2007-04-16, 13:07
i've moved the nntp communication stuff and deEnc stuff to use file pointers instead of doing anything in RAM, so i am almost ready to set up the functionality for multithreading... i've still got some issues however!
there is some curruption in the yDecoded content that i am downloading & decoding... with mp3s, this results in little skips that you can here. the mpg123 system is also complaining (at the same time as these skips) that it can't rewind by X bits... if i do a diff on the mp3 file; compiring the decoded content i am getting to the same file downloaded through newspro, there are a few differences alright... i spent most of yesterday trying to debug this but all i know so far is that i think my decoder subroutine is ok!.. doh!
gonna check next that the downloaded encoded content is correct and the corruption isn't introduced before the decoder gets a chance to do its work! if the answer doesn't become obvious to me within the next hour or two of looking at it, i'm going to skip it for now and revisit it later, while i impliment the multi threaded-ness.
i figured also that it's not really strictly necessary to use multi-threaded nntp streams for the reasons i had originally thought... the yEnc'd content can be decoded on the fly pretty easily without this... it was mentioned however that some free/isp nntp services give you 4 connections with each of the connection speed capped to a certain bandwidth... i will therefore impliment the threading functionality anyway to make sure we can get the fattest data stream possible, and to make sure that as a result, streaming will work for the most people....
will post when i have more news!
Patrick.
there is some curruption in the yDecoded content that i am downloading & decoding... with mp3s, this results in little skips that you can here. the mpg123 system is also complaining (at the same time as these skips) that it can't rewind by X bits... if i do a diff on the mp3 file; compiring the decoded content i am getting to the same file downloaded through newspro, there are a few differences alright... i spent most of yesterday trying to debug this but all i know so far is that i think my decoder subroutine is ok!.. doh!
gonna check next that the downloaded encoded content is correct and the corruption isn't introduced before the decoder gets a chance to do its work! if the answer doesn't become obvious to me within the next hour or two of looking at it, i'm going to skip it for now and revisit it later, while i impliment the multi threaded-ness.
i figured also that it's not really strictly necessary to use multi-threaded nntp streams for the reasons i had originally thought... the yEnc'd content can be decoded on the fly pretty easily without this... it was mentioned however that some free/isp nntp services give you 4 connections with each of the connection speed capped to a certain bandwidth... i will therefore impliment the threading functionality anyway to make sure we can get the fattest data stream possible, and to make sure that as a result, streaming will work for the most people....
will post when i have more news!
Patrick.