Article 33356 of comp.sys.amiga.programmer: Path: news.csos.orst.edu!flop.ENGR.ORST.EDU!ogicse!uwm.edu!wupost!darwin.sura.net!jabba.ess.harris.com!mlb.semi.harris.com!usenet.ufl.edu!cybernet!conners From: conners@cse.fau.edu (Sean "Captain Napalm" Conner) Newsgroups: comp.sys.amiga.programmer Subject: Re: SOURCE: 8-Way-Smooth-Scrolling Message-ID: <1993Jun12.072953.28871@cybernet.cse.fau.edu> Date: 12 Jun 93 07:29:53 GMT Article-I.D.: cybernet.1993Jun12.072953.28871 References: <1v9ki1INNcv1@fstgds15.tu-graz.ac.at> <1993Jun12.045539.5338@sol.cs.wmich.edu> Sender: news@cybernet.cse.fau.edu Organization: Florida Atlantic University, Boca Raton, FL, USA Lines: 59 In article <1993Jun12.045539.5338@sol.cs.wmich.edu> 22dussia@sol.cs.wmich.edu (Demetri) writes: >In article <1v9ki1INNcv1@fstgds15.tu-graz.ac.at> hollosi@fm11ap02.tu-graz.ac.at (Arno Hollosi) writes: >>Hi ! >> >>If you are interested in an example source for an >>smooth 8-Way-Tile-Scroller, then check out 'alt.sources.amiga'. >> >>It isn't recommended for OS-fanatics, because of hardware-banging. > ^^^^^^^^^^^ > hehe, And I thought you said you where doing this WITH the OS routines! > Looks like I misread your posts! I always knew going direct to Copper to > change the screen pos was fast... :') > >> >>It uses the copper-wrap-method and is really FAST. >>On my A4000 it needs 10 (TEN!) rasterlines to scroll in >>X and Y direction with 4 pixel/frame. >> >>So check it out. >>Comments and criticisms are welcome. >> >>Cheers >>Arno >> > I think lots of people will check it out, and I am sure you will hear from >them! > And here's one of them 8-) Yes, I got the code, downloaded it, took some time to re-work some of it so it would assemble on my 68000 based Amiga 500 (with 3M (512K Chip) RAM and 50M harddrive still running 1.3). It's nice to know that it runs on your A4000, but it promptly Guru'ed out on my system (with 00000004.xxxxxxxx - Illegal instruction). On closer in- spection of the code, I found the following instruction: clr $88(a6) ; A6 - custom chips BTW ... May I quote from the Amiga Hardware Manual that I got for 1.3? NOTE When strobing any register which responds to either a read or a write, (for example copjmp2) be sure to use a MOVE.W #$00, not CLR.W. The CLR instruction causes a read and a clear (two accesses) on a 68000, but only a single access on 68020 and above. This will give different results on different processors. And $88(a6) is copjmp1 - BAD PROGRAMMER! BAD PROGRAMMER! 8-) 8-) And no, when I fixed that did it work. It still bombed. As far as I can trace it, it appears that you are calling 0(execbase) (from the PC of the illegal instruction, that's the base of Exec on my system). I just now found out it was my assembler (but partially your code, the assembler I use is a bit fussy about spaces). But I'll try again and call back later 8-) -spc (Calling up my Unix box at work, editing your code, and assembling, why do you like killing off this wonderful OS? 8-) Article 1221 of alt.sources.amiga: Path: news.csos.orst.edu!flop.ENGR.ORST.EDU!rutgers!uwm.edu!cs.utexas.edu!math.ohio-state.edu!darwin.sura.net!paladin.american.edu!news.univie.ac.at!fstgds15.tu-graz.ac.at!fm11ap07.tu-graz.ac.at!hollosi From: hollosi@fm11ap07.tu-graz.ac.at (Arno Hollosi) Newsgroups: alt.sources.amiga Subject: Re: SOURCE: 8-Way-Smooth-Scrolling Message-ID: <1vhlt3INNe0u@fstgds15.tu-graz.ac.at> Date: 14 Jun 93 11:02:59 GMT References: <1v9juvINNcv1@fstgds15.tu-graz.ac.at> Organization: Technical University of Graz, Austria Lines: 21 NNTP-Posting-Host: fm11ap07.tu-graz.ac.at X-Newsreader: TIN [version 1.2 PL0] Arno Hollosi (hollosi@fm11ap02.tu-graz.ac.at) wrote: : This is an example of how to code an smooth 8-Way-Tile-Scroller. : It's hardware banging, so it isn't recommended for OS-fanatics. ATTENTION! There's a bug in '8wayscroller.lha' that will cause a GURU on any machine with an CPU <'020. Somewhere in the 'Init_copper'-routine I do a CLR.W COPJMP1. (look for clr $88(a6)). As I wasn't aware of this problem, CLR makes a READ and WRITE cycle on an 68000. But COPJMP1 is only a STROBE-Register, so the READ-cycle can crash (and does so) your machine. So please change it to: MOVE.W #0,$88(a6) It should now work properly. Cheers Arno Article 33448 of comp.sys.amiga.programmer: Path: news.csos.orst.edu!flop.ENGR.ORST.EDU!gaia.ucs.orst.edu!umn.edu!news-feed-1.peachnet.edu!gatech!paladin.american.edu!news.univie.ac.at!fstgds15.tu-graz.ac.at!fm11ap01.tu-graz.ac.at!hollosi From: hollosi@fm11ap01.tu-graz.ac.at (Arno Hollosi) Newsgroups: comp.sys.amiga.programmer Subject: Re: SOURCE: 8-Way-Smooth-Scrolling Date: 15 Jun 1993 10:38:55 GMT Organization: Technical University of Graz, Austria Lines: 85 Message-ID: <1vk8rvINNt03@fstgds15.tu-graz.ac.at> References: <1v9ki1INNcv1@fstgds15.tu-graz.ac.at> <1993Jun14.103149.29744@cc.ic.ac.uk> NNTP-Posting-Host: fm11ap01.tu-graz.ac.at X-Newsreader: TIN [version 1.2 PL0] Hi again! I updated my source-code, so that it can be assembled with MACRO68 too. This means it doesn't contain any SEKA/AsmOne- specific features anymore. Ofcourse the bug was fixed. I uploaded this new version (V1.04) onto AMINet today. You should find it in pub/aminet/dev/src in some days in an AMINet-site near you. If you can't wait, check out 'alt.sources.amiga'. Someone asked me how to include copperbars (Turrican,Lionheart,..) Here is two possible methods: FIRST (slower) method: ====================== For each line you have: WAIT $yy0f,$fffe MOVE $0180,color (copper-instructions ofcourse). But you have to do the copperwrap too! So you insert the following at the Y-wrap-postion: MOVE $00e0,pl0pth MOVE $00e2,pl0ptl .... MOVE $00f2,pl4ptl But to insert something into to right position can be very nasty. SECOND (I think this one is faster) METHOD: =========================================== For each line you have: WAIT $yy0f,$fffe MOVE $0180,$fffe MOVE $00e0,pl0pth MOVE $00e2,pl0ptl ... MOVE $00f2,bpl4ptl Ofcourse you need only to set the BPL-pointers at the copper-wrap-position so you do the following: In every line you do instead of: MOVE $00e0,pl0pth you do this: MOVE $0082,addr_of_next_WAIT_command ($82 is the LOW-word of the Copper1-PC) This MOVE performs a JUMP to the next WAIT. (no move into COPJMP1 necessary!) BUT ATTENTION: the whole copperlist must be in the same 64K (this means the high-word must always be the same.) Just to summarize: Build copperlist with MOVE $0082,addr_next_WAIT In every frame do: restore MOVE $0082,addr_next_WAIT at the old wrap-postion. at new wrap-position: change MOVE $0082 INTO MOVE $00e0,pl0pth (And the best: you never have to change or rebuild the other BPL-pointers ($e2-$f2)) This method only needs approx. 1 rasterline! FAST!! Cheers Arno P.S: my EMail account is valid until 1.7.93 only. Article 1226 of alt.sources.amiga: Path: news.csos.orst.edu!flop.ENGR.ORST.EDU!gaia.ucs.orst.edu!news.reed.edu!batcomputer!caen!uwm.edu!cs.utexas.edu!usc!howland.reston.ans.net!xlink.net!scsing.switch.ch!news.univie.ac.at!fstgds15.tu-graz.ac.at!fm11ap01.tu-graz.ac.at!hollosi From: hollosi@fm11ap01.tu-graz.ac.at (Arno Hollosi) Newsgroups: alt.sources.amiga Subject: Re: SOURCE: 8-Way-Smooth-Scrolling Date: 15 Jun 1993 10:00:10 GMT Organization: Technical University of Graz, Austria Lines: 341 Message-ID: <1vk6jaINNt03@fstgds15.tu-graz.ac.at> References: <1v9juvINNcv1@fstgds15.tu-graz.ac.at> <1vhlt3INNe0u@fstgds15.tu-graz.ac.at> <1993Jun14.115031.13139@news.uit.no> NNTP-Posting-Host: fm11ap01.tu-graz.ac.at X-Newsreader: TIN [version 1.2 PL0] Hi again! This should be the final version of my 8-Way-Tile-Scroller. I fixed the bug. After some work it can now be assembled with MACRO68 too. This means it is free of any SEKA/AsmOne-features. Enjoy! Cheers Arno P.S: Comments and criticisms welcome. (my account is valid until 1.7.93) [DR's Notes: The UUENCODED source file that was here has been removed and UUDECODED for you, it can be found in Files/Miscellaneous as 8WayScroller104.LHA] Article 33483 of comp.sys.amiga.programmer: Path: news.csos.orst.edu!gaia.ucs.orst.edu!wsrcc.com!fastrac.llnl.gov!usenet.ee.pdx.edu!cs.uoregon.edu!ogicse!uwm.edu!cs.utexas.edu!math.ohio-state.edu!caen!nigel.msen.com!yale.edu!yale!gumby!wmu-coyote!22dussia From: 22dussia@sol.cs.wmich.edu (Demetri) Newsgroups: comp.sys.amiga.programmer Subject: Re: SOURCE: 8-Way-Smooth-Scrolling Message-ID: <1993Jun15.231306.911@sol.cs.wmich.edu> Date: 15 Jun 93 23:13:06 GMT Article-I.D.: sol.1993Jun15.231306.911 References: <1v9ki1INNcv1@fstgds15.tu-graz.ac.at> <1993Jun14.103149.29744@cc.ic.ac.uk> <1vk8rvINNt03@fstgds15.tu-graz.ac.at> Organization: Western Michigan Univ. Comp. Sci. Dept. Lines: 37 In article <1vk8rvINNt03@fstgds15.tu-graz.ac.at> hollosi@fm11ap01.tu-graz.ac.at (Arno Hollosi) writes: >Hi again! > >I updated my source-code, so that it can be assembled with >MACRO68 too. This means it doesn't contain any SEKA/AsmOne- >specific features anymore. Ofcourse the bug was fixed. Junk SEKA/AsmOne! Macro68 is the best. As for the bug...well...sorta fixed (or a new bug...) What happens to my system (NTSC A2000/'040 Kick2.04 OS2.1) is that AFTER your run and EXIT your "demo", The copper is not set correctly, since I just get a BLACK screen, but I can still type commands in, and they do execute... Now, looking at your code, find the part where you have : jsr LoadView(a6) move.l copinit(a6),$dff080 And INSERT these lines to fix the problem... move.l #$dff000,a6 move.l gfxbase,a0 move.l $26(a0),cop1lc(a6) move.l $32(a0),cop2lc(a6) And this will fix the copper correctly by using the SYSTEM copper view... Now, when I run your demo, with those lines added, I get my screen back on exit! Now a comment on that demo, I have done a much faster pixel scroll to achieve a "dizzy" effect, why do you limit the max speed to 2 pixels? I like *REAL* fast scrolling, I think you should add in some code to allow a pixel delay from 0 to 15 pixels... :') Later.. >Cheers >Arno > >P.S: my EMail account is valid until 1.7.93 only. Oh, where can we reach you after this date? Article 33502 of comp.sys.amiga.programmer: Path: news.csos.orst.edu!flop.ENGR.ORST.EDU!rutgers!gatech!howland.reston.ans.net!paladin.american.edu!news.univie.ac.at!fstgds15.tu-graz.ac.at!fm11ap01.tu-graz.ac.at!hollosi From: hollosi@fm11ap01.tu-graz.ac.at (Arno Hollosi) Newsgroups: comp.sys.amiga.programmer Subject: Re: SOURCE: 8-Way-Smooth-Scrolling Message-ID: <1vms6mINNfml@fstgds15.tu-graz.ac.at> Date: 16 Jun 93 10:21:10 GMT References: <1v9ki1INNcv1@fstgds15.tu-graz.ac.at> <1993Jun14.103149.29744@cc.ic.ac.uk> <1vk8rvINNt03@fstgds15.tu-graz.ac.at> <1993Jun15.231306.911@sol.cs.wmich.edu> Organization: Technical University of Graz, Austria Lines: 35 NNTP-Posting-Host: fm11ap01.tu-graz.ac.at X-Newsreader: TIN [version 1.2 PL0] : move.l #$dff000,a6 : move.l gfxbase,a0 : move.l $26(a0),cop1lc(a6) : move.l $32(a0),cop2lc(a6) Never heard of this problem before. Hmm. My program NEVER changes COP2LC. So why must you set it back ?? I thought Loadview(OldView) would do this for me!!! BTW, I only used PAL-Amigas to check it. : to achieve a "dizzy" effect, why do you limit the max speed to 2 pixels? I don't know if you've read the README-file, but there you would have read the following: PRESS FIRE to double your speed !!!!! So it means MAX. 4 pixel/frame, which is really FAST. If this is still too slow for you, then just change 'blt_cntx', 'blt_cnty' and the according tables 'blt_tabx' and 'blt_taby'. You should be able to manage MAX. 16 pixel/frame. So try it yourself. It isn't hard to do. I PROVIDED THE SOURCE-CODE, SO THAT YOU CAN MAKE YOUR OWN CHANGES. : >P.S: my EMail account is valid until 1.7.93 only. : Oh, where can we reach you after this date? From July to September I only can be reached via SNAIL-MAIL. Address: Arno Hollosi, Oberndorf 313, A-6322 Kirchbichl, AUSTRIA In October I'll get a new account here at my university. My username will be hollosi too, but the @fm11ap01.tu-graz.ac.at will change. Cheers Arno Article 33600 of comp.sys.amiga.programmer: Path: news.csos.orst.edu!flop.ENGR.ORST.EDU!ogicse!uwm.edu!math.ohio-state.edu!sdd.hp.com!portal!danb From: danb@shell.portal.com (Dan E Babcock) Newsgroups: comp.sys.amiga.programmer Subject: Re: SOURCE: 8-Way-Smooth-Scrolling Message-ID: Date: 18 Jun 93 04:45:37 GMT Article-I.D.: unix.C8svw6.C2F References: <1vk8rvINNt03@fstgds15.tu-graz.ac.at> <1993Jun15.231306.911@sol.cs.wmich.edu> <1vpjubINN5hs@fstgds15.tu-graz.ac.at> Sender: news@unix.portal.com Organization: Portal Communications Company -- 408/973-9111 (voice) 408/973-8091 (data) Lines: 22 Nntp-Posting-Host: jobe In article <1vpjubINN5hs@fstgds15.tu-graz.ac.at> hollosi@fm11ap03.tu-graz.ac.at (Arno Hollosi) writes: >Demetri (22dussia@sol.cs.wmich.edu) wrote: > >For everyone who uses MACRO68! >In the README-file I wrote, that you shouldn't use resident MACLIBs. >This is, because MACRO68 seems to have some problems with 'em. >Demetri mailed me again, and as I thought it was the following: > MACRO68 assembled the following line incorrect. > > move.l copinit(a6),$dff080 > MACRO68 (using resident MACLIBs) made > > move.l $00(a6),$dff080 (0 ???!!) > instead of > > move.l $26(a0),$dff080 > >So be careful with the resident MACLIBs. >I don't know, if MACRO68 has got a BUG, or if the MACLIBs were not >compiled correctly. Neither. Your code is wrong. The symbol you want is "gb_copinit". "copinit" is indeed zero (defined in graphics/copper.i). Dan