Missing Link UPDATE* Compiler files missing - Exxos

[EstTeeEfEm]

Hi all,
the Missing Link Compiler files are different from the missing link STOS files in MISLINK.ZIP (Full package)....

There are several commands from the registered updated version that can be run internally from STOS but cannot be compiled due to the updated commands not being present in the COMP files /ECQ/ECR/ECS

This leads me to believe that the "MISLINK.ZIP" that is available everywhere is missing the updated compiler files.

Here are the updated commands

Updated command list for The Missing Link (Registration version).
(C) 1993 - Top Notch


Quick reference
---------------
replace range madr,min,max,blk

r = win block amount (madr,x1,y1,x2,y2,blk)
win xy block madr,x1,y1,x2,y2,xadr,yadr,blk,num
win replace blocks madr,x1,y1,x2,y2,blk1,blk2
win replace range madr,x1,y1,x2,y2,min,max,blk

bullet scr,x,y,col

many bob x1,y1,x2,y2,0,0,0,0,0,1
many bob scr,gadr,imgadr,xadr,yadr,statadr,xoff,yoffnum,0
many joey x1,y1,x2,y2,0,0,0,0,0,0,1
many joey scr,gadr,imgadr,xadr,yadr,statadr,coladr,xoff,yoff,num,0
many bullet scr,xadr,yady,statadr,coladr,xoff,yoff,num
many spot scr,xadr,yady,statadr,coladr,xoff,yoff,num

r = many overlap (x1,y1,xadr,yadr,wid1,hig1,wid2,hig2,statadr,
imgadr,stval,imgval,num)

many add xadr,vadr,num,lval,uval
many inc xadr,num,lval,uval
many dec xadr,num,lval,uval

freq = hertz
set hertz freq

raster flag,coladr,line,wid,num,col

display PC1 adr,scr

a = mostly harmless(1,2,3,4,5)

{and also an updated version of...}
t = musauto (adr,num,sz)

Command list

Sorry for the long quote but it seemed like the best way to do it.
RASTER is one command that looks pretty good in STOS as a BASIC run program but cannot be compiled with COMP 2.06 with an extension not present error (Even though Missing Link Ext IS present!) and the COMP 3.00 compiles it but the actual PRG file it generates crashes...

Does anyone have these updated compiler files.. ?

And do we all need to update our MISLINK.ZIP archives?

Thanks in advance..

I've also included the entire ZIP archive of my program for your OWN testing... and to make sure I'm not going mad
PLEASE NOTE: You will need to alter the location within the listing of the files themselves if you wanna run it.

[exxos]

I've seen that many version of the missing link floppies its hard to know whats going on, assume you downloaded the ones off my site ? http://www.exxoshost.co.uk/atari/STOS/E ... /index.htm

Odd that the "many bob" command is missing, assume you got all the link complier and stos extension numbers the same ? Also it got confusing at some point as some commands where in the "Misty" extension.

Its possible at some point when I merged all the missing link files into one folder I ended up with a demo version, I do still have my master disk somewhere, so I can update the archive if needs be.

IIRC you had to run a = mostly harmless(1,2,3,4,5) to "register" the extension. I know I had it in my autoexec.bas but can't remember if it effected the compiler part or not.

Looking on my PC, the only difference I see in files is link1.ecs where I have 17k and 18k file size versions, but the 17k one is dated 1998, the 18k 1995.

[EstTeeEfEm]

I just took MISLINK.ZIP from your achive again, just to be sure as I couldnt be sure it didnt come from another source.

Odd that the "many bob" command is missing, assume you got all the link compiler and stos extension numbers the same ?

That, I couldnt tell you I afraid, what i CAN tell you is that within your archive the UPDATE.DOC file is there so I assume it maybe a simple case of the wrong version of the compiler extension vs stos extension?

Just loaded the DEMO version from STformat to be sure it wasnt the demo
The a=Mostly Harmless(1,2,3,4,5) ... Yeah, registration only got rid of that annoying message after every 10 lines or something Also, in the "demo" version the WORLD commands are disabled so I at least know I have (and you have in your archive) a full version at least.

It does seem more likely to be an older version of the compiler extensions.

Anyway, It will probably take some time because if you have several versions then you'll have to compile something with the new commands in, just use my source, at least it contains "updated" commands and works in runtime.

[exxos]

You have a few problems..

Here is my quick hash of your version..

Code: Select all

10 key off : curs off : hide : mode 0
20 remmouseoff
30 gosub 5030
40 reserve as screen 14
50 load "tiles.pi1",14
60 get palette (14)
70 load "rastblue.pi1",logic
80 load "blocks.mbk",5
90 load "level1_1.mbk",6
100 world 16,0,304,200,0,1
110 logic=back
120 WX=0 : WY=0
130 raster 1,start(15)+2,1,15,13,15
140 world logic,start(5),start(6),WX,WY,0
150 A$=inkey$
160 if A$=" " then goto 200
170 if A$="l" then WX=WX+2
180 screen swap : wait vbl 
190 goto 140
200 raster 0,0,0,0,0,0
210 remmouseon
220 default : end 
5030 reserve as work 15,512*2
5040 for R=0 to 7 : for G=0 to 7 : for B=0 to 7
5050 doke start(15)+H,R*256+G*16+B
5060 H=H+2
5070 next B : next G : next R
5080 return 

You basically confuse the raster data with the background image, this is what is causing the crash once compiled. You are loading your image into bank 15, then calling the raster function to read colour values, which are probably not in the format the raster code is expecting. What I did was to load your background image into the logic screen, then reserve bank 15 and doke some values in there as per the ML example code. This compiles and works, but it needs fixing more..

ml_00001.png

I ran it compiled in the editor that way, the PRG seemed to get stuck on the rasters alone, but its only a quick hash. It actually flickers like hell, but that is because some things are messed with But at least it should give you a step in the right direction.

I think your raster colour (15 I think) should be black and the whole of your background image should be black with colour 15. This way the rasters should show up more as you expect. You will also need to work out your colour values and doke them into bank 15 (lines 5030+).

So its not actually a compiler problem, its basically as the raster function doesn't have correct data and crashes in the compiled version. As to why it works in the editor is probably a fluke

The colour format I assume matches this format "xRRRxGGGxBBB" So you may want to try some things out to confirm that is the format the raster function uses.

I have also re-done the mslink.zip on my site, I have taken my ML floppy (as supplied by goodmans) and unpacked and extracted them back into the zip on my site just to be sure everything on there is off the master disk. I think the copy you had should be ok anyway.

[EstTeeEfEm]

Thanks for that and sorry for the LOOONG delay in replying

You basically confuse the raster data with the background image, this is what is causing the crash once compiled. You are loading your image into bank 15, then calling the raster function to read colour values

Yea, I basically made a binary file with the colour data in now, should be OK when compiled but haven't tested it yet.

Went for this format

Code: Select all

10  RESERVE AS WORK 1,96
20 BLOAD "RASTER.BIN",1

The raster bin is in the format xRGBxRGB as the raster function requires.. which i tried to do by nicking the colours from a background PI1 ( "start+2" .. the first 16 bytes) .. probably screwed it up

RASTER allows you to "rastorize" one colour out of the on-screen or physical palette so i decided to use colour 1, if you use colour 0 then its the background hence why i originally used colour 15 because the tiles used colour 0 as its OWN background. jeez this gets complicated!!

Thanks for adding the missing link files into your site, even if they are the same, its always better to be safe than sorry, at least we KNOW we're working from the same page, so to speak.

And don't even get me started on the stars extension with the raster code... everything went a little Jeff Minter

[EstTeeEfEm]

Here's an updated PRG you can run,

HISCORE.rar

demonstrating the Raster as a compiled program...

DEFINITELY the screwy way i picked the colours out of a PI1 file that knacked the compiled version!!

Damn right a fluke, pretty good one mind, had me fooled for ages ?!

Just hit space to exit, its a test HISCORE page I may use...

[exxos]

One thing you might want to test out with your rasters is they are very CPU timing dependent. When you start with moving stuff on the screen such as sprites or something, your rasters might "judder" So you may want to double check before doing to much with it..

[exxos]

Here's an updated PRG you can run,
demonstrating the Raster as a compiled program...

That looks pretty awesome

DEFINITELY the screwy way i picked the colours out of a PI1 file that knacked the compiled version!!
Damn right a fluke, pretty good one mind, had me fooled for ages ?!

Yeah, I've been caught out by silly things like that in the past