[Paranoia-dev] libcdio paranoia status

[R. Bernstein]

Folks:

I just wanted to let folks know where things stand with cdparanoia which
I think influences when the next libcdio release comes. In short, I'm
hoping a release sooner than later.

As I've been adding more paranoia regression tests I'm more confident
that it works. And I've also been finding problems - I'm not sure so
much with the paranoia code as perhaps my testing framework. 

I've been able to run this on computers where I have access a CD-ROM:
GNU/Linux, Solaris, and cygwin. It works best on GNU/Linux (same as
cdparanoia).

It works on Solaris too, but with the simulated jittering, I get a
weird Segmentation Fault on a calloc of 28 bytes!  Probably I've
corrupted memory somewhere earlier; I'll run valgrind (on GNU/Linux)
to reduce the wild stores that may be there.

Steve Schultz reports that something was playable on his great Apple
G5. However Derk-Jan Hartman reports on the vlc-devel list problems in
reading audio after replacing a CD drive with a DVD drive. And there's
evidence from Scott Wood that with the Darwin complexity and device
dictionary scanning I've messed up somewhere. I'll post a separate
note to libcdio-devel about the Darwin situation.

In version 0.71, audio reading on FreeBSD CAM access was not working
where the CDIOCREADAUDIO ioctl is not available. A pity this wasn't
tested before the libcdio 0.71 release as it was easy to fix. After
fixing by simply using the common SCSI-MMC read routine, Heiner
reports having something working on FreeBSD 4.10. Library hell
prevented him from compiling on FreeBSD 5.3. But I don't have any
reason to believe that there will be big problems there.

cygwin works -- painfully slowly and with lots of errors. I think BSDI
works. (Right Steve? And Steve is not interested I have a hard time
believe anyone else on the planet is.)

The goal of the next release is not necessarily a flawless cdparanoia
on all platforms. That will come in time, depending on interest. 

I also don't expect people to drop whatever cdparanoia they are using
and use this one (except in cases where they don't have a cdparanoia
or where folks are libcdio friendly -- fat chance). That's one reason
I have provision for both to coexist and allow the name of the
cd-paranoia program to be specified on configuration.

In its initial libcdio cd-paranoia release, I am reminded of Gerd
Knorr's cdinfo program. When I started porting this to libcdio, cdinfo
was much richer than libcdio's cd-info. In fact there are still some
things it reports that cd-info doesn't. However at this point, I'd
venture to say that overall cd-info does more and is in fact more
widely used. And similar to cdparanoia vs. cd-paranoia the I suspect
there may be more frequent updates to the program ;-)