On Sun, Aug 14, 2011 at 4:44 PM, Stef Bon <[email protected]> wrote:

> 2011/8/14 Rocky Bernstein <[email protected]>:
> > On Fri, Aug 12, 2011 at 9:45 AM, Stef Bon <[email protected]> wrote:
> >>
>
> >> Maybe you should try it.
> >>
> >> (as you can read I'm not ready with it, but it works )
> >
> >
> > I just tried compiling from git sources.  I get
> >     make: *** No rule to make target `cdfs-io-thread.h', needed by
> `cdfs.o'.
> >  Stop.
> > And there is no C header file cdfs-io-thread.h. Perhaps you meant to add
> > this to the git repository? (As I write this there is only one initial
> > commit to the git repository.)
> > As far a the C coding style and organization, the code looks good!
>
> Thanks!
> I removed a special io thread from the fs, which I use in almost every
> fuse fs. I've removed it recently from fuse-cdfs, but I forgot this
> one. I already fixed this. The new version in gitorious is corrected.
>
> Please let me know how it works!
>

Works fine for me for the iittle I have tried.

Two small things regarding the example invocation
http://linux.bononline.nl/wiki/index.php/Fuse-cdfs

In

--cache- directory=/home/sbon/.cache/fuse-cdfs/

 There seems to be a stray space after --cache-. Also, one can use libcdio
to find the a device that has a CDDA loaded in it via
cdio_get_devices_with_cap. In many situations there will only be one such
device with a CDDA loaded, but of course it is good to provide an
command-line switch to allow one to be specific.


>>
> >> Now I have some questions:
> >>
> >> a. is there a way to identify the cd uniquely.
> >>
> >> Right now it stores
> >> cached files in
> >>
> >> %cache_directory%/tracknr-xx.wav
> >>
> >> Now when I unmount, I ant to be able to keep the cached file. I would
> >> like to have an unique id per readable cdrom, like an uuid per
> >> partition on a harddisk. Is there one?
> >
> > A CD can have a ISRC also known as a MCN  (Media Catalog Number). In
> > practice many CDs do not have this. If I wanted to ensure uniqueness, I'd
> > probably create a SHA1 hash for each track and do something with
> combining
> > that for each of the tracks to form a disk hash.
> >
>
> Yes that's possible, but to get the SHA1 code per track you need to
> have to read the whole file first, and that's what I want to avoid,
> you understand? I would like to know how the fuse-fs can take the file
> from the cache instead of reading the cdrom.


> So yes the ISRC code you mention can be usefull here.
>

Also, you might be able to use or adjust the acoustic fingerprinting
routines from libofa <http://packages.debian.org/lenny/libofa0> for a track
fingerprint.


>
> >
> > The libcdio call read_audio_sectors allows you to read a large number of
> > sectors in one call. Depending on the CD-ROM driver this may wind up into
> > several smaller calls. If the CD-ROM support MMC, then read_audio_sectors
> > turns into a MMC READ_CD call. I don't know that one can be more direct
> than
> > this.
>
> OK, this stays in the fridge for now. This requires some digging.
>
> Stef Bon
> the Netherlands
>
_______________________________________________
Libcdio-help mailing list
[email protected]
https://lists.gnu.org/mailman/listinfo/libcdio-help

Reply via email to