Really Interesting Rendering Bug

[Onar4241]

Alright, I got to the bottom of the problem. It only happens via ACS. In DECORATE, it does not happen.

[edward850]

I had to close my computer for 3 days, that's why that happened.

"Closing computers" does not generate critical kernel faults.

[Edward-san]

... can you please provide a damn link to the wad? The one I have has skulltag-only actors, so it can't be opened with zdoom.

[PWAD Pete]

Definitely interesting to see how zdoom can go balls to the wall ballistic when it goes past its limits. I wonder what other things can happen when more boundaries are broken.

[Onar4241]

... can you please provide a damn link to the wad? The one I have has skulltag-only actors, so it can't be opened with zdoom.

https://sites.google.com/site/realgunsh ... /downloads

[Onar4241]

I had to close my computer for 3 days, that's why that happened.

"Closing computers" does not generate critical kernel faults.

I have a faulty-type of RAM and when I don't turn the memory down to zero (Shut down my computer) it is most likely to crash and restart.

[MaxED]

Code: Select all

#include "zcommon.acs"

script 1 ENTER
{
    while(true)
    {
        SetActorPitch(0, GetActorPitch(0) - 100);
        Delay(1);
    }
}

Code: Select all

#include "zcommon.acs"

script 1 ENTER
{
    while(true)
    {
        ChangeActorPitch(0, GetActorPitch(0) - 100, true);
        Delay(1);
    }
}

Both of these will do the trick, and in both directions (up and down)...

[Edward-san]

I know what's going on, but I don't know a 'proper fix' yet.

[edward850]

turn the memory down to zero

Please tell me you don't work with computers.

I know what's going on, but I don't know a 'proper fix' yet.

You can say that again. A_SetPitch is clearly supposed to be clamping the view pitch, and yet if this report is to be believed, that apparently isn't happening.

[Edward-san]

You can say that again. A_SetPitch is clearly supposed to be clamping the view pitch, and yet if this report is to be believed, that apparently isn't happening.

You're looking at the wrong place. See this fix I managed to find. The issue is that SetActorPitch in p_acs.cpp just calls AActor::SetPitch without taking into account the player's limits. In the end, I went for unifying the player's handling code from thingdef_codeptr.cpp inside AActor::SetPitch so that it could work for both decorate and acs.

[edit] BTW: I believe that MinPitch and MaxPitch can part of userinfo code.

[edward850]

That would make more sense. I thought we were talking about DECORATE, which is why people need to remember to make intelligible bug reports.

[edit] BTW: I believe that MinPitch and MaxPitch can part of userinfo.

Probably not a good idea, considering the pitch limits exist in the first place due to the renderer.

[Edward-san]

True, but gzdoom has gl_maxpitch which can be set to an arbitrary value (ah, graf, btw: why can gl_maxpitch be negative? It should be constrained between 0 and 90, otherwise It could mess with the network code...).

So, why not allowing to set the max pitch also in zdoom, knowing the sw renderer's limits? It could be interesting for the user to limit the pitch for example between -10 and 10 degrees.

[Onar4241]

That would make more sense. I thought we were talking about DECORATE, which is why people need to remember to make intelligible bug reports.

First post above says that it only happens in ACS

[Onar4241]

So from what I understand, clamping is now optional via a flag?

[edward850]

First post above says that it only happens in ACS

You mean the post nobody can actually fucking read?

So from what I understand, clamping is now optional via a flag?

No.

(ah, graf, btw: why can gl_maxpitch be negative? It should be constrained between 0 and 90, otherwise It could mess with the network code...).

the player submits their own viewpitch anyway. Not having it on hand, It might not be a problem, as long as gl_maxpitch doesn't try to override what was received but rather only what is sent.