[ExtractStream] Random SplitStream Thoughts...

[Roger Merchberger]

--- Mark Jeffcoat <jeffcoat@a...> wrote:
> On Sat, Feb 09, 2002 at 07:41:58AM -0800, Roger Merchberger wrote:
> > The ones I understood were *super* helpful - a couple went over
> me
> > head... such a long road, but like the roads of Basic, 6809
> assembly,
> > perl, cobol, APL, JCL and to a lesser extent Pascal, Lisp,
> Logo...
> > one must start learning & keep using it enough to get it beat
> into
> > this tired old brain of mine... (why is 6809 index register
> > redirection & writing 8-bit position independant code so easy for
> me,
> > but C pointers so frelling hard???)
> 
> C is easy, but most of the programs that show up here are
> difficult, crammed with magical numbers and unmotivated 
> algorithms necessary to appease the MPEG gods. Ick.

C *could* be easy, if I did more than 2 weeks programming in it every
3 years or so... That's been my problem - I've needed to "re-learn"
it on occasion, but I've never had enough time to fully "learn" it...

Yes, and harkening back to my halycon days of youth... back when you
used to get a "memory map" which explained graphically the layout of
the system memory - I really could use one of those to explain MPEG2
for me...

I looked at the mplex source, and I heard this odd Wooooooshhhhing
sound again... Right over my head... :-/

> I've been toying with splitstream recently, too; mostly
> trying to cut parse_chunk() into bit-sized pieces. I may
> post something more formal in a few days, but feel free
> to send me an email if you'd like to see the work in progress--
> you may find it more easily digestible than what you're
> looking at now.

At least from what I saw, a goodly chunk of parse_chunk() was
error-checking & splitting out what it percieves as bad chunks... but
I'd need to go further, deeper...

> The memory problems you're looking at are probably red
> herrings. Someone else has pointed out that you operating
> system should clean up un-freed memory no matter how the 
> program exits, so the only real concern is making sure
> that programs don't end up wasting so much memory while
> they're running that they start using unnecessary swap space.

Yea, but remember: This *is* Bill Gates we're talking about...

And at least in Dos-mode-based programs, I don't think they can
allocate swap space... but I can't test that as I don't have a
computer with "little" enough RAM to find out, it seems...

> I don't think this is a major concern for splitstream, but
> if you're interested in a permanent fix, look into giving
> up manual memory management alltogether:
> http://www.hpl.hp.com/personal/Hans_Boehm/gc/

It prolly isn't, but 1) I've always been good at debugging, and 2)
things that might not be bad in a small program can be *disastrous*
in a large program, and I don't want to get in the habit of doing bad
things & have it turn around and bite me in the tookus...

I took a quick peek at that page, and learned just enough to realize
that I need to look more at that page... ;-)

Anyway - I must be mozeyin...
Roger "Merch" Merchberger

__________________________________________________
Do You Yahoo!?
Send FREE Valentine eCards with Yahoo! Greetings!
http://greetings.yahoo.com