Unicorns season: Prince of Persia for the A8!

[xxl]

1 hour ago, tmp said:

i give up

? I wonder if he is aware of his condition. but let's be honest the statistics are inexorable, there are such people in every environment

[Mazzspeed]

1 minute ago, xxl said:

? I wonder if he is aware of his condition. but let's be honest the statistics are inexorable, there are such people in every environment

This isn't even worthy of a reply. What an outright pathetic response.

 

[+DrVenkman]

The bottom line with these ATRs is that xxl’s file system is marginal at best. It works on his machine and a subset of other real hardware in the world. It does NOT work reliably with the current most popular SIO device in the world (FujiNet) even on hardware that is demonstrably more stable, with cleaner power and ground planes than any vintage Atari ever made and with a far sharper and cleaner clock (the 576NUC+). 

 

One can rightly wonder - as @DjayBee did last night - if FujiNet “properly” models the track stepping timing of a real drive, though I don’t know if anyone but @tschak909 or @mozzwald can rightly answer that. On the other hand, FujiNet seems to have zero issue loading nearly any other properly-constructed ATR, ATX, XEX or CAS file so should it be necessary for a device in use by thousands of hobbyists have to emulate xxl’s 130XE + disk drive combo to be considered “accurate?” Beats the heck out of me but I tend to doubt it. 

 

If this game can’t load from a SIDE Loader in use by thousands of hobbyists, and can’t run reliably from a FujiNet in use by thousands of hobbyists, maybe that ought to be addressed.

[xxl]

you have to understand that when a 4-year-old child who has no idea what the adults are talking about, interjects in the conversation, he / she gets petted on the head

[ggn]

Right, so I apologise if I break the current thread mood of doom and gloom and flamewars. That said...

 

I really want to congratulate @rensoup and everyone involved in this project - it is a labour of love and it really shows! I wish the team all the best and hope to see a final version soon!

 

The main purpose of my post is just to give my bit of feedback testing the 20210619 bugfix version on my various configurations here. Mainly:

• 800XL with Ultimate 1Mb and dual POKEY extensions installed
• Stock XF551
• Fujinet v1.3
• AVG cart

The combinations I tried are:

• Fujinet loading the DD version
• XF551 loading the DD version
• AVG cart loading all 3 cartridge images

I'm happy to report that everything passed with flying colours :). Reading this thread the past couple of weeks I was expecting a crashfest, but I was gladly proven wrong. So I'm extra happy.

 

That's all from me, again congratulations and keep up the good work!

 

(You may now return to flaming each other)

[+DrVenkman]

4 minutes ago, ggn said:

The combinations I tried are:

• Fujinet loading the DD version
•  

See if the game will run through its self-play/demo mode for more than an hour. 

[mozzwald]

26 minutes ago, DrVenkman said:

One can rightly wonder - as @DjayBee did last night - if FujiNet “properly” models the track stepping timing of a real drive, though I don’t know if anyone but @tschak909 or @mozzwald can rightly answer that

iirc,

@jamm ported the XEX code to Fujinet from SDrive or at least used it as a guide. I have no useful input on how it works but I can test it and see if I have the same issues. I'll also check if there's anymore debug info we can get from Fujinet.

[+DrVenkman]

2 minutes ago, mozzwald said:

iirc,

@jamm ported the XEX code to Fujinet from SDrive or at least used it as a guide. I have no useful input on how it works but I can test it and see if I have the same issues. I'll also check if there's anymore debug info we can get from Fujinet.

To be clear, the files involved here are the DD ATRs posted in the first post of the thread, as well as several test versions posted by @rensoup in the last couple days on previous pages. I can compile them for you and send you a PM later tonight after work if I get some time.  

And to further detail, the newer test demo versions DO load and will run apparently normally for ~30 - 60 minutes before eventually crashing, usually with an animated mouse graphic on a blue background. According to rensoup, this is his “file not found” indicator. Since the demo versions do apparently load and only crash after some time, each test cycle involves a lot of waiting … 

 

[+DjayBee]

57 minutes ago, DrVenkman said:

One can rightly wonder - as @DjayBee did last night - if FujiNet “properly” models the track stepping timing of a real drive

I do not wonder if, I am quite sure that it does not.

Just like all the other disk emulators out there (SDriveMAX, RespeQt, ...).

 

IMHO it would make no sense to do so because ATR ist just about delivering disk sectors.

Obviously this needs to happen with the correct SIO speed chosen by the loader in the Atari but not so for anything surrounding the plain SIO protocol.

 

ATR is not about exact timing. For these purposes we already have ATX.

AFAIK ATX currently does not support DD but somebody should be able to create either an ED image with the latest version or (if ED still has not enough free space) a set of two SD images. Tis could be used by @DrVenkman.

 

I still believe that there is some sort of race condition between more than one interrupt due to the interrupts rensoup's loader raises/causes.

Since @rensoup has reduced the number and/or the length of his interrupts it seems that it happens less often but it still does.

 

From my understanding this heavy usage of interrupts is not possible with the Atari's standard OS, therefore I think it's unfair to blame XBIOS for something it allows to be leveraged (which not at all can be done with other DOSes).

[mozzwald]

1 hour ago, Mazzspeed said:

However, as stated, hopefuly a developer can clear up as to whether the web GUI option in FujiNet-PC does in fact mean that COMMAND is implemented correctly to Atari specifications

Yes, Fujinet implements COMMAND according to Atari specs but always a possibility of a bug.

 

The web gui option to change is only available on fujinet-pc for the different USB to serial converters. True Fujinet hardware uses a gpio pin.

[Faicuai]

11 minutes ago, DjayBee said:

IMHO it would make no sense to do so because ATR ist just about delivering disk sectors.

Bam! Bingo!

 

Not only reading .ATRs, but also from SIO on actual disk-hardware that dynamically combines / interleaves magnetic-reads with sector-ram-buffer reads (e.g. Indus/GT with SDX's TrackBuffer-code enabled + RamCharger).

 

That is a very significant test, because sector-read times do change with such interleaving of media+bF reads... and it works like a charm with PoP DD-image (and built-in X-bios, for as long as I could test).

 

It is the SIO-chain emulation of affected solutions the one we need to zero-in, before adjusting supporting+application SW. That would be the rationally-honest thing to do.

[rensoup]

1 hour ago, DrVenkman said:

And to further detail, the newer test demo versions DO load and will run apparently normally for ~30 - 60 minutes before eventually crashing, usually with an animated mouse graphic on a blue background. According to rensoup, this is his “file not found” indicator. Since the demo versions do apparently load and only crash after some time, each test cycle involves a lot of waiting … 

Just to reiterate the "file not found error" occurs because it fails with 3 retries on reading a directory sector according to a log posted by @Level42. @xxl mentioned that he set the retry count to be 64.

 

1 hour ago, DjayBee said:

still believe that there is some sort of race condition between more than one interrupt due to the interrupts rensoup's loader raises/causes.

Since @rensoup has reduced the number and/or the length of his interrupts it seems that it happens less often but it still does.

interrupt usage is very low, it's just the VBi. which is now fast enough with the latest test (but possibly not with the _bugfixed version from the 1st page). XBIOS only uses polling, so there's no IRQ.

[rensoup]

this was posted by ebigguy 

 

Quote

RespeQt automatically switches back to 19200 after 2 retries. And then all the remaining sectors of the ATX is read at normal speed

 

 

I have no idea if that's relevant or not since this is for ATX 

 

 

[rensoup]

In other news @TIX is tweaking all the animations again ?, that deserves props ?

 

(almost feeling off topic!)

[Heaven/TQA]

Another 2 cents by me:

 

game runs on cart? Check

game runs on sio2SD? Check

game runs on real media? Check

game runs on xl and 800 and pal/ntsc? Check

on mono/dual pokey? Check

 

so wtf guys… why is there such a heat here? 
 

i as coder would not care if the game crashes after running 1 hour or longer (!!!) in demo mode… seriously? If I would care… as demo coder I would switch off the attract mode and made my life easier… same with those all exotic hardware expansions… yes I call blutooth etc exotic… 

[rensoup]

17 hours ago, DrVenkman said:

EDIT: Okay, I've loaded up the game via SDrive-MAX now and I'll see if it crashes.

btw what was the result of this ?

[+Stephen]

13 minutes ago, Heaven/TQA said:

Another 2 cents by me:

 

game runs on cart? Check

game runs on sio2SD? Check

game runs on real media? Check

game runs on xl and 800 and pal/ntsc? Check

on mono/dual pokey? Check

 

so wtf guys… why is there such a heat here? 
 

i as coder would not care if the game crashes after running 1 hour or longer (!!!) in demo mode… seriously? If I would care… as demo coder I would switch off the attract mode and made my life easier… same with those all exotic hardware expansions… yes I call blutooth etc exotic… 

This reminds me of a blog post from Jeff Minter way back in the Nuon days.  VM Labs made him fix a bug in Tempest 3000 where the game crashed if running in attract mode for something like 2 days.  I think my response would have involved such words as "game, ass, shove, up" - maybe not in that order.

[Yautja]

27 minutes ago, Heaven/TQA said:

Another 2 cents by me:

 

game runs on cart? Check

game runs on sio2SD? Check

game runs on real media? Check

game runs on xl and 800 and pal/ntsc? Check

on mono/dual pokey? Check

 

so wtf guys… why is there such a heat here? 
 

i as coder would not care if the game crashes after running 1 hour or longer (!!!) in demo mode… seriously? If I would care… as demo coder I would switch off the attract mode and made my life easier… same with those all exotic hardware expansions… yes I call blutooth etc exotic… 

So anyone has time to let a game running its attract mode for one hour?

 

[xxl]

if PoP stops working after one hour, it is a sign that it is time to replace the fan in the device because it has overheated.

Edited June 28, 2021 by xxl

[+DrVenkman]

14 minutes ago, Yautja said:

So anyone has time to let a game running its attract mode for one hour?

 

Sure thing, Mr. Wonka - absent bugs in the game, the file system or the loading device, the game SHOULD run indefinitely- days or years. Ever notice the burn-in on old CRT arcade monitors? That’s because the machine owners didn’t boot the game to make sure it loads, then turn it off. These machines and devices are deterministic. There’s no intentional randomness built into the hardware designs. Games should not crash with a file system error within 30 minutes of loading. 

 

 Now why don’t you go back to your chocolate factory?

[Heaven/TQA]

10 minutes ago, DrVenkman said:

Sure thing, Mr. Wonka - absent bugs in the game, the file system or the loading device, the game SHOULD run indefinitely- days or years. Ever notice the burn-in on old CRT arcade monitors? That’s because the machine owners didn’t boot the game to make sure it loads, then turn it off. These machines and devices are deterministic. There’s no intentional randomness built into the hardware designs. Games should not crash with a file system error within 30 minutes of loading. 

 

 Now why don’t you go back to your chocolate factory?

The question as manager like me would Be how many money and time goes into fixing this 0,01% relevant bug? Bug of whatever in the chain causing the bug?

Edited June 28, 2021 by Heaven/TQA

[Level42]

My 2 cents:

 

The developer (rensoup) seems to be willing and interested in getting PoP bug free and DrkVenkman has the hardware AND TIME to try out stuff and provide feedback. I see both communicating in a constructive way so we should keep out of this unless we can also provide valuable (test) information.

 

if rensoup would decide to call it a day on fixing the bugs (if any are his) then we’ll have to accept that and be happy with whatever we will get and adjust our setups accordingly.

 

By the way Rensoup, I didn’t provide any logs AFAIR, so the credit for that goes to someone else.

 

[Level42]

21 minutes ago, Heaven/TQA said:

The question as manager like me would Be how many money and time goes into fixing this 0,01% relevant bug? Bug of whatever in the chain causing the bug?

Are you a manager at Boeing maybe ?

Edited June 28, 2021 by Level42

[+DrVenkman]

16 minutes ago, Heaven/TQA said:

The question as manager like me would Be how many money and time goes into fixing this 0,01% relevant bug? Bug of whatever in the chain causing the bug?

This is a hobby programming project and I’m a hobbyist. My children are adults and grown. My wife has a real job and hobbies of her own. I have a real job that lets me work from home and allows me to have the toys I want within reason. All of the above means that if I want to set up a system and let it run a game demo as a stability test for 24 hours straight, I have no one to answer for but myself.

 

You’re welcome to productively contribute or not. That said, I have run a number of your demos for extended periods of time and they generally don’t crash. And I can run them from a CF card or SIDE3 as well. *shrug*

[+CharlieChaplin]

19 hours ago, rensoup said:

@CharlieChaplin didn't you mention it worked with your 1050 ? could you try the above version:

 

 

and let it loop for a while ?

 

 

Will do.

Besides, both SD and DD versions that did several retries worked on my PAL 800XL (under XL-OS Rev. 2) with 512k SRAM enhancement by tf_hh and 1050 with Speedy, very slow but they worked. The bugfixed SD and DD versions also worked on my 1050 with Speedy and much faster (approx. 4x faster than the version with lots of retries).

 

Tested the CAR versions with my AVG cart. and PAL 800XL (again under XL-OS rev. 2) with 512k SRAM enhancement by tf_hh (set to 512k, so there was always 576k RAM available). The older CAR versions all worked fine (Atarimax, Megacart and SIC!) and the newer bugfixed CAR versions also worked fine. I could always load the game fully and finish level 1. The only problem I ever had and still have is that I suck at this game and so it is always Game Over for me in level 2... ;-)  However, I did not test the demo infinitely and so I cannot tell if it crashes after a while.