[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Gnump3d-users] Test the pre-release?

From: Steve Kemp
Subject: Re: [Gnump3d-users] Test the pre-release?
Date: Thu, 31 Aug 2006 17:55:14 +0100
User-agent: Mutt/1.5.9i

On Thu, Aug 31, 2006 at 05:08:54PM +0100, Steve Kemp wrote:
> On Thu, Aug 31, 2006 at 11:59:51AM -0400, Stuffed Crust wrote:
> > On Thu, Aug 31, 2006 at 03:16:31PM +0100, Steve Kemp wrote:
> > >   Perhaps rename the 3.0 pre-release to 2.9.9 and make a 3.0 codebase
> > >  from scratch using SQLite + HTML::Template, etc?
> > 
> > I'd keep it simpler, actually.  Migrate to SQLite first, make that v3.0. 
>   The migration to sqllite would be fairly simple, agreed.

  Actually thinking about it a little more this falls into two

    1. Changing the tag cache to be a database.

    2. Creating a database schema.

  The latter might be a challenge.  I'm wondering if it makes sense
 to actually serve files by path in the future, since that is what
 allows the basic-auth sidestepping.

  In my newer program I have a schema like this:

  table directory
    int id;
    int parent
    char name

  table songs
    int id;
    int directory;
    char name

  # I can never decide whether to have a table with entries for
  # each song such as "artist", "title", "album", "genre".  So I
  # do it like this which allows future-proofing .. kinda.
  table tags
     int id;
     char name
     char value

   This leads to entries like this:

   # directories beneath /home/mp3
   (1, 0, "Abba" )   => /home/mp3/Abba
   (2, 1, "Hits" )   => /home/mp3/Abba/Hits

   # files
   (1, 1, "Testing.mp3") => /home/mp3/Abba/Hits/"
   (2, 0, "") => /home/mp3/Abba/"

   # tags
   (1, "Artist", "Abba")
   (1, "Album", "Hits")

   (2, "Artist", "Abba" )
   (2, "Album", "Misc collection")

  I then actually stream songs via the file ID, so instead of

    http://server:port/Abba/Blah.mp3   or

   We use links like this:

   I'm not sure if that makes sense globally or not - downsides is that
  there could be spidering done easily.  (Atlhoug "track" could be user
  selected?).  Plus sides is that we don't ever care about HTTP encoding
  decoding, etc.


reply via email to

[Prev in Thread] Current Thread [Next in Thread]