>> Would this be an acceptable solution? Should I look further at it? As
>> is now, I have a simple Win32 build that works...
>Any patches in that direction would be very welcome. There is already
>an initial implementation of a basic Ogg parser in the apps/metadata.c
>file - this was written without reference to the Tremor sources, but
>does (hopefully) find the correct sample count (which is very important
>if you want the correct track length to be displayed).
Wouldn't time information in millisecond resolution be enough? That still works
>The important issue to consider is the requirement for Rockbox to use as
>little memory as possible. I haven't looked at the Tremor
>implementation of tag parsing, but we need a streamlined parser that has
>very minimal memory usage.
Well, Tremor is designed for embedded use, so it shouldn't be that bad, I
hope. But I haven't looked much at it yet, I must admit. :) And there's
that thing with Rockbox and malloc (or the lack thereof) that needs to be
handled as well.
>By all means reuse code already in Tremor, but it may be a better idea
>to copy and paste into new source files, rather than using the source
Actually, I looked a bit at that first, but it wasn't really a good option either.
You need most of the stuff in the four files I mentioned, so it seems easier,
from a maintenance point of view, to just add that relatively small number
of defines (a dozen, to be exact).
Received on Mon Jun 20 13:26:01 2005