennington 2004-09-18 19:36:50 UTC 
Several problems here really:

1. if the device in the gconf key is wrong or unset, sound-juicer 
exits at startup; but then you can't get to prefs to set up the 
device.

2. by default we have the device wrong, so it looks like #1 always happens

3. we don't seem to have permissions for users to access /dev/cdrom
by default; we should have that working somehow

Anyway, the basic thing I think we need to work for release is:
 - fresh default install
 - plug in USB CD-ROM drive
 - be able to open sound juicer and juice the songs
with no command line...

Not sure how much of it not working on my system is due to rolling
upgrade rather than fresh install.
Comment 1 Havoc Pennington 2004-09-18 19:41:30 UTC 
of course the same applies to gnome-cd, etc., not only sound-juicer
Comment 2 Colin Walters 2004-09-20 21:17:40 UTC 
1) Pretty sure this is fixed by 0.5.12, just uploaded to rawhide
2) I think this one might be fixed too
3) Mrmph.  udev.  I need to install a fresh rawhide system and
investigate that.
Comment 3 Colin Walters 2004-09-20 21:32:26 UTC 
johnp and I talked about this a bit but we weren't sure - what's the
story with permissions for dynamically plugged in CD devices? 
Currently pam_console chowns the device when you log in, but that
isn't enough for devices plugged in after login.

Nalin, do you know?
Comment 4 Colin Walters 2004-09-21 21:28:15 UTC 
David says the permissions fix was a udev bug, didn't make fc3t2, but
should work in the latest rawhide.
Comment 5 David Zeuthen 2004-09-21 22:18:52 UTC 
I've never used sound-juicer before on my system and 0.5.12 doesn't
work for me; I'm seeing issues 1. and 2. Specifically I'm getting this

[david@laptop ~]$ sound-juicer

** (sound-juicer:5029): WARNING **: Can't read /proc/scsi/sg/device_strs

** (sound-juicer:5029): WARNING **: Can't read /proc/scsi/sg/device_strs

However, udev-027 does fixes issue 3 and I can confirm udev-026
(shipped with test2) didn't.
Comment 6 David Zeuthen 2004-09-21 22:22:08 UTC 
After 'modprobe sg' as root, I get this

[david@laptop ~]$ sound-juicer
This CD could not be queried: Cannot open '/dev/hdd'

That's probably because my audio disc is in /dev/sdc0 which is my
external USB2.0 optical drive.
Comment 7 David Zeuthen 2004-09-21 22:32:22 UTC 
Hmm, when selecting the right drive in the preferences dialog of sj,
I'm getting this

[david@laptop ~]$ gconftool-2 --get /apps/sound-juicer/device
/dev/scd0
[david@laptop ~]$ sound-juicer
This CD could not be queried: Cannot open '/dev/hdd'

Inserting the drive into /dev/hdd and still having /dev/scd0 as the
preference makes ripping work.
Comment 8 Harald Hoyer 2004-09-22 08:24:02 UTC 
latest udev should fix the dynamic cdrom permission issue and also sr0
-> scd0 renaming
Comment 9 Colin Walters 2004-09-22 18:51:04 UTC 
I have:

[root@nexus walters]# rpm -q udev
udev-032-1

But my USB cdrom still shows up as /dev/sr0.

And I get a kernel oops unplugging it.  Whee, off to file that bug.
Comment 10 Colin Walters 2004-09-22 21:16:40 UTC 
As for 2), it looks like sound-juicer stole some code to detect CD
devices which supports either HAL or linux-specific bits, and it just
uses the linux bits.  I'm trying to enable HAL now to see if that helps.
Comment 11 Colin Walters 2004-09-22 21:33:28 UTC 
harald: nevermind, it is /dev/scd0, i was confused by dmesg saying it
was sr0.
Comment 12 Colin Walters 2004-09-23 02:35:17 UTC 
Ok, this is a deep, deep mess.  I'm working with the latest
sound-juicer CVS which fixes some bugs, and I've got it using HAL to
pick up cdroms instead of the linux-specific grubbing, but this is all
for naught because apparently the latest cdparanoia doesn't understand
/dev/scd0.  This is bug #133312.
Comment 13 Havoc Pennington 2004-09-25 03:12:37 UTC 
On my current system:
kernel-2.6.8-1.541
hal-0.2.98-4
udev-032-1
sound-juicer-0.5.12-1
cdparanoia-alpha9.8-21

I've just been rebooting a lot and to make things work I have to do
the following:
 - do not plug in the USB CD at boot time, or the kernel oopses;
   seems safe to plug it in after boot
 - MAKEDEV sg0
 - chmod hp.disk sg0

If I do those things then sound-juicer works fine. udev only creates
scd0 which as discussed above doesn't work.

This probably isn't helpful, but I thought I'd add the data point.
Comment 14 Colin Walters 2004-09-25 04:56:03 UTC 
With Peter's modified cdparanoia, and the sound-juicer from CVS,
things work for me reasonably well.  I'm hoping Ross will do a new s-j
release soon, if not I'll just do a cvs snapshot.

Havoc, does your kernel oops look at all like this?
https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=133249
Comment 15 Colin Walters 2004-09-27 17:25:33 UTC 
Ok, I just built a new sound-juicer snapshot into rawhide,
sound-juicer-0.5.12.cvs20040927-1.  Havoc, could you try installing
Peter's cdparanoia-alpha9.8-21 RPM from the URL below and see if
things work for you?

http://people.redhat.com/pjones/cdparanoia/
Comment 16 Colin Walters 2004-10-14 20:07:54 UTC 
This works for me, please reopen if you still have a problem.
Note You need to log in before you can comment on or make changes to this bug.