The Sinclair QL Forum

Just checked my QDOS Companion too. My notes for IO_OPEN (Page 39) are as follows:

• D2 may not be preserved on exit. See QL World, December 1986 Page 43.
• A1 is not preserved unless a TRAP #4 was done first!!!

It seems I had problems too, some time back.


Cheers,
Norm.

NormanDunbar wrote:Just checked my QDOS Companion too. My notes for IO_OPEN (Page 39) are as follows:

• D2 may not be preserved on exit. See QL World, December 1986 Page 43.
• A1 is not preserved unless a TRAP #4 was done first!!!

It seems I had problems too, some time back.


Cheers,
Norm.

There is a Trap#4, just before the Trap#2. And I get the problem.

Martin_Head wrote:

NormanDunbar wrote:Just checked my QDOS Companion too. My notes for IO_OPEN (Page 39) are as follows:

• D2 may not be preserved on exit. See QL World, December 1986 Page 43.
• A1 is not preserved unless a TRAP #4 was done first!!!

It seems I had problems too, some time back.


Cheers,
Norm.

There is a Trap#4, just before the Trap#2. And I get the problem.

That's Norman's docs that say "preserved after TRAP#4" - Other's don't. I guess that's more of an observation than a system commitment.

Here is updated MDI driver. I have done a quick test on the CONVERT2_MDI command. But it seems to work OK in Qemulator with JS and Minerva 1.98 ROMs.
It looks like I also did a fix for a block checksum problem in the FORMAT routine some time ago, But I don't remember doing it

If I use the ROM version of the driver with Minerva, The game displays the opening page. Then I see something resembling the QL 'tweed' pattern, and Qemulator complains with two 'Bad channel block address' errors.

I don't know if it's a problem with my driver, But I suspect it's a game compatibility issue with Minerva.

I have not tried to look into the problem about using the same drive as the source and destination when converting images on a SuperGoldCard. My QL's are in a cupboard, and my SuperGoldCard died some time ago.

tofro wrote: That's Norman's docs that say "preserved after TRAP#4" - Other's don't. I guess that's more of an observation than a system commitment.

From a JS disassembly I can determine that A1 isn't preserved at all. The only thing a TRAP#4 does is to set a bit in the job header, which is read by the subsequent TRAP#2. Unfortunately, the entry code of the latter trap uses A1 to locate and read this bit (for those interested, it starts at $32A2). Ironically, A1 is saved in the code afterwards which is of course of no use.

Of course, Minerva or SMSQ/E might have different behaviour...

Jan

Hi Martin,

I tried the updated version of MDI on a Tetroid Trump Card with Minerva v1.98, JS, JM roms, the CONVERT2_xxxx command now works perfectly on each of the roms.

I can not get the game to run due to the fact it need 128K of ram, the Trump Card is probably taking some for the disk controller, but using the RES_128 command I was able load the first screen, but next files loaded into the computer from floppy disk emulating a MDV with the FLP_USE command, crashed the computer with the microdrives runnng.

I will try other configurations to get the game running, maybe vDriveQL is the best bet.

Thank you for the trouble in finding the problem, I will update the Q68 distribution with MDI v1.09.

Way back when the thread was about Sam the Little Spaceman, there were some comments about the game not running from disk - giving out of memory or getting so far on Minerva and crashing.

Back in the depths of my memory, I don't think I ever got Sam the Little Spaceman to work from disk.

The reason?

The disk drives use up some memory for the disk toolkit and slave blocks. RES_128 on the QL does not give the QL the full 128K (before removing the 32K screen memory).

Using Minerva's CTRL F4 (or SHIFT F4) give different memory options and results so it is always worth trying.

Also remember to try Minerva in dual screen mode, if a program appears to run but you cannot see it on screen!

Hi Rich,

I started to disassemble the game, which I have only small amount of work on this.

But the game uses fixed System Variables addresses, so if the Second Screen is enabled, the System Variables are at $28000

The game also thinksvtge screen is at $20000, so will only work on systems that can fix the Screen and System Variables. Like QPC2, SMSQmulator, Qemulator, Q68...

so any progress???

i got it run from flp1_ (js rom + tk2 2.20 )
cst qdisk 1.16 128k ql
(in second load just type.... RUN)