Path: chuka.playstation.co.uk!news From: "Steve Dunn" Newsgroups: scee.yaroze.programming.codewarrior Subject: Re: MWDebugIO Date: Thu, 2 Jul 1998 11:12:01 +0100 Organization: Backroom Software Ltd. Lines: 96 Message-ID: <6nfmlu$60l1@chuka.playstation.co.uk> References: <359A3B82.2267A83D@easynet.co.uk> <359A43E3.CD8101E3@hinge.mistral.co.uk> <359A4F13.6927AD1A@easynet.co.uk> <6ndqd5$sub14@chuka.playstation.co.uk> <01bda59b$bafd0560$f30b0a0a@Angela1.intelligent-group.com> NNTP-Posting-Host: usera507.uk.uudial.com X-Newsreader: Microsoft Outlook Express 4.72.2106.4 X-MimeOLE: Produced By Microsoft MimeOLE V4.72.2106.4 >I don't think __rts_info_t__.h appears in the Yaroze distribution Steve. >Perhaps >you're seeing things yes? > erm, think it did. If it didn't come with the original, it must have been put there when I inadvertently visited black-bag or something. >You're talking a load of shit. But the re-order the .data to the end is a >good idea, aha, excuse me while I visit Ladbrokes to pick up my winnings! '1-5' on Craig being the first person to say I was in fact talking 'sh!t' ! :) (Didn't expect the 'good idea' bit at the end though ! :) > >Stop being such a wuss and get the serial ARS support done - I might even >give you a hand >and do the PSX side libs if you're really lucky. > Would be nice to get it done. As soon as I'm sacked from my day job (shouldn't be too long now), I'll get straight to it. > >Craig Steve Craig Graham wrote in message <01bda59b$bafd0560$f30b0a0a@Angela1.intelligent-group.com>... > > >Steve Dunn wrote in article ><6ndqd5$sub14@chuka.playstation.co.uk>... >> Just a 'wild' (very wild) thought, don't know if it'll work or not... >> >> use __declspec( bss ) for all 'disposable' C structs created with >> bin2hex, and then use __declspec( sbss ) for the rest of the unitialised >> data... >> >> For instance... >> __declspec(bss) char c[]={"1234567890"} ; (this would be the tim file as >an >> array) >> __declspec(sbss) int nLives ; > >Don't do this. the bss is the uninitialised data segment (things are layed >out in it >by the linker, but no data is placed in it and I don't think it even >appears in the PXE >file - in fact, a bug in one revision of the CW linker was that the bss >actually did appear >in the final file, causing huge PXE's to be produced. They fixed that >problem though, >and I don't think it ever appeared in a Yaroze release of the compiler >anyway. > >> Next thing to do is load all the disposable stuff into VRAM that's >contained >> in the bss section and (here comes the main bit I'm not sure about) tell >the >> psx that the heap starts at the 'start' of the bss section. >> Something like this... >> InitHeap( _fbss, newheapsize ) ; >> >> Another 'possible' alternative is to shuffle everything around (data, >sdata, >> sbss, and bss), so that '.data' is at the end, then init your heap all >over >> it when you've finished loading what was in it. Look in >__rts_info_t__.h. >> There lots of interesting pointers in there. > >I don't think __rts_info_t__.h appears in the Yaroze distribution Steve. >Perhaps >you're seeing things yes? > >> As I said, just a thought. Someones soon gonna come along and tell me >I'm >> talking a load of sh!t :) > >You're talking a load of shit. But the re-order the .data to the end is a >good idea, >only problem is in guaranteeing that the TIM's are at the end of the data >section >(link in the file with them in it last in the link order pane?). > >> Steve > >Stop being such a wuss and get the serial ARS support done - I might even >give you a hand >and do the PSX side libs if you're really lucky. > >Craig.