> Exessive interupt over-heads will be apparrent by screen output being > noticably > slowed down whilst data is comming in at high-speed and sudden emptying of the > ISR ring-buffer and fast printing when the other end stops sending after a > screen full.
Yeah, but the fifo will play havoc with my XON/XOFF flow control :( Or maybe it won't... hmmm... > [Screen handling] > Interupts during a PUSH HL... style CLS routine don't corrupt the screen as > on an interupt the PC and whatever reg's get pushed onto the stack which does > cause momentry blobs BUT they're popped back off afterwards and the next CLS's > PUSH HL's erase the blobs.... The only thing to watchout for is IF an interupt > occours just after the final PUSH HL at the top of the screen (as viewed > on-screen) then we'd have to make sure we don't have any important data > there.. Yeah... there shouldn't be anyway... The main problem I've got at the moment is a 30% reduction in efficiency when I put the shadow-screen which I use for the scroll-back buffer. Bother. > > The scripts will be compiled anyway, so it should make things quite nice > > and fast... > > Vroooooom.......... *smiles* I'm even thinking of putting a peephole optimiser in there :) > > [Sam Fax] > No response as yet... The mailer frontend that I'm using on this PeeCee has > an earlier versions C source available and may have the Fax receive functions > in it as it can receive raw Fax T.4 files itself! I'll get hold off the source > tomorrow and have a look see:-) Oooh brillo! Simon
