| |
zzarko
Registered: Feb 2003 Posts: 77 |
Starting a program after copying from cartridge
I am working on a 3.0 version of Magic Desk Cartridge Generator and I have come up with a problem with three programs (testing is done in VICE 3.3, and on a real C64):
Turbocopy V5.0
Turbo Nibbler V5.0
64 Tester
that simply would not start when copied from cartridge to $0801 (the other ~140 programs that I have tried worked just fine).
Anyhow, cold start is like this:
stx $d016
jsr $FDA3 // IOINIT, init CIA,IRQ
jsr $FD50 // RAMTAS, init memory
jsr $FD15 // RESTOR, init I/O
jsr $FF5B // SCINIT, init video
cli
jsr $E453 // load BASIC vectors
jsr $E3BF // init BASIC RAM
jsr $E422 // print BASIC start up messages
ldx #$FB // init BASIC stack
txs
Then I setup IRQ, VIC and CIA registers to my liking. After a program is selected from cartridge menu, I restore original IRQ and call these:
jsr $FDA3 // IOINIT, init CIA,IRQ
jsr $FD15 // RESTOR, init I/O
jsr $FF5B // SCINIT, init video
cli
jsr $E453 // load BASIC vectors
jsr $E3BF // init BASIC RAM
jsr $A68E // set current character pointer to start of basic - 1
ldx #$FB
txs
then I copy the program to its address ($0801), set $2d/2e to end of program, and then:
jsr $A871 // clr
jsr $A533 // re-link
jmp $A7AE // run
Does someone knows what initialization am I missing here (or have too much of it)? Complete source code will be available soon, but I would like to fix this before the release, if possible. |
|
... 11 posts hidden. Click here to view all posts.... |
| |
Krill
Registered: Apr 2002 Posts: 2980 |
Since you have both a working and a failing setup for the same program, you can do a side-by-side comparison.
Get both invocations (from cartridge and from disk) in two VICE instances to the same point, that is, some loop where it waits for user input, same program counter, at least.
Then inject a keypress. I guess writing something into the keyboard buffer would do, but you'll see how it checks for input anyways.*
Then step through both instances simultaneously until they diverge, at which point you might have a vague idea what the problem might be. Watch out for stray or spurious interrupts, too.
Analyse memory and other state, redo for more hints, repeat until success. =)
* Finding out how to inject a keypress with the program performing the desired action can be done beforehand with the working setup. |
| |
Count Zero
Registered: Jan 2003 Posts: 1932 |
Why use an AR bloated version of that compiler at all? |
| |
zzarko
Registered: Feb 2003 Posts: 77 |
One reason is I haven't found cracked version of 1.03. I have tried to make executable from disk (got it from http://commodore.software/index.php/downloads/download/53-basic..), but I wasn't successfull (I obviously lack knowledge), so a friend made an AR snapshot that works from floppy.
Second reason is that it was in my set of testing programs and it failed. Maybe some other program could fail to work from cartridge for the same reason(s).
But, good thing is that 145 out od 146 tested programs now work :) |
| |
Count Zero
Registered: Jan 2003 Posts: 1932 |
Cracked version?
You mean a single .prg stripped from that auto start and the text or so?
Couldn't find some kind of real protection. |
| |
Hoogo
Registered: Jun 2002 Posts: 105 |
You could prefer the "practical" solution and replace the problematic file with a nicer version.
Or you could be happy that you've found such a problematic example and really understand and fix the problem.
Not sure what I would do in this case... |
| |
zzarko
Registered: Feb 2003 Posts: 77 |
Quote: Cracked version?
You mean a single .prg stripped from that auto start and the text or so?
Couldn't find some kind of real protection.
The image in 192K, and after starting it loads (or tries to load) a bunch of blocks from tracks 36-39 (if I remember correctly). I tried to track down what is being loaded where, but, as i said, couldn't figure it out. |
| |
map
Registered: Feb 2002 Posts: 27 |
As it was mentioned before, the protection seems not to be very hard, so I spend some time yesterday evening to create a onefile version of this thing. If you like I can send it to you. Maybe it helps.. |
| |
zzarko
Registered: Feb 2003 Posts: 77 |
Yes please! (I've sent you a PM) |
| |
Oswald
Registered: Apr 2002 Posts: 5094 |
if it only has input problems maybe its cia initialization problem ? (reading keyboard) |
| |
zzarko
Registered: Feb 2003 Posts: 77 |
It is not just an input problem, apart from showing the menu, almost nothing works as it should. Something in initialization is missing, but I do not know what. I tried running cartridge and disk version side by side in VICE, they both enter some kind of loop, but as it is a compiled basic, I was unable to determine what is going on.
And I found one more program that works from floppy, but not from cartridge.
Marko and I have decided to release the current code, as it works in most cases:
Magic Desk Cartridge Generator V3.0
Source code is available, and if someone who knows more than we can help, please do.
I'd like to thank all of you for your advices and help, code is better than it was thanks to you. |
Previous - 1 | 2 | 3 - Next |