On Wed, Mar 19, 2003 at 03:02:39PM +0000, Matthew Toseland wrote:
> On Wed, Mar 19, 2003 at 05:46:30PM +1100, fish wrote:
> > I've thrown the code up at http://artificial-stupidity.net/~fish/newstream/.
> > There is old archives there as well... tehy lack the threaded HttpBuffer
> > stuff,
> > which makes some applications (read: freeamp, winamp) happier by handing
> > buffering in a nicer way, because i had some issues with it - it seems to be
> > all fixed now, but who knows.
> >
> > there is a nice readme.txt and stuff, and it should be easy enough to
> > understand, but probably isn't, so, uhm, i expect many stupid fuckups -
> > I do think I got rid of most of the worst fuckups, tho (the aborted
> > rewrite was... ill advised at best. not participating on the network
> > will do that to you :-p I spent all yesterday backing out all of my
> > so-called "clever" ideas :-p)
>
> The README seems to suggest that if you set the client buffer to 10
> blocks, then it will do all 60 requests in parallel rather than queuing
> them... is this correct?
that's correct - basically, there is only one queue size, for both the
output queue and the request queue. it would be easy to change
this, it's just an arbitary choice that I made - all that would have
to change, is make outputBuffer use it's own length parameter rather
than using requestQueueLength or whatever the variable was
- jj
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 230 bytes
Desc: not available
URL:
<https://emu.freenetproject.org/pipermail/devl/attachments/20030321/9715c8b3/attachment.pgp>
