Chuky, can you not use a driver cleaner and clear out your drivers.?

005 can log something, now after loading demo i hear start of music, and it hangs, i can't spot if its resolution change or not.
it hangs only music plays.

RaVeN-05 wrote on 2023-10-18, 19:08:

005 can log something, now after loading demo i hear start of music, and it hangs, i can't spot if its resolution change or not.
it hangs only music plays.

Are your loading screens fullscreen without a randomly coloured mirrored L section on the right ? The resolution switch occurs at the beginning. If they're fullscreen then it's switched correctly.

I'll check out the logs tomorrow.

BTW I have issues running rrosetta.exe with regular rready. Unless I set it to run in win98 mode rredline is randomly unavailable. The source does have a comment about rredline issues on winnt. Maybe I need to modify the binary.

without a randomly coloured mirrored "yes"

RaVeN-05 wrote on 2023-10-18, 19:28:

without a randomly coloured mirrored "yes"

Then the resolution switch is working for you.

This will still crash, but the logs should have less junk (V_QueryCmdBuffer) and should be easier to trawl through.

https://1drv.ms/u/s!As-dKk-N73dSlj5cAgcyGi4N_V5l?e=4j9AQT
MD5: 69cb7e3dca72a24cd9716ee6e076496d

Source:
https://1drv.ms/u/s!As-dKk-N73dSlj5cAgcyGi4N_V5l?e=NhjsCo
MD5: cfc626aadc18b92bca7d1bd1ac04d234

last 3 entries

05:52:39.110 V_AddToCmdList_

05:52:39.160 Massaged Callback (Entry)

05:52:39.160 V_IssueCmdBufferAsync_

RaVeN-05 wrote on 2023-10-18, 20:57:

last 3 entries […]
Show full quote

last 3 entries

05:52:39.110 V_AddToCmdList_

05:52:39.160 Massaged Callback (Entry)

05:52:39.160 V_IssueCmdBufferAsync_

Thought this was the problem, but couldn't be sure. When a command buffer overflows it calls a user supplied function which issues the commands to the Verité and returns a different empty buffer, which is where the V_AddToCmdList adds its commands.

I had issues with this with rready as well, but resolved it. I think using those same fixes in v2k. Rready hasn't quite worked.

No more builds for tonight will fix it tomorrow.

About music, in Windows 98 I connected the sound card to the cd drive and now music works in the d3d version, but the rendition version still crashes without music for me in 1024*768 (while Raven said he had music when he crashed). This is with version 18-10-003.corrected.

Then I formated again and I installed Windows ME so usb works without hassle now.

- In Windows ME everything seem to be the same as before in win98. In 1024*768 the game crashes after loading and I don't hear music. I still can't read the error message behind the game.
- With versions 18-10-003.corrected, 18-10-004, 18-10-005 and 19-10-001, if the desktop is in 640*480 then the loading screen is fine, I can navigate in the menus with music, when I start a game I can see the starting grid screen with the 3d checkered flag behind, but when I press Enter to start the race, the game freezes/crashes/hangs on a black screen. The computer is not responsive, I can't press anything and I have to switch the computer off. If the desktop is in another resolution like 1024, it's the same as before, the game doesn't change resolution, it crashes and I can't read the message.
-18-10-004 doesn't output any log in 640*480, but outputs logs in 1024*768.
-18-10-005 and 19-10-001 output logs in 640 and 1024. In 640, I'm not sure if logs keep growing when the game hangs with a black screen and I have to switch off the computer. One time the log file was 1 Mb, the other time it was 2 Mb. I included both logs.

Some notes
- I restarted the computer every time before a new test.
- Menus look pretty slow compared to d3d, especially the fade to black when you start a quick game.
- Maybe the checkered flag isn't actually in 3d and just fakes the 3d. In Tomb Raider 2, the Playstation version has a red flag when you open the menu while playing, just like this but it's actually 2d. There's a shader that fakes the 3d. On pc it's flat.
- I don't know what demo Raven is talking about.

There are 2 patches for Formula 1 that replace the exes. The first one says it's for Cyrix cpus, the 2nd one says it fixes joysticks. The 1st one has all 3 versions (3dfx, d3d and rendition), but I only found the 2nd one for the 3dfx version. The readme file says rendition and d3d versions will come if this patch is fine. I don't know if they ever released it for Rendition and d3d.
The Cyrix patch is here https://archive.org/download/pelitcd2/PCDA.IS … T%2FF1CREND.EXE The 3dfx one is in a different file.

The Cyrix patch seems to behave the same as the retail version with your wrapper, with the desktop resolution in 1024 and 640. In 640, my logs were 3 Mb but it probably doesn't matter. I included logs with the patch in case there's something.

Here's a video in 640*480 of the unpatched game with the version 19-10-001 of the wrapper
https://youtu.be/y5hKVWzEG7E

I'm attaching logs for each resolution for the 4 versions of the wrapper, with 2 different logs in 640 with 19-10-001, and logs with the Cyrix patch.

Chuky,

Your problem's exactly the same as Raven's:

02:40:34.420 Massaged Callback (Entry)
02:40:34.420 V_IssueCmdBufferAsync_
02:40:34.420 Massaged Callback (Exit success)

The callback isn't working correctly and corrupts bits of RAM. I think you have a different CPU/OS to Raven, which is why the callback comes back, possibly without doing its job properly. Execution continues after that with a corrupted stack, which is why you get different results.

The slow speed is because of the logging, which I made even slower yesterday to prevent empty log files. Once that's removed (once we get everything working), it should hopefully be faster than a V1000.

The game definitely switches resolution. It does that on Win10/11 (even Win7) with RReady and works for Raven on an actual V2k.

Could you try switching to whatever driver revision Raven's got?

Try this:

https://1drv.ms/u/s!As-dKk-N73dSlkHIAiE0JbRECypo?e=AbWkr0
MD5: 0392255fc589c4d07b2c7f23fa6d17e0

Source:
https://1drv.ms/u/s!As-dKk-N73dSlkKqVcMaTH8xDj5p?e=Qa4ZZI
MD5: 004c50edb69d58115d24ad8f2df79bd7

It should run for a while longer now. Not entirely sure about the polygon commands (yet).

This should also work on a V1000 with stock drivers in exactly the same way.

Raven can you check your desktop resolution just to be sure ? You said you couldn't spot a resolution change and that you didn't have random colors, but it could also indicate that your desktop resolution is 640x480. If it's 640x480, can you try a higher resolution ?

edit: Like I said, yesterday I tried various drivers and it was always the same issue, and I ended up formating again.

checked my desktop is 640x480

from desktop 1024x768 starting game leads to have corruptions and it crashes before main menu

RaVeN-05 wrote on 2023-10-19, 09:05:

checked my desktop is 640x480

from desktop 1024x768 starting game leads to have corruptions and it crashes before main menu

The plot thickens.

This should switch res:
https://1drv.ms/u/s!As-dKk-N73dSlkP2Ikd3vNf4u2qH?e=Kc9WcQ
MD5: ae38c7b89079673445c574f2fe902c98

I think the problem was with the bit depth (V_SetDisplayMode_). I wasn't reading it in from the ecx register. RReady wasn't affected because it just ignored it and assumed 32 bit. It's been fixed in V2k.RReady (passed to V_SetDisplayMode) and regular RReady (assigned to the variable, but still being ignored).

Source:
https://1drv.ms/u/s!As-dKk-N73dSlkRiEDnPleePnrgS?e=VVgi84
MD5: 0678edf8b55a0ac377cae3950b19a987

now it changes from desktop 1024x768 60hz to 640x480 75hz

RaVeN-05 wrote on 2023-10-19, 11:58:

now it changes from desktop 1024x768 60hz to 640x480 75hz

Delete your log files before starting. Should probably automate that.

Don't try this. I made a mistake.
Try:
https://1drv.ms/u/s!As-dKk-N73dSlkV2o8dFmbpcuRzC?e=1c7oND
MD5: ab4ba41418a730a9bd4bd10b161bb6d8

Source:
https://1drv.ms/u/s!As-dKk-N73dSlkbkkikhVxjm-fnv?e=jCohod
MD5: 6aac4b94ca975d93b2b31be9c98e666d

Different command buffers have different callbacks (I think). There're two lots of command buffers created at different times. This caches a per command buffer callback. The raw callback can't be passed to system dlls. They expect different behaviour from the callback. So the system DLLs get the Massaged callback which calls the native callback. Before there was a single global callback. So if a later command buffer had a different func all buffers got the new func.

This version might fail spectacularly because I'm using STL std::map.

switches resolution, i see main menu, waited for demo , see demo loading screen , if fade to black, music start playing, hangs in black screen with music, no game sounds. PC unresponsive, hard reset.

Try this instead:

Ignore these:
https://1drv.ms/u/s!As-dKk-N73dSlkiJLvyGsB80lJrf?e=PgMfR4
MD5: b407e49eead2546f2865b19369c0fbc5

Source:
https://1drv.ms/u/s!As-dKk-N73dSlkkoQd-3jsB5CsHz?e=u7OblT
MD5: 30c5414cc1b353adb6a891395c7a21ed

Try these instead (This is a slow-mo build. It'll be very very slow. But it'll log everything.):
https://1drv.ms/u/s!As-dKk-N73dSlkrwg_Z1V10scbss?e=bhezH9
MD5: 7c7e57e93cc771492d7509c1e36d4ee4

Source:
https://1drv.ms/u/s!As-dKk-N73dSlktCPaZt-oFix4Lh?e=VeacXt
MD5: 39d01c9f93b17acf770b5df93bdce6ae

Might be easier if I committed these to github or sourceforce..

I won't be able to provide any more builds for a couple of hours. Downtime.

everything the same , very slow, i know its ok)

RaVeN-05 wrote on 2023-10-19, 13:59:

everything the same , very slow, i know its ok)

I'm not at my pc. What're the last lines?