Unicorns season: Prince of Persia for the A8!

[Nickolasgaspar]

1 minute ago, rensoup said:

yes

There's not much I can do to debug this... I'd be curious to know if there are any read retries...

 

Perhaps it crashes always after x attract loops ?

 

No worries. Maybe changing  the storage medium might solve the issue. But I think that  there aren't any read retries, at least they are not registered by RespeQt. The one thing I noticed (new to me) while loading I can see RespeQt reading sectors but without the loading sound(it stops after a while)

[_The Doctor__]

you might try making sure you are running the latest respeqt and turning the sio speed down or double checking the serial port setting in reference to flow control and falling edge etc.... from time to time people mention such issues and it sometimes turns out they have the device set at a higher divisor than certain software can handle or a control method that has issues with speeds and synchronization with the edge of the signal falling at the wrong moment.

[Nickolasgaspar]

4 hours ago, _The Doctor__ said:

you might try making sure you are running the latest respeqt and turning the sio speed down or double checking the serial port setting in reference to flow control and falling edge etc.... from time to time people mention such issues and it sometimes turns out they have the device set at a higher divisor than certain software can handle or a control method that has issues with speeds and synchronization with the edge of the signal falling at the wrong moment.

I appreciate  for your suggestions!

It turned out I had the r4 release of RespeQt...which happens to have the best performance with this file! I slowed down the sio speed ( I only see differences in XEX loading speed). I tried RespeQt r5.2 and 5.3 (I think) but  unfortunately the loading process doesn't even reach the first Screen with those two releases!  I even made  a second sio2pc cable with a new serial board. I only saw an improvement  in the loading sound of the game. Now I can hear a really weak loading sound that the previous cable didn't pick up for a weird reason! But again no success.

I talked to Jurgen (tfhh)and he suggested to send him any files that  displays loading/run  issues so he can run some tests on his setup. He sounds  really interested in improving his products at every opportunity!

 

The  thing is that I have managed to find a way to run every  problematic game/demo/prg by  doing a different kind of trickery for each one, so I am starting to think that the Freddie chip must  play a role here. Does any one else has a Freddie  PAL machine with Jurgen's memory expansion?

 

 

 

[Mazzspeed]

1 hour ago, Nickolasgaspar said:

I appreciate  for your suggestions!

It turned out I had the r4 release of RespeQt...which happens to have the best performance with this file! I slowed down the sio speed ( I only see differences in XEX loading speed). I tried RespeQt r5.2 and 5.3 (I think) but  unfortunately the loading process doesn't even reach the first Screen with those two releases!  I even made  a second sio2pc cable with a new serial board. I only saw an improvement  in the loading sound of the game. Now I can hear a really weak loading sound that the previous cable didn't pick up for a weird reason! But again no success.

I talked to Jurgen (tfhh)and he suggested to send him any files that  displays loading/run  issues so he can run some tests on his setup. He sounds  really interested in improving his products at every opportunity!

 

The  thing is that I have managed to find a way to run every  problematic game/demo/prg by  doing a different kind of trickery for each one, so I am starting to think that the Freddie chip must  play a role here. Does any one else has a Freddie  PAL machine with Jurgen's memory expansion?

 

 

 

SIO has to be at bog stock speeds, there can be no HSIO at all. I can load the DD version of the game faultlessly using RespeQt at standard SIO speeds on a 600XL equipped with SIDE3, U1MB (both running the latest firmware), 64k upgrade, Sophia 2 and a home made SIO2PC adapter. RespeQt running on a Pi400 @ 2.4 GHz under TwisterOS.

 

Even when running bog stock SIO speeds, the lightweight DOS replacement used appears overly finiky regarding timing issues on certain configurations.

[+DrVenkman]

17 minutes ago, Mazzspeed said:

Even when running bog stock SIO speeds, the lightweight DOS replacement used appears overly finiky regarding timing issues on certain configurations.

This.

 

And with all due respect, this is why - while emulators are undoubtedly indispensable for most modern development - testing on real hardware absolutely should be part of the process. Real hardware is rarely as precise or predictable as an emulator.

[Nickolasgaspar]

1 minute ago, Mazzspeed said:

SIO has to be at bog stock speeds, there can be no HSIO at all. I can load the DD version of the game faultlessly using RespeQt at standard SIO speeds on a 600XL equipped with SIDE3, U1MB (both running the latest firmware), 64k upgrade, Sophia 2 and a home made SIO2PC adapter. RespeQt running on a Pi400 @ 2.4 GHz under TwisterOS.

 

Even when running bog stock SIO speeds, the lightweight DOS replacement used appears overly finiky regarding timing issues on certain configurations.

So after many efforts and tests I think its a matter of cable"noise".
Even if both home made SIO2PC cables perform excellent with all other files, the problems with this specific file start from this sector:

 

[Disk 1] Read Sector 646 ($286) #16 in track 35 ($23)
[Disk 1] Get status: $ff
[Disk 1] Speed poll

 

This is where the loading sound  drops and failures to read specific sectors begin.(multiple attempt [x2,3,4] )

For some reason I have some random success  by enabling non-standard speeds and setting the high speed mode POKEY divisor to 0.(1 success every 10-20 loading attempts...hardly a success)

 

So the fact that the game can load and the errors occure in random convinced me that the memory module or my freddie machine has nothing to do with this.

 

I will start by  building  a more robust cable with a better serial adaptor  and see if that changes anything.

 

  

[rensoup]

6 hours ago, Nickolasgaspar said:

[Disk 1] Read Sector 646 ($286) #16 in track 35 ($23)

interesting because there's nothing there I think, it's basically the last used sector on the disk ?

 

Spoiler

directory of image 'PoP_DD_20210619_bugfix.atr'

  0 42 ATTRCOV2BIN 00004 00015
  1 42 ATTRCOVLBIN 00019 00008
  2 42 BG1D    DTA 00027 00013
  3 42 BG1D    TBL 00040 00002
  4 42 BG1P    DTA 00042 00013
  5 42 BG1P    TBL 00055 00002
  6 42 BG2D    DTA 00057 00007
  7 42 BG2D    TBL 00064 00001
  0 42 BG2P    DTA 00065 00009
  1 42 BG2P    TBL 00074 00001
  2 42 CH1     DTA 00075 00020
  3 42 CH1     PMG 00095 00009
  4 42 CH1     TBL 00104 00004
  5 42 CH2     DTA 00108 00019
  6 42 CH2     PMG 00127 00010
  7 42 CH3     DTA 00137 00014
  0 42 CH3     PMG 00151 00006
  1 42 CH4F    DTA 00157 00013
  2 42 CH4F    PMG 00170 00004
  3 42 CH4F    TBL 00174 00001
  4 42 CH4G    DTA 00175 00013
  5 42 CH4G    PMG 00188 00004
  6 42 CH4G    TBL 00192 00001
  7 42 CH4K    DTA 00193 00009
  0 42 CH4K    PMG 00202 00001
  1 42 CH4K    TBL 00203 00001
  2 42 CH4S    DTA 00204 00010
  3 42 CH4S    PMG 00214 00001
  4 42 CH4S    TBL 00215 00001
  5 42 CH4V    DTA 00216 00011
  6 42 CH4V    PMG 00227 00003
  7 42 CH4V    TBL 00230 00001
  0 42 CH5     DTA 00231 00014
  1 42 CH5     PMG 00245 00007
  2 42 CH6A    DTA 00252 00014
  3 42 CH6A    PMG 00266 00005
  4 42 CH6A    TBL 00271 00002
  5 42 CH6B    DTA 00273 00014
  6 42 CH6B    PMG 00287 00003
  7 42 CH6B    TBL 00290 00001
  0 42 CH7     DTA 00291 00002
  1 42 CH7     PMG 00293 00001
  2 42 CH7     TBL 00294 00001
  3 42 CIRCLE  APL 00295 00033
  4 42 CUTOVL  BIN 00328 00006
  5 42 FLYTEXT APL 00334 00027
  6 42 FONTOVL BIN 00370 00010
  7 42 FUJICHS APL 00380 00005
  0 42 LAUNCHEROBX 00385 00007
  1 42 LEVELS      00392 00041
  2 42 LZSSOVL1BIN 00433 00011
  3 42 LZSSOVL2BIN 00444 00013
  4 42 PMGOVL  BIN 00457 00004
  5 42 POPCORE BIN 00461 00100
  6 42 PROOM   RAW 00561 00015
  7 42 SFX1    BIN 00576 00004
  0 42 SFX2    BIN 00580 00005
  1 42 TITLE   PCO 00585 00009
  2 42 TITLES  PCO 00594 00014
  3 42 VSIN    APL 00608 00029
  4 42 XBIOS   COM 00637 00010
74 free sectors
 

 

 4 42 XBIOS  COM 00637 00010

 

so it tries to read just beyond XBIOS.COM  (starts at sector 637 and is 10 sectors long) ?

 

Can anyone else with a working respeQt setup confirm that ?

 

 

 

 

 

 

 

[_The Doctor__]

I don't see those retries at all using respeqt, but I am using a real rs232 port with respeqt, and APE with usb port only PC's... I use WIN XP WIN8.1 and WIN10 for a quick test of each and did not have the retry issue...

 

What is the exact topography of your set up Nickolasgaspar?

As for Aspeqt? wow I wouldn't use that. It's not quite up to the job, I never had good results using the original Aspeqt. Respeqt corrected most of it's issues but a few still linger. With that in mind, timing is everything for both, I do remember having to choose different usb port and polling rates on a laptop or two in order to keep it chugging along.

 

oh and another issue that cropped up over time required some FTDI to be configured not to sleep or conserve power on long file transfers for back ups of larger disks and partitions... @Kyle22... do you remember the tools we used and what we set to slay your FTDI beast when that was going on?

[rensoup]

30 minutes ago, _The Doctor__ said:

I don't see those retries at all using respeqt, but I am using a real rs232 port with respeqt, and APE with usb port only PC's... I use WIN XP WIN8.1 and WIN10 for a quick test of each and did not have the retry issue...

but do you see it trying to read sector 646 ?

[xxl]

In addition to being one of the best games, PoP also has an educational value - the player will learn how to use the computer and its peripherals. will find out that equipment that does not meet the standard is considered to be non-qualitative and that a self-respecting consumer of Atari content uses premium quality equipment only. you have an Atari, you belong to the elite.

 

 

 

 

 

Edited June 24, 2021 by xxl

[VinsCool]

PoP is really gorgeous.
Only had tried it on emulator but this is even better on the real thing!
NTSC Atari 800xl, S-Video, 256k RAM upgrade, AVGCART, works perfectly so far!

I really like the loading times from the cartridge, this makes the experience really fluid

 

 

[MEtalGuy66]

On 6/22/2021 at 4:08 PM, rensoup said:

Sure we were all kids once, no need to insult the man who created this masterpiece though...

Noone insulted anyone. 

[Mazzspeed]

I see no sector retries as bad as those shown by @Nickolasgaspar using RespeQt via USB. The speed polling is indicative of communication issues. I agree with your decision to go over your cable, possibly trying another FTDI board.

 

I'm also getting tired of a certain users comments, and still question the use of a certain lightweight DOS substitute when it's actually creating compatibility issues as opposed to resolving them.

Edited June 24, 2021 by Mazzspeed

[+DrVenkman]

If I don't end up having to work all weekend and get motivated enough, I'll connect one of my FujiNets via USB cable to my PC during game loading and while the demo is running. Connected this way, the FujiNet firmware flasher program also gives you the option to capture serial debug data straight off the SIO bus connection to the ESP32 inside the FujiNet.

 

Since the same "blue screen with mouse" lockup has now occurred with people using SIO2PC adapters as well as with my multiple FujiNet devices on NTSC machines, I think there's some kind of timing related bug in the game's serial I/O code. 

[rensoup]

2 hours ago, VinsCool said:

Unbelievably some of those crap Atari joysticks survived ?

[VinsCool]

Just now, rensoup said:

Unbelievably some of those crap Atari joysticks survived ?

That's one of the two that came with one of my 2600's!

It also still works mostly fine.

 

The other one is not in a very good condition, it's got missing plastic cover, and as far as I remember it didn't seem to work lol

 

[xrbrevin]

1 minute ago, rensoup said:

Unbelievably some of those crap Atari joysticks survived ?

ive seen many with teeth marks on the rubbery stick tho...

[xxl]

55 minutes ago, DrVenkman said:

I think

excellent code, checks the biological footprint and refrains from working in two people who find it lacking the correct parameters. checks what comments you wrote and returns status: fail.

[Level42]

10 hours ago, DrVenkman said:

This.

 

And with all due respect, this is why - while emulators are undoubtedly indispensable for most modern development - testing on real hardware absolutely should be part of the process. Real hardware is rarely as precise or predictable as an emulator.

Don’t turn things around my friend. If an emulator doesn’t emulate the hardware exactly like the hardware behaves, it’s the emulator which isn’t precise.

 

[Mazzspeed]

5 minutes ago, Level42 said:

Don’t turn things around my friend. If an emulator doesn’t emulate the hardware exactly like the hardware behaves, it’s the emulator which isn’t precise.

 

You can't emulate the production variances of real hardware, and this seems to be the issue here.

[Level42]

30 minutes ago, rensoup said:

Unbelievably some of those crap Atari joysticks survived ?

Funnily, I just received a Chinese repro of the 2600 stick (AliExpress for less than 10 euros) which I ordered just “for the original looks” and I was REALLY surprised by it. It looks great and feels great. It has a nice click, not like micro-switches though and I actually played quite a bit of PoP with it. The game really lends itself because it’s like a “click-move” game.

the button is a bit too stuff for my taste though, but since you hardly use it in PoP.....don’t think it’s going to be comfortable with fast shooters....

but all in all I am seriously impressed by that Chinese knock-off, it is WAYYYYY better than the original Atari ones that came with our VCS 4-switcher BITD....I also hope it will last longer....the originals broke done after only about 7 months....just after warranty expired....

 

.....it’s also cool because who would have expected back then you could still get these brand new in the year 2021 ?

Edited June 24, 2021 by Level42

[Level42]

2 minutes ago, Mazzspeed said:

You can't emulate the production variances of real hardware, and this seems to be the issue here.

Which is why emulators suck.

[flashjazzcat]

11 minutes ago, Level42 said:

If an emulator doesn’t emulate the hardware exactly like the hardware behaves, it’s the emulator which isn’t precise.

Whether the emulator is precise or not is not the point. The point is that it if is assumed to be precise and the importance of real hardware testing is not appreciated, you end up... well, here.

Edited June 24, 2021 by flashjazzcat

[Mazzspeed]

1 minute ago, Level42 said:

Which is why emulators suck.

They're great as a development tool, but they need to be backed up by testing on a wide range of real hardware under varying configurations.

[+Stephen]

14 minutes ago, Level42 said:

Don’t turn things around my friend. If an emulator doesn’t emulate the hardware exactly like the hardware behaves, it’s the emulator which isn’t precise.

 

Which is why, knowing that all emulators are reproductions and therefore not original, it was said to test on the target (real) hardware.  We don't have to argue bullshit semantics do we?