Chango 1.0
jsl@absynth.com
2000 July 03


Chango is a system for cacheing MP3 files from multidisc cdrom drives.
Since you can't actually mount different discs in the same physical
device simultaneously, this software will cache the files you need
off to a predetermined cache, so that all of the needed files will be
available simultaneously.

Prerequisite software:
	properly configured cdrom drives.
	'eject'
	'id3ed', by Matt Mueller  freshmeat index 922229360



    This set of scripts is free.  If you use it, try to give me some
    sort of credit.  If you find it useful, drop me some e-mail!

    No warranty is either expressed or implied.  If this somehow kills
    your grandmother, it's not my fault.  Don't do dumb things with this.


    btw.  it's pronounced "change-oh".


Here's how to configure it for your system;

- Edit the 'drives.conf' file in accordance to how your machine is set up.
  The included .conf file shows two drives, each holding a max of 4 discs.

	/dev/cdrom0:/mnt/cdrom0:4
	/dev/cdrom1:/mnt/cdrom1:4 

  I setup symlinks in /dev  from /dev/cdrom0 to /dev/hdc  and a second one
  from /dev/cdrom1 to /dev/hdd.  Make sure these are read/execute by the 
  user who will be running the software.   I am using a RedHat distro, so
  all mounted devices will end up in /mnt.  so, my /etc/fstab shows the 
  following:

	/dev/cdrom0    /mnt/cdrom0     iso9660 noauto,owner,ro 0 0
	/dev/cdrom1    /mnt/cdrom1     iso9660 noauto,owner,ro 0 0

  So, by looking at both files, you should see that there is a corrolation
  between the two.

- At this point, you should be able to 'mount /mnt/cdrom0' as whatever 
  user will be running the script.  As well, you should be able to change
  disks using 'eject -c 2 /dev/cdrom0' to swap disks. 

- Run the 'scandrives.pl' script.    'perl scandrives.pl'
  This will scan the CD's in the drives, and create the database files 
  based on the mp3s.  It should show you the directories it is working 
  through as its working.   If it notices that the cd is the same as the
  one it had already scanned, it will skip it.  (This is so that if you
  change just one disc, you can re-run the app, and it will only do 
  the full work on the new or changed discs.)

  The files it creates (in "db/") are of the following form:

  The filenames themselves are the mountpoint, then a period, then the 
  disc number.  The path for the mountpoint has the slashes changed into
  colons.  So, if you wanted to look at the file for /mnt/cdrom1, disk 3,
  you would look in "db/:mnt:cdrom1.3"

  The files themselves are of the following form:
      Header information
      a blank line
      database entries

  The header information currently is just a single line, which is composed
  by the amount and content of the disc.  This is how I identify one disc
  from another.  It's simple, and is pretty effective.

  The blank line is there so that in the future, if more header information 
  is added, it will be easy to seperate it out from the rest of the file.

  Here's an example:

/dev/cdrom1##::##/mnt/cdrom1##::##2##::##/mnt/cdrom1/Hack/08 Hard Currency.mp3##::##Hard Currency##::##Information Society##::##Hack##::##1999##::##, AG# E6A59D0A##::##Pop(13)

  The database entries are fields seperated by the string "##::##".  That is,
  pound, pound, colon, colon, pound, pound.  A string that probably won't
  appear anywhere in the filenames.

  The fields, in order are:  (with examples)
	Device path		/dev/cdrom1
	Mount point		/mnt/cdrom1
	Disk Number		2
	path to file		/mnt/cdrom1/Hack/08 Hard Currency.mp3
	ID3 Title		Hard Currency
	ID3 Band		Information Society
	ID3 Album		Hack
	ID3 Year		1999
	ID3 Comment		, AG# E6A59D0A
	ID3 Genre		Pop(13)

  Writing a playlist editor using these files should be trivial.

- In any event, by this point, you should have valid database files, all 
  filled out with all of the above info (assuming the MP3's have correct
  ID3 tags.)

- To cache in a file, just send cache.pl the database entry (like the above
  line).  ie:

      perl cache.pl /dev/cdrom1##::##/mnt/cdrom1##::##.....etc....

  Or, say you want to cache all files with the word "Talking" somewhere in
  their entry, you could do this:

      grep --no-filename -i Talking db/* | perl cache.pl

  If the file that it is about to copy is already in the cache, it will just
  touch the file instead of re-copying it.

- It is assumed that some external timeout mechanism or expiration mechanism
  will be used to remove the files from the cache.