[Fixed] r2010 Player now better at climbing slopes

Bugs that have been investigated and resolved somehow.

Moderator: Developers

r2010 Player now better at climbing slopes

Postby Enjay » Mon Nov 30, 2009 4:03 pm

Not sure when this first appeared, but I *think* it is relatively recent.

Try the attached file. In older versions, (eg the last official release) the player cannot climb the slope but just sort of judders around at the bottom. In recent versions, he climbs it with ease.

I don't know about any other maps, but this messes quite royally with my Burghead mod where the old behaviour was very much expected and where being able to climb the slopes can mean missed teleport lines, failure to trigger scripts and the ability to get to parts of the map that the player was never meant to get to and also to get horribly stuck. Also, enemies that should not climb the slopes can now do that too.

[edit] And it's obviously a very subtle change because raising the height of the slope by just a single unit blocks the climb.[/edit]
User avatar
Enjay
Everyone is a moon, and has a dark side which he never shows to anybody. Twain
 
Joined: 15 Jul 2003
Location: Scotland

Re: r2010 Player now better at climbing slopes

Postby Graf Zahl » Mon Nov 30, 2009 4:12 pm

I suspect the new float->fixed conversion code which may produce very subtly different results. Apparently your slope was right at the threshold.
User avatar
Graf Zahl
 
Joined: 19 Jul 2003
Location: Germany

Re: r2010 Player now better at climbing slopes

Postby Enjay » Mon Nov 30, 2009 4:15 pm

Is it something likely to be changed from the engine side or should I go about fixing the map? (I'm doing a bugfix anyway to take account of another few problems - though this particular fix may be a little more complex than most).
User avatar
Enjay
Everyone is a moon, and has a dark side which he never shows to anybody. Twain
 
Joined: 15 Jul 2003
Location: Scotland

Re: r2010 Player now better at climbing slopes

Postby Xaser » Mon Nov 30, 2009 4:17 pm

I dunno... are you by chance referring to a perfect 45-degree slope?(can't check at the moment) -- I've always assumed those to be non-climbable, and I'm sure other mappers have done the same if that's the case.
User avatar
Xaser
secretly a supercomputer being a government
 
Joined: 20 Jul 2003
Location: .plɹoʍɹǝʌǝu.

Re: r2010 Player now better at climbing slopes

Postby Enjay » Mon Nov 30, 2009 4:22 pm

It is a 45 degree slope, yes.
User avatar
Enjay
Everyone is a moon, and has a dark side which he never shows to anybody. Twain
 
Joined: 15 Jul 2003
Location: Scotland

Re: r2010 Player now better at climbing slopes

Postby Xaser » Mon Nov 30, 2009 4:24 pm

Hmm, that's what I was afraid of. I'm sure such shapes are relatively common in the Dooming world, so it may be necessary to tweak the threshold on the engine side to accommodate for this. I doubt Burghead's the only mod to exhibit the error now.

It's 'just' a rounding thing, anyway, so hopefully it isn't too hard to correct a bit. Though I'm speaking out of my league on that point. :P
User avatar
Xaser
secretly a supercomputer being a government
 
Joined: 20 Jul 2003
Location: .plɹoʍɹǝʌǝu.

Re: r2010 Player now better at climbing slopes

Postby Graf Zahl » Mon Nov 30, 2009 4:33 pm

In that case the code needs adjusting. Apparently the new code rounds this critical value in the wrong direction.
User avatar
Graf Zahl
 
Joined: 19 Jul 2003
Location: Germany

Re: r2010 Player now better at climbing slopes

Postby Enjay » Mon Nov 30, 2009 5:03 pm

Well, I've started fixing up the map where this really causes a problem anyway. With a few scripts to switch on/off player blocking as and when required it should be an easier fix than I thought and it will allow me to solve a few minor inconsistencies that have bugged me for a while with the map anyway.

However, I agree that the likelihood of this affecting more than just my map is quite high and, therefore, it should be addressed.
User avatar
Enjay
Everyone is a moon, and has a dark side which he never shows to anybody. Twain
 
Joined: 15 Jul 2003
Location: Scotland

Re: r2010 Player now better at climbing slopes

Postby NeuralStunner » Mon Nov 30, 2009 5:09 pm

Which stinks, in a way. (I would consider 45 degrees to be climable, albeit slowly. Much over that though, not so.) Any chance at all this could become a parameter in MapInfo...?
User avatar
NeuralStunner
Magnets.
 
Joined: 21 Jul 2009
Location: Secret Bunker (where the magnets are)

Re: r2010 Player now better at climbing slopes

Postby Enjay » Mon Nov 30, 2009 5:42 pm

Funny thing is, when I first discovered that 45 degrees was not climable I thought that it was a bit shallow (I mean, I could easily do it, but I can't run at 90mph ;) ). However, I just got used to it. Like all parameters, I suspect most experienced mappers have checked the limits of movement (eg, checking the lowest wall that will block the player for both walking and jumping) and then mapped to those parameters.

Anyway, the map fix turned out to be a lot simpler than I feared it might be. I'm all done now, aside from testing it a few times to check that I didn't miss any important blocking lines and, perhaps, checking it against a newer exe just in case anything I have done gets messed up by the change back of the rounding - if that happens.
User avatar
Enjay
Everyone is a moon, and has a dark side which he never shows to anybody. Twain
 
Joined: 15 Jul 2003
Location: Scotland

Re: r2010 Player now better at climbing slopes

Postby NeuralStunner » Mon Nov 30, 2009 5:47 pm

If slopes were "slippery" it could possibly be a matter of the player's actual velocity needing to overcome the steepness of the slope. I just don't know how viable that would be with the existing physics code. Sadly, a lot of good ideas just aren't possible due to the tendency to break other things. :(
User avatar
NeuralStunner
Magnets.
 
Joined: 21 Jul 2009
Location: Secret Bunker (where the magnets are)

Re: r2010 Player now better at climbing slopes

Postby randi » Mon Nov 30, 2009 7:20 pm

Wow. All this paranoia over something that's easily changed. For the curious, the value in question, before conversion to an integer, is 46340.94921875. The old float->fixed point macro rounded toward zero. The new macro rounds toward nearest. (Or more specifically, it uses banker's rounding.) So I just need to increase by one the magic number used for determining what a steep slope is.
User avatar
randi
Site Admin
 
Joined: 09 Jul 2003

Re: r2010 Player now better at climbing slopes

Postby Xaser » Mon Nov 30, 2009 7:53 pm

randy wrote:Wow. All this paranoia over something that's easily changed.


Wouldn't you be disappointed if we weren't? :P
User avatar
Xaser
secretly a supercomputer being a government
 
Joined: 20 Jul 2003
Location: .plɹoʍɹǝʌǝu.

Re: r2010 Player now better at climbing slopes

Postby Graf Zahl » Tue Dec 01, 2009 12:45 am

Let's just hope that nobody ever created slopes which are slightly less than 45° and now fall below the threshold and are climbable... ;)
User avatar
Graf Zahl
 
Joined: 19 Jul 2003
Location: Germany

Re: r2010 Player now better at climbing slopes

Postby Enjay » Tue Dec 01, 2009 3:01 am

randy wrote:Wow. All this paranoia over something that's easily changed.

:lol: Fair comment.

Sadly, I suppose, being able to climb a 45 degree slope is actually the preferable situation, or rather it would be if it had been possible from the start. Mind you, with Doom's messed up 1:1.2 ratio, 45 degree slopes look steeper than 45 degrees anyway.

Graf, you're just being paranoid now. ;)
(Yes, I know that was the joke.)
User avatar
Enjay
Everyone is a moon, and has a dark side which he never shows to anybody. Twain
 
Joined: 15 Jul 2003
Location: Scotland


Return to Closed Bugs

Who is online

Users browsing this forum: No registered users and 2 guests