Robert Kaye <[EMAIL PROTECTED]> writes: > On Fri, 2002-01-25 at 16:18, michael wrote: > > 1. AddFile makes some assumptions that won't work well for my mp3 > > setup and I'll likely have to roll my own script to populate the db. > > What assumptions is it making, and why won't it work for you?
ok, so here's my weird setup.
I have lots and lots of cd's and not a whole lot of disk space. and I
rip them all with high-quality settings.
so I have a song "foo". it appears on the album "bar". it's also on
the album "the best of bar" and on the compilation "punk bands perform
your favourite polkas". rather than have three identical copies of
the same song on my hard drives, I just have one. those who insist on
havering the album name in the directory path, can accomplish this
with symlinks, but in my case I realized that an album is just a fancy
form of playlist, and store my files under artist-name/track-name.mp3 and
have albums represented as album.m3u.
(needless to say re-mixes and live version have to be carefully marked
as such.) as a result, the id3 tag is good for artist name and track
name, but album name and track number are only representative of
whichever of the above albums I ripped most recently. for that I'd
have to parse the .m3u file.
granted that there are a large number of tracks for which the above is
no problem, but there are quite a few tracks for which it is.
(also I don't have id3 V2 tags, just V1. so a lot of the titles are
chopped badly if you go by that over what's in the filename or m3u
file.)
> > 2. Compilation discs were not well supported.
> > I read back in the email archives and found the patch for compilation
> > discs from a couple months back and (after realizing that the
> > ui_branch was not in working shape) I extracted it from the mail
> > message and applied it to the 0.5.0 code. apart from a few minor
> > aesthetic grips (such as not being able to change the name
> > "Compilation"), it seems to be working well, and I hope that this
> > patch makes it into the main code branch soon.
>
> Unfortunately I haven't had much time to devote to Obs lately. But,
> there are a number of things stacking up that I need to address real
> soon. Once I dedicate some hours to obs, I'll look into this matter.
cool. glad to hear it. :)
> > overall I think this is a fabulous product, even at this stage of
> > development, and I'm interested to see where it goes from here.
>
> The best way to make the project move forward is to jump in an help out
> a bit. :-)
yeah, ok then.
so based on my rambling above, I discovered that if I have two entries
in "Track" with the same Url and different Album and TrackNum, obs
will deal with it just fine. but I also realized that this goes
against how db's are supposed to work, and so I tried to do things a
bit more "properly" and created a table called "AlbumTracks" which
contains three columns. "Album" is and Id from the Album table,
"Track" is an Id from the Track table, and "TrackNum".
(I was able to populate this with a "select into" from the existing
info in the db.)
I modified the queries I found in most of the cgi-bin files that
weren't under admin/ (including the ones related to the compilation
changes from the above mentioned patch) and it seems to work ok.
(this also means that things like "TimesPlayed" will follow the song,
regardless of which album you played it off of. I think that makes
more sense than having separate entries for the same song, but others
may disagree.)
if no one here tells me that I'm completely off my rocker and going
about this totally backwards, I'll finish checking the sql in the
admin/ dir and do an alter table to drop those two fields from the
Track table. if that all works well, I'll post the patch here.
assuming of course that this is something that anyone besides me is
interested in.
after that's all done, maybe I'll go back and tweak AddFile to parse
m3u files, but one thing at a time.
since we're on the subject of duplicate files, and obs seems closely
in touch with freeamp, are there any plans to stick a (relatable?)
fingerprint into the Track table somewhere so that something like
AddFile can make more intelligent decisions about what constitutes a
duplicate entry?
what do you all think?
-michael
--
"I do not go out of my way to boldly split infinitives,
nor do I actively seek prepositions to end sentences with."
-James W. Zubrick
msg00527/pgp00000.pgp
Description: PGP signature
