Rsync and Jemplode

Posted by: TedP

Rsync and Jemplode - 03/01/2003 13:16

Hello,

I've been reading the RSYNC thread, and it's something I've been wanting to do for quite a while. Also, I am loving the "programmable soups" of Jemplode. My question is.. can I do both? I would run RSYNC to sync the MP3 files, and then run JEMPLODE to generate the soups. Would this work? and if so, what would happen on subsequent RSYNC runs?

Oh yes.. I will be running on WIN2K. Which packages of cygwin will I need to install? I assume I will need GCC to compile the program.

Thanks
-Ted
Posted by: image

Re: Rsync and Jemplode - 03/01/2003 13:42

rsyncing w/ the mp3tofid tree will delete your playlist file, so all jemplode soups will be gone. i personally do a two step process, where i rsync then recreate the soups.
Posted by: MP3944

Re: Rsync and Jemplode - 06/01/2003 09:01

I don't know if this has been mentioned before, but, is there any way rsync functionality can be added to jEmplode?

Sorry if this is a repeat question.
Justin
Posted by: mschrag

Re: Rsync and Jemplode - 06/01/2003 10:36

Cloning and syncing are pretty popular requests, and I've considered the rsync route ... I haven't really decided how I will end up solving that problem yet, though.
Posted by: TedP

Re: Rsync and Jemplode - 06/01/2003 11:35

iMaGe: It looks like you're doing what I want to do, but I'm not sure I quite understood the process.

#1: you rsync the database (step 1). the first time you do this, all soups and playlists are on the RIO right?

#2: you run jemplode and recreate the soups (step 2).

now, what happens on subsequent rsync runs? are the jemplode soups preserved?

thanks
-ted
Posted by: image

Re: Rsync and Jemplode - 06/01/2003 12:55

no they arent. jemplode soups are just playlists that are added by jemplode to match a criteria. thats all.
Posted by: pim

Re: Rsync and Jemplode - 06/01/2003 19:17

I will be running on WIN2K. Which packages of cygwin will I need to install?

Get at least the base package, cygrunsrv and rsync (of course).
Also get a decent editor, one that knows about unix/windows/mac
text files. It does not have to be a cygwin-based editor (like vim).

I assume I will need GCC to compile the program.

Not unless you need to change anything. A precompiled binary for
the cygwin environment is included in the package.

Pim
Posted by: pim

Re: Rsync and Jemplode - 06/01/2003 19:23

Hmm. Given that Jemplode soups are just playlists, it would be very possible to implement soup mode for mp3tofid by just creating a symlink tree within the mp3 directory tree. mp3tofid will recognize the symlinks and build fid playlists from them.

The symlink tree could be created by mp3tofid or by another binary or script.

I'll take a look at jemplode to see what it does ...

Pim
Posted by: TedP

Re: Rsync and Jemplode - 06/01/2003 23:41

hey, thanks for the replys PIM! you all are talking a little above my head, but if i understand how JEMPLODE works, it dynamically regenerates each soup after a resync by storing info about the soup filters in the tags. if this is the case, then your suggestion of linking into the JEMPLODE playlist might not have the desired effect unless you copy the tag info along with them.

im looking forward to getting this all installed.

thanks
-ted
Posted by: image

Re: Rsync and Jemplode - 07/01/2003 08:11

if you can put REG EXP functionality into mp3tofid, it would be a great help. symlinking all the files that match a criteria manually would be too tedious. i was thinking that if mp3tofid comes across a directory with just a specially named text file containing a regexpression, then it would add everything that matches to a playlist which is named the same thing as the dir. (look at emptool source for regexp code, i would imagine).

btw, have you been able to implement the crc check that would be compatible w/ jemplode?
Posted by: mschrag

Re: Rsync and Jemplode - 07/01/2003 08:24

i didn't thoroughly read what he was trying to do, but for reference, jEmplode stores a tag attribute at the top level only that tells the system that the playlist is a soup playlist (soup="search:refs=0", for instance). When you sync with jEmplode, it looks at the soup definition and looks for anything that is 1) missing from the def and adds it and 2) shouldn't be there (doesn't match anymore) and removes it. So if you put other things into the soups that weren't there before, they should be automatically removed by jEmplode. However, you sort of can copy subsections of the soups to other places (with the caveat that when that section becomes completely empty <i.e. you remove all the artists with a given name>, it will be discarded by the soup and not kept up to date -- so if you added a new artist with that name after clearing all the others out, it would make a new playlist for the artist).

ms