[mythtv] mythmusic w/o id3 and setting to limit how often rereads dir

[Isaac Richards]

On Saturday 07 December 2002 02:06 am, Chris Pinkham wrote:
> Couple of ideas that people might like for mythmusic.  I've already
> implemented one and will do the 2nd real soon probably.
>
> I don't have ID3 tags on most of the 2000+ mp3 files that I've ripped from
> my wife and my CDs.  My car mp3 jukebox computer doesn't need ID3 tags so
> I never bothered to add them to the files I ripped.  I store the files in
> paths ending like GENRE/ARTIST/ALBUM/TITLE.mp3.  Since I don't have ID3
> tags these show up in mythmusic a little weird (ie, not by genre, artist,
> album, title order).  I added a little code to mythmusic (in maddecoder.cpp
> I think) to fake the ID3 info if it's not there.  It assumes the files
> are stored in the above directory structure and uses that info to populate
> the correct variables before populating the database.  I was wondering if
> others might be interested in something like this or if they have better
> ways to do it.  My code is enabled/disabled by an "ignoreID3" variable in
> settings.txt

Sounds good.  It'd probably be nicer if the directory/naming scheme were 
specified in the settings file, but I don't have a problem hardcoding it like 
that.

> My second idea concerns mythmusic scanning the directory every X times
> that it starts up.  I'd prefer that it never did the scan unless I wanted
> it to (ie, picking a menu item), but will probably take the easy road
> out right now and make the just make it a setting in the config file
> to scan every X startups.  I think someone said it does it every other
> time right now which can get to be really slow.  I have all my mp3 files
> stored on an nfs/samba server so mythmusic reading them can be slow.
>
> Comments/suggestions/flames?? :)

It _might_ be more worthwhile into seeing why it takes so long with mp3 files 
-- most of the time sink on initial database import is because it's 
calculating the length of each song, and that takes awhile to do properly.  
But, each scan after that should just be testing for the presence of the 
file, essentially.  For me startup time is pretty much instantaneous, though 
I only have .oggs and .flacs, and I'm using it locally, not over a network..  
Your call, though, as a setting would work just fine.  I'll commit any 
patches you send my way.

Isaac