On (19 Jan 95) [EMAIL PROTECTED] wrote in response to [EMAIL PROTECTED] > From: Simon Cooke <[EMAIL PROTECTED]> > Date: Thu, 19 Jan 1995 12:03:41 +0000 (GMT)
> Right, well, I'm using RXRDY mode at the moment -- I didn't think I could > get away with FifoFull, in case only 2 bytes were received -- it'd just > hang, with no data coming back. I know that this isn't likely, but it > came to mind... Oh you're quite right it could be a problem... But not that serious;-) When you check for character availability you first check the Ring-buffer and if none there then disable the interupts and check the UART RxRDY bit then re- enable the interupts... When Reading a character do the same as above but get the character from which- ever has a character character first.. More software overhead in bytes but they're only executed *when* you check/read a character and not whilst you're trying to process the data... > Got to remember that screen clearing and scrolling in my code is all done > using LDI's -- I can't afford to switch interrupts off for it. And the > colour encoding gets in the way a tad... still, it's all got to be > optimised, but I reckon the optimisations will be cancelled when I put > the shadow text screen in (for the scroll-back buffer, and string > grabbing, etc) Under ProDos1923 apart from momentary DI's during context switches the only time I disable the interupts is during the disk-allocation find-first free block routine as that places the stack on top of the actual allocation vector bit-map and uses non-standard methods to search through it.. Interupts whilst in that routine would be fatal! Note I don't even Disable the interupts when doing the fast PUSH HL... style screen clear routine. I'd suggest that when hooking in scripts that you give the screen the least priority ie force the most time efficient emulated screen mode ie No colour, ANSI filtering rather than calculations and a 24x64 or 85 if there's time... Or even avoiding screen output until an error ocours in the script! ie keep the script-mode code seperate from the hands-on mode or both will suffer! [Sam Fax] > *grins* Yeah, Class 2 is a lot better documented than Class 1 as well in > the docs I've got, so it's not only easier to use, it's also easier to > understand! >> I know where I can get hold of the author of the MS-DOS BG-FAX or the QL's >> FAX program. Would you like me to see if they have any doc files that'll >> assist you? > Pretty pretty please!!!! It'll take some time as they're on Fidonet, I'll get right on it:-) (I hope the docs aren't too big;-) > Can you drop me your home address so I can send you an alpha copy of > Termite and see if you've got any ideas??? Will email it:-) regards Johnathan. ... Taglines: Your last chance to say something stupid! -- |Fidonet: Johnathan Taylor 2:2501/307.9 |Internet: [EMAIL PROTECTED] | | Standard disclaimer: The views of this user are strictly his own.
