Bad Mood : Falcon030 'Doom'

[leech]

Yes, I attached them as the current manual (version 5) on the site shows. http://hencox.gildaelectronics.se/homep ... manual.pdf

I was able to play some other games, like Worms and Alice's Mom's Rescue without graphical issues.

[leech]

Hmm, did anyone else notice the unstable the HUD is on CT60? It's literally flashing even though it's perfect in 030 mode (even the same binary - ctbmct6.ttp).

I also had this issue (see above posting) did you also have the VIDEL clock sync set up? That is if you have the SuperVidel connected.

[dml]

By the way, I sort of want to do a comparison between the music output by the Falcon and various other systems, I'm pretty sure it has the best Doom music I've ever heard.

It was one of the longer parts of the whole project... getting the sound right.

BTW music cheats...

type 'idmusXX' where XX=number from 11 to 64 (or some fairly high number). this lets you change the music at will.

If I read that correctly, there is even a way to output the music to a MIDI keyboard? (I have one laying around, I need to hook it up to test some of the other games that support it too! I think I did it for King's Quest at one point and the music was phenomenal)

Yes alongside the F030 software Codec device there's also a hardware MIDI device - the latter being handled by Saulot's library. It will output MIDI direct to a keyboard or synth. You can change this via the config file. The MIDI device uses less CPU too.

[Eero Tamminen]

The 060 build hasn't been properly tested so it is flaky - I can't really do much with it until I set up my own and I don't have time for it yet. I will get around to it at some point though. I can probably rebuild the 060 version to fix the flickery HUD but the crashes won't be resolved without debugging it locally here. So for now 060 support is just W.I.P.

Hatari's Mercurial version has had quite noticeable 040 & 060 improvements recently, as Nicolas has been synching Toni's WinUAE CPU core work. It should now have data cache emulation with optional MMU support also for 040 & 060, not just 030. And 030 mode is now supposed to support both cache & MMU emulation at the same time (earlier it emulated either of them not both at the same time).

I haven't yet tried that, but I'm planning on checking it with BM soonish.

[leech]

Would be cool to get SuperVidel support in Hatari. Not sure how possible that is (I assume it'd require some of the code that went into SV, and I'm not sure how they'd feel about that or not).

In the 8bit world, VBXE emulation was implemented into the Altirra emulator. It seems to support all the awesome hardware upgrades out there.

I'll try to make a video tonight of both the MIDI keyboard music and the weird HUD issue.

[Atari030]

Firstly, I am downloading everything on a PC, unzipping it there and transferring to the Falcon via USB Parcp. I have ensured I have DOOMU.WAD and it's the correct filesize (12,408,292) when checking in Windows. I am using Bad Mood v0.37.

Do you also have the other files present? BMT1.WAD and the three shpp* files?

If it can't locate PNAMES its having trouble reading one of the WADs. This would normally mean a corrupt WAD. However if you transferred it with PARCP (and CRCs enabled!) that would leave the disk device or driver software in question :-/ or the way the disk is formatted perhaps.

I then downloaded the BMC zipfile in this thread, hoping that removing the need to set the WAD file up would resolve the issue, but it has not. It still crashes at the same point, with the same error.

It does sound to me like there is a problem reading or writing large files.

My NVRAM needs changing and it's running in 60hz, not that I would imagine that matters.

I have the same problem with my Falcon and it doesn't get in the way.

It might be good to locate a hashing/checksumming program (e.g. MD5) which works at both ends to compare results. This would tell you if the WAD files are intact after transfer/writing, without transferring them back again.

It seems I have the same problem on my 4.01/14MB/8GCF Falcon. Not just badmood but PM Doom too. It does look like the WAD's are the root of the problem. My CT2 Falcon works fine with the same files (albeit off a 20G 3.5" IDE HDD). I tried a few variations on disk drivers but no joy. It might be time to upgrade from HDDriver 7, I think. I won't give up just yet and I might try another CF card with different partitioning.

[Fletch]

I am having issues getting midi music working on my MT32. I have set the mid_replay_mode to 2. midi_device_type to 0 (also tried 1). midi_channel to 2 (the MT-32 default).

During the boot up sequence the MT-32 flashes but no audio is played.

Any suggestions?

-Pete

[dml]

Not sure if anyone has tried HW MIDI since I built v0.37 so I'll try to test it next chance I get with the Falcon set up - in case something is wrong there. It did work ok in v0.36.

Meanwhile if anyone else has success with it I'd like to know.

[Fletch]

Not sure if anyone has tried HW MIDI since I built v0.37 so I'll try to test it next chance I get with the Falcon set up - in case something is wrong there. It did work ok in v0.36.

Meanwhile if anyone else has success with it I'd like to know.

Thanks Doug. I am attending an Atari Party this weekend and was trying to get that working for the show. No big deal, the built in playback will do just fine.

Awesome work on this.

-Pete

[dml]

I had a look and the MIDI support does seem to be broken. Can't get it to work at all on my Falcon. It also causes my copy of Hatari to crash. Will probably need to go back a bunch of revisions to see where it last worked and figure out what happened. Might just be breakage from gcc6 compiler upgrade.

[Eero Tamminen]

I tested the latest version from here with latest Hatari:
http://www.leonik.net/dml/sec_bm.py

Some observations:
* Normal Falcon version (bm030 & bm030s) starts now with Mercurial version of Hatari when MMU is enabled, but screen isn't updated properly
* Strict DSP sync version (bm030a), gets into problems earlier, nearly immediately after sound starts
* bmct6.ttp: 68060 unimplemented opcode 4C43, PC=000834f4 -> seems to be real BM bug, DIVS doesn't have HW support in 060?
* bmct2*.ttp: crash very early on BM startup
* bmhat.ttp works fine, but statusbar doesn't show, even with natfeats enabled, either with v0.37 or v0.36. I need to go back to ~Jan 2016 version to get Statusbar
* Otherwise still as amazing as I remembered

It would be nice to get statusbar working again with Hatari. I guess the screen update issue with MMU are Hatari problem, real Falcon doesn't have problems like that?

I took a look at the READMEs, and noticed few omissions:
* Hatari as one of the tested systems
* Free/mouselook feature that wasn't in original Doom
* Changelog doesn't list what was added in v0.37

[mikro]

bmct6.ttp: 68060 unimplemented opcode 4C43, PC=000834f4 -> seems to be real BM bug, DIVS doesn't have HW support in 060?

Although CT60 TOS contains an exception handler for unimplemented instructions, this shouldn't be present in an 060 build, agreed. -m68020-60 doesn't generate unimplemented integer instructions like DIV{S,U}.L <ea>,Dr:Dq so perhaps a wrong compilation flag was used.

[dml]

Although CT60 TOS contains an exception handler for unimplemented instructions, this shouldn't be present in an 060 build, agreed. -m68020-60 doesn't generate unimplemented integer instructions like DIV{S,U}.L <ea>,Dr:Dq so perhaps a wrong compilation flag was used.

A big part of the project is assembly language - it was written for 030 (with long muls and divs), so the 060 has to emulate it. There isn't really an 060-specific build, just a version which is made more compatible with it (mainly dsp timing, mmu etc).

[dml]

I tested the latest version from here with latest Hatari:
http://www.leonik.net/dml/sec_bm.py

Thanks Eero - I'll look at the hud thing next time I build it. Something funny with the new ELF/GCC makefile I think.

BTW I wanted to ask you about Hatari symbol loading direct from the Atari binary - as it has trouble with the new binaries - I can't use the NM route anymore because the symbols direct from ELF have different offsets from those in the final TTP. So it would be good to figure out if its a problem with our symbol table generation or a problem with Hatari symbol reading. You might be able to tell whats going on there.

[diamondspringwater]

Is it possible to link 2 falcons and do a co-op game of doom?

[leech]

IPX for the win?

[diamondspringwater]

How about through the midi ports like midimaze, using something like powernet.

[dml]

There's no network support in BM currently. There could be, but no work was done on it. Also it would probably only work Falcon-to-Falcon and some of the patches would need to be bypassed or undone to keep the game worlds in sync.

Midi port could probably work though as a sort of IPX hack.

[Ragstaff]

How about a bot to play against? :p

[Atari030]

NetusBee or Ethernec network support would be pretty cool too. Not that you haven't done enough already, Doug.

[dml]

Quick update - we figured out what went wrong with the MIDI replay and the ongoing problems with BMC update pass getting stuck on some (missing!) textures. Those things are all fixed now so v0.38 zips will be along soon.

[Ragstaff]

Yes, I hope my "tongue out" emoticon wasn't missed. BadMood is, as it is, far more than Falcon users could have wished for a few short years ago

[Trixster]

BTW music cheats...

type 'idmusXX' where XX=number from 11 to 64 (or some fairly high number). this lets you change the music at will.

I only just noticed this! Where do you type this in exactly?

[Eero Tamminen]

BTW I wanted to ask you about Hatari symbol loading direct from the Atari binary - as it has trouble with the new binaries - I can't use the NM route anymore because the symbols direct from ELF have different offsets from those in the final TTP. So it would be good to figure out if its a problem with our symbol table generation or a problem with Hatari symbol reading. You might be able to tell whats going on there.

Hatari supports only DRI/GST binary symbol tables. You can test them by using Hatari's external symbol handling program, it works exactly like Hatari debugger "symbols" command.

Here's the symbol table reading:
https://hg.tuxfamily.org/mercurialroot/ ... cii.c#l205

And here code for checking the binary sections:
https://hg.tuxfamily.org/mercurialroot/ ... cii.c#l411

For large programs (e.g. one's with code ported from other OSes), NM route is superior to using DRI/GST symbols from program itself because:
* latter has limits on symbol length
* DRI/GST symbol tables normally lack a lot of symbols that are in NM output (static function/variables?)
* with external symbols file, unwanted symbols can be filtered out (for profiler)

Btw. Is CT60 TOS available somewhere, or some program that would add missing 060 opcodes emulation on top of TOS v4? I'd like to test BM "060" version with Hatari 060 emulation.

[Eero Tamminen]

And of course I'm waiting eagerly for v0.38 with the MIDI & Hatari statusbar fixes!