|
Notes |
|
|
|
|
Can you please try attached slave and tell me if it fixes the problem, thanks. |
|
|
|
|
|
I have updated the patch and fixed a few problems, if it still doesn't work properly on your system, reopen this ticket. |
|
|
|
|
Thanks for your quick replies @StingRay. (And sorry I could not check your fixes yet, but I have been away from my Amiga.)
I'll let you know how this worked and reopen the ticket if necessary. |
|
|
|
|
|
Tried the updated slave, unfortunately though it does not make a difference. The demo still immediately blanks the video output once started (loss of signal). No audio is heard, and the A1200 only responds to a reset. |
|
|
|
|
|
I have attached a new slave, please try this one and tell me what happens (you should see a red screen, if so, press left mouse button, WHDLoad should then quit with "Illegal instruction at $3a3f6"). |
|
|
|
|
Tested the updated slave. It behaves as you correctly describe above: starts with a red screen, then upon pressing LMB WHDLoad quits with an Illegal Instruction ($10) at the same address ($3a3f6) you mentioned.
btw – I tried to get it to produce a coredump, it starts writing to the hdd and then the same "illegal instruction" pops back up again with no corecump created (that is, unless I am looking in the wrong place as I am assuming it would be in cwd) |
|
|
|
(0008170)
|
|
StingRay
|
2020-04-15 18:10
(Last edited: 2020-04-15 18:11) |
|
Cored dumps are saved in C: but the core dump is not needed. I made the slave crash on purpose as I need to find the part of the code which causes the problem on your machine (and I bet it's the ACA's fault, wouldn't be the first time).
I have attached 4 new slaves, please try them and tell me if the problem still occurs with all of them, thanks.
|
|
|
|
|
Just tried the first slave – it shows a red screen and continues with the demo upon pressing LMB. Going to let it run a bit to see if any other display issues occur.
btw - I hear you wrt. to the ACAs, this is my second one as the first one actually had a cpu-fault which was diagnosed by me and by AmigaStore (after quite a bit of haggling w/ Jens who claimed everything else as the culprit – from PSUs, FPU-libs and cellphone interference) ;)
I came a bit late to the A1200 game, as I obviously would have preferred a Blizzard card but not at the extortionate prices they are at now. (And I would suspect the Vampire cards would introduce a whole lot of incompatility issues on their own..)
Anyway, back on topic. The demo seems to run ok with the "No136" slave (red screen in between every section, so to mousepresses needed to advance). I'll try the others as well to compare. |
|
|
|
|
|
Hmm. The 136 slave has slight gfx-glitches (vertical bars in both scrollers) at the end w/ the Madness soundtrack (with the three skulls) |
|
|
|
|
Ok, preliminary summary: ;)
- 136 runs but has gfx-issues in the Madness/3x skulls section.
- 13a shows the problematic section correctly (no vertical bars in scrollers) w/ no other issues I could detect.
- 13e does not have any sound during the first intro part(!) (w/ the Alcatraz logo). No gfx-issues with the Madness segment here though, it runs perfect.
- 142 appears to be identical to 13a wrt. the way it behaves. All sections run properly w/ no vertical bar or sound issues.
All show red screens between the demo segments, needing two LMB presses. (But that probably comes as no surprise.) |
|
|
|
|
Very strange as I expected at least 1 of the "No1xx" slaves to trigger the problem on your config (I have just disabled all interrupt routines one by one in part 1). Somehow this doesn't make it any easier.
I have attached a new slave, what happens when you try the "NoVBI" slave?
|
|
|
|
|
The vertical bars issue appears to be intermittent, as it now showed on both 13a & 142 slaves when I was comparing the two.
Did a soft reboot and tested the sections again with 13a, 142 & 136 and it works properly. Seems WHDLoad does not completely clear the memory after each run.
EDIT: just saw you replied right before I was going to submit this. I will try the NoVBI slave as well. |
|
|
|
|
|
The NoVBI slave appears to work fine. It did show vertical bar issues in the skulls section, but I haven't tried rebooting the A1200 yet to make sure it is a "clean" test. |
|
|
|
|
Did a soft-reboot, but the glitches in the skulls section persisted. Launched the demo with the 142 slave (w/o rebooting) and it runs it ok.
Could be that this is triggered due to some other part of the demo running for a longer amount of time, as I obviously tested the NoVBI slave for longer. |
|
|
|
|
Confirmed – when skipping through the parts directly to the "skulls-section" both text scrollers now look ok with the NoVBI slave.
Ah, the weirdness. |
|
|
|
|
Left the problematic section running after I wrote the last comment, and it still runs perfect.
Wish I had not recently sold my Typhoon 030 card w/ SCSI & FPU, as it would have been quite interesting to compare its functionality with the ACA on this 1d board! |
|
|
|
|
|
OK, so it definitely is interrupt related. Try one last slave, does this one work OK on your machine? |
|
|
|
|
|
The UserMode slave works properly; even with the two scrollers in the Madness/skulls section (although that might be random). However, I left some sections running a bit longer and didn't reboot the Amiga in between testing this slave and the NoVBI one. |
|
|
|
|
|
OK, switching to user mode did the trick then, this will then be version 1.03 of the patch which will be available on the WHDLoad site shortly. Thanks for testing and fast replies. |
|
|
|
|
Thanks the same and for your blazing fast replies. I was a bit worried at first that this was some weird hw-defect with the old 1d board, since the way it crashed with the original slave - killing the video signal (& WHDLoad) - was not very reassuring.
At least now this true classic of a demo can be properly enjoyed; even on ACA boards. ;) |
|
|
|
|
Luckily all is fine now and you can enjoy this classic on your machine too. :)
The updated patch is available on the WHDLoad site so I'll close this ticket. |
|
Megademo2.slave (2,008) 2020-04-15 08:37