Rockbox mail archive

• Next message: Pedro: "Re: Some Jukebox and Rockbox issues"
• Previous message: Greg Haerr: "Re: cvs patch for loadable fonts"
• In reply to: Daniel Stenberg: "RE: Rockbox for the player"
• Next in thread: Ed Wilson: "RE: Rockbox for the player"
• Messages sorted by: [ date ] [ thread ] [ subject ] [ author ] [ attachment ]

Since we know the seed used to shuffle, can we somehow
backtrack through the shuffle process and find out
what the original index was without storing all of the
original indices?

Basically have a function like this:
int playlist_original_index(int seed, int
shuffled_index);

I think we would need to find the inverse of the seed
used. I don't remember much from my number theory
class and don't know much about the Mersenne twister
we use, so I can't "show you the code" just yet.
(Sorry, Linus :)

-Lee

--- Daniel Stenberg <daniel_at_haxx.se> wrote:
> On Thu, 12 Sep 2002, Ed Wilson wrote:
>
> > 1) When showing the playlist number 4/1567 etc..
> the 4 is the current count
> > of how many tracks have been played, but on the
> old firmware it referred to
> > the position of the song within the playlist which
> was much easier to
> > remember so that the song could be removed from
> the playlist at a later
> > date.
>
> The superb (if I may say so) efficiency of playlist
> in Rockbox of course come
> with a price.
>
> We only keep a single list in memory, and that list
> is shuffled when you
> select shuffle. When advancing down this list, we
> have no knowledge at all
> about what "original" position a particular index
> has.
>
> To be able to know that, we'd have to store an
> additional 16 bit number for
> each playlist entry. Quite doable, but not done
> today.
>
> --
> Daniel "Bagder" Stenberg -- http://rockbox.haxx.se/
>

__________________________________________________
Do you Yahoo!?
Yahoo! News - Today's headlines
http://news.yahoo.com

• Next message: Pedro: "Re: Some Jukebox and Rockbox issues"
• Previous message: Greg Haerr: "Re: cvs patch for loadable fonts"
• In reply to: Daniel Stenberg: "RE: Rockbox for the player"
• Next in thread: Ed Wilson: "RE: Rockbox for the player"
• Messages sorted by: [ date ] [ thread ] [ subject ] [ author ] [ attachment ]