Compileable version of Alien Breed 3D II - The Killing Grounds

[amipal]

That's weird. Maybe there was a fix in one of the libraries that TKG references?

[pipper]

TKG is using really rudimentary interfaces, nothing fancy one would believe.

[alexh]

Originally Posted by pipper

One more data point: using 3.1.4 ROMs in WinUAE makes the 'hires' executable start up instead of crashing. I have no idea why... TKG is not using any fancy 3.1.4 features...

Maybe unrelated but memory allocation was completely overhauled in v3.1.4

Originally Posted by Thomas Richter

In 3.1.4, the prime sources for memory segmentation were already removed, and those are ClipRects (layers V45 fixed that) and Regions and RegionRectangles (graphics V45 fixed that)

[abu_the_monkey]

I can confirm that it does indeed work with ks3.1.4 but stepping through with the winuae debugger it still shows many exceptions like

Exception 26, PC=46240A5E
Exception 8, PC=00F80C68
Exception 27, PC=4624F340

anyone know how to interoperate these?

edit: forget the above. Pipper has explained it in the winuae debug thread. or it seems to be the same thing I am seeing.

[Karlos]

I still think something may be wrong with the taglist. A tag value being interpreted as an address or something. Just a hunch.

[pipper]

Yes, I agree, but I have stared at this for hours. I cannot figure out what is supposed to be wrong here:

>m MainWindowTags 7
7000B088 8000 0064 0000 0000 8000 0065 0000 0000  ...d.......e....
7000B098 8000 0066 0000 0000 8000 0067 0000 0000  ...f.......g....
7000B0A8 8000 0070 702A 1100 8000 006B 0003 1940  ...pp*.....k...@
7000B0B8 8000 0089 0000 0001 8000 0088 0000 0001  ................
7000B0C8 8000 008A 0000 0001 8000 0087 0000 0001  ................
7000B0D8 8000 008C 0000 0001 8000 0085 0000 0001  ................
7000B0E8 0000 0000 0000 0000 0000 0000 0001 0001  ................
>m MainScreen 1
7000B008 702A 1100 0002 AFB0 0003 EFD8 0000 0000  p*..............
>

WA_CustomScreen is the only pointer in the window taglist and provides the pointer to the screen with the right tag item.

[Karlos]

Is it some property of the Screen that is passed in that it doesn't like then? That would be my next guess.

[Karlos]

I need to try this WinUAE debugger. I've not used it before.

[Karlos]

Regarding the fullscreen floor texture issue, I've noticed that it seems to be the case the floor texture shifts slowly to the right relative to the player's forwards vector as you move around. I'm intrigued.

[pipper]

I think what’s wrong there is that the perspective foreshortening for floor and walls mismatch, I.e the „FoV“ is not the same. Likely to me expanding the rendering from 288 to 320 pixels

[Karlos]

I'm going to try and look into it. Pardon the pun. When the kids eventually hit the hay.

[pipper]

Originally Posted by Karlos

I'm going to try and look into it. Pardon the pun. When the kids eventually hit the hay.

I filed https://github.com/mheyer32/alienbreed3d2/issues/9 and left some additional info there

[Karlos]

Originally Posted by pipper

I filed https://github.com/mheyer32/alienbreed3d2/issues/9 and left some additional info there

Good information. Looks like it will be tomorrow before I can take a look.


Off topic, I notice there are quite a few integer divisions, for example signed divide by 3. If memory serves, the 68060 can do integer multiplication much faster than it can division. This makes me think that some of these, depending on required precision, might be optimised, e.g:

divs.l #3, d6

could become

muls.l #341, d6
asr.l #10, d6

You can use larger values for the multiply and shift to better approximate 1/3

This is an example, but the same strategy can be applied for other divisors.

[abu_the_monkey]

these are not really a problem as they have been this way from the start, but, maybe could do with being named slightly differently as they do get flagged by vasm

warning 41 in line 6028 of "hires.s": label name conflicts with directive
>end:

warning 41 in line 425 of "cd32joy.s": label name conflicts with directive
included from line 6402 of "hires.s"
>Data:

warning 41 in line 1021 of "objectmove.s": label name conflicts with directive
included from line 6628 of "hires.s"
>DX: dc.w 0

warning 40 in line 756 of "hireswall.s": label name conflicts with mnemonic
included from line 6635 of "hires.s"
>store: ds.l 500

warning 41 in line 4185 of "objdrawhires.s": label name conflicts with directive
included from line 9593 of "hires.s"
>offset:

warning 41 in line 11425 of "hires.s": label name conflicts with directive
>even

warning 41 in line 12278 of "hires.s": label name conflicts with directive
>Echo: dc.w 0

line numbers are from the 'try_fix_crash' branch.

[Karlos]

It's probably an OCD thing of mine but the mixture of different cases for labels and offsets in the code grinds my gears.

[abu_the_monkey]

there seems to be at least 3 different menu systems in the source code one from ab3d1 and from a quick scan over it appears redundant and two for ab3d2 which seem to have some cross over (not helped by the similar naming due to their purpose) like 'mnu_MYMAINMENU' and 'mnu_mainmenu'

example: in 'menunb.s' in 'mnu_start:'

ln 48: mnu_loop: lea mnu_mainmenu,a0

but this is the wrong menu system which explains why as per the comment it loads the wrong menu if this is called in 'controlloops.s'
however there are others as when changed to point to the correct 'mnu_MYMAINMENU' it breaks the menu. that is it works and displays the correct menu but you can't select anything.

[pipper]

Yeah, the new menu system seems bolted on last minute; while the old menu system is still in there. In fact, if you compile “cheesey” in “cheesesauce”, the resulting executable will show up with the old AB3D1 menu system.

You seem to have fun digging around :-D

[abu_the_monkey]

I find it interesting to 'try' and understand how things work

[Karlos]

@pipper - YouTube just gave me this [ Show youtube player ]

I thought it was interesting because the aspect and perspective issues were really conspicuous, both in fullscreen and 2/3 size. However, it was posted a fortnight ago so I don't know which version it is. The effects are a lot less noticeable on my machine.

[Karlos]

@pipper - YouTube just gave me this [ Show youtube player ]

I thought it was interesting because the aspect and perspective issues were really conspicuous, both in fullscreen and 2/3 size. However, it was posted a fortnight ago so I don't know which version it is. The effects are a lot less noticeable on my machine.