[Danukem]

James, on Oct 10 2009, 04:37 AM, said:

Is there any way I can use readgamevar to create unlockable stuff in multi? I've noticed that readgamevar only reads the local player's variables in multiplayer - which is fine for things that don't affect gameplay like certain computer screens and achivements I've coded...but I want to make unlockable weapons and since this modifies gameplay it desyncs the game.

I mean I've coded in a computer system that reads saved gamevars and displays them onscreen as a status thing. When I press K to see my ally's stuff it just reads the gamevars in my config file rather than his; but of course from his point of view it's displaying his stuff and pressing K when I'm using the PC displays his again as well. When I try and read if the weapon is unlocked, since it's unlocked for me it works ok but since it's still locked on his system it comes out negative and so desyncs since the game doesn't understand that our configs would have different saved vars or something.

What if you did something like this:

gamevar initplayer NO 1
gamearray weapmods 8 // each element is a bitfield storing all weapon mods for a player
gamevar myweapmods 0 1 // this is what is saved to the ini file
gamevar temp 0 0


state initp

  setvar initplayer YES
  readgamevar myweapmods
  getactor[THISACTOR].yvel temp
  setarray weapmods[temp] myweapmods

ends

actor APLAYER MAXPLAYERHEALTH PSTAND 0 0

ifvare initp NO state initp

So what you do is read the array element corresponding to the player number, rather than reading the gamevar from the file. Then you can display each player's stuff separately. When you get a new weapon mod, you add that to myweapmods, save that gamevar, and then update the array accordingly.

EDIT: It would make things messy if players changed their IDs in mid-game, but the basic concept seems sound.

[Jblade]

Cool! I'll give that a shot as soon as I can - thanks

[Danukem]

I have a request. I would like CON access to which sprites are actually being rendered (i.e. which ones have tsprite), and also to which sectors/walls are being rendered. AFAIK, there is currently no way to access this information in CON. Since the information is already in the system, checking it would allow for some optimizations. For example, it could be a substitute or a precondition for expensive visibility checks. Also, if you want to know whether a certain wall is being rendered (say, because you want to use the DISPLAYROOMS invalid texture trick to project another part of the map onto it and you want to avoid the extra rendering if the wall is not being rendered), then direct access to the information is vastly superior to the alternatives (such as checking visibility to a nearby sprite).

[Plagman]

How exactly would that system that tells you which sprites are being rendered would differ from just looking at the tsprite array today?

For walls and sectors, there are several things to consider:

• The game code doesn't know which walls and sectors are getting rendered. Before implementing a way for the game code to tell CON, we would need to implement a way for the engine code to tell the game code. This would make the engine <-> game limit even more broken than it already is.
  
• Rendered doesn't necessarily mean visible, and possibly vice versa.
  
• These results would vary depending of the renderer that's being used, and would even vary with different performance settings in the same renderer. That means CON scripts could potentially make wrong assumptions and only work on the renderer they were developed against.

I know why you need it and I agree that it would help in some cases, but it probably needs a little more thought than just giving that kind of information.

[Danukem]

Plagman, on Oct 14 2009, 10:05 AM, said:

How exactly would that system that tells you which sprites are being rendered would differ from just looking at the tsprite array today?

For walls and sectors, there are several things to consider:

• The game code doesn't know which walls and sectors are getting rendered. Before implementing a way for the game code to tell CON, we would need to implement a way for the engine code to tell the game code. This would make the engine <-> game limit even more broken than it already is.
  
• Rendered doesn't necessarily mean visible, and possibly vice versa.
  
• These results would vary depending of the renderer that's being used, and would even vary with different performance settings in the same renderer. That means CON scripts could potentially make wrong assumptions and only work on the renderer they were developed against.

I know why you need it and I agree that it would help in some cases, but it probably needs a little more thought than just giving that kind of information.

Thank you for the detailed reply.

About tsprite access:

I know that there's an event EVENT_ANIMATESPRITES which works sort of like EVENT_GAME, except for tsprites. But tsprites get processed in that event only if the corresponding game world sprite has had bit 4 (16) added to its mdflags member. I was confused and I also thought that sprites would be forced into the tsprite array by that flag whether they were rendered or not (which is why I thought it was impossible to determine whether they were actually being rendered). Fortunately the reality is more sensible than my assumption!

[Chip]

Simple question: What sort of problems can occur by not using the scriptsize command?
Are they dire? are they things which can alter an actor's behaviour or sometihng else?

[Plagman]

Not really, the automatic script buffer increasing code should take care of that for you.

[Danukem]

Plagman just fixed a bug in the showview command which would cause it to mess up moving sectors (e.g. rising doors) when used in EVENT_DISPLAYROOMS. It wasn't working properly with the movement interpolation code when called from that event. The problem was framerate dependent, which means that any use of showview in EVENT_DISPLAYROOMS would break sync in multiplayer.

Here's a link to the fixed build.

Once again, Plagman, thanks!

[Fox]

I am having an error with new snapshot (EDuke32 2.0.0devel 20091017)

Quote

42154928
33317: 31337 {
33318: 1899 getactor 253 19 332 230 332 14
33325: 1901 divvarvar 332
33327: 4 guniqhudid 253 17
33330: 1902 cos 385
33332: 4 guniqhudid 253 15
33335: 1903 sin 386
33337: 4 guniqhudid 253 15
33340: 1904 hitscan
33341: 4 guniqhudid 253 0
33344: 4 guniqhudid 253 1
33347: 4 guniqhudid 253 2
33350: 4 guniqhudid 253 13 385 386 332current actor: 288 (2605)
g_errorLineNum: 0, g_tw: 386

This is the CON part that causes the error:

Quote

getactor[THISACTOR].zvel ATEMP
mulvar ATEMP 16384
divvarvar ATEMP sprite[THISACTOR].xvel
cos ANGX sprite[THISACTOR].ang
sin ANGY sprite[THISACTOR].ang
hitscan sprite[THISACTOR].x sprite[THISACTOR].y sprite[THISACTOR].z sprite[THISACTOR].sectnum ANGX ANGY ATEMP
RETURN RETURN ATARGET RETURN RETURN RETURN 4294901808

[Jblade]

I've got a question about build script - what can it be used for? there's lots of instructions on the wiki but no tutorials or anything that tells me what it can do.

[Lord Misfit]

Time for a new question [I think ]:

I'm developing a shield buff [like a status buff] for NR, which is supposed to in theory "block" and sometimes "reflect/deflect" mystic-based attacks [one example can be the Octabrain's/Cycloid's mindblasts. I got it to "block" okay, but I was curious if anyone had ever tried the "deflect" part in their own code runs before.

Basically what I'm hoping to pull off is that when a projectile that can be blocked/deflected hits the shield, the original Projectile copies certain properties to a new projectile of identical strength and stuff but with the player's owner id, and bounce it away from them. For the best effect in my opinion, I need to know if it's possible to make a projectile disappear mid-impact[as in it won't show it's impact/explosion animation and just vanishes instantly]. I'm not sure if it can be done with custom or hard-coded projectiles or not.

I'm going to eventually work at this, but I got other tweaks for earlier stuff to worry about before I do, and I figured on dropping this set of questions on you guys in the meantime.

[Fox]

htg_t may help. I am not sure, but it may be acessible before the projectile impact.

[Danukem]

Lord Misfit, on Oct 22 2009, 07:54 PM, said:

Basically what I'm hoping to pull off is that when a projectile that can be blocked/deflected hits the shield, the original Projectile copies certain properties to a new projectile of identical strength and stuff but with the player's owner id, and bounce it away from them. For the best effect in my opinion, I need to know if it's possible to make a projectile disappear mid-impact[as in it won't show it's impact/explosion animation and just vanishes instantly]. I'm not sure if it can be done with custom or hard-coded projectiles or not.

I've made a few player force fields that deflect projectiles. Just change the angle of the projectile, it's zvel and owner. You don't need to fire a new one. You can spawn an animation sprite at the point where it deflects and play a sound. Maybe the real problem is you don't know how to make the projectile deflect without it dying. You can have the projectile detect distance to the force field and hitscan for it, then turn it around before it actually hits anything. Another method is to allow it to hit, but use EVENT_KILLIT to cancel its death and then turn it around from there.

[lycanox]

Is there a way to kill actors when they are asleep.
I am trying to create a remote controlled missile. But it is unable to kill most monsters unless the player wakes them first.

It also seems that monsters behave differently when asleep. So is it possible to blow up Liz troopers when asleep.
But not possible to blow up Liz mans.

[Lord Misfit]

DeeperThought, on Oct 23 2009, 12:46 AM, said:

I've made a few player force fields that deflect projectiles. Just change the angle of the projectile, it's zvel and owner. You don't need to fire a new one. You can spawn an animation sprite at the point where it deflects and play a sound. Maybe the real problem is you don't know how to make the projectile deflect without it dying. You can have the projectile detect distance to the force field and hitscan for it, then turn it around before it actually hits anything. Another method is to allow it to hit, but use EVENT_KILLIT to cancel its death and then turn it around from there.

I should let you guys know I have gotten most of this all working now with deflecting stuff. It's not the exact method above, but I do have EVENT_KILLIT involved with a lot of it, and I still let the original projectile die and spawn a new one[but I can destroy the originals instantly too]. The only minor snafu with it is that when it deflect's a COOLEXPLOSION1, I try to get it's count and actorcount just before it's deleted through htg_t 0 & 2, but they always come back as 0 so when the new projectile is spawned, it spawns at it's default animation frames instead of part-way into the animation already.

Because of this I've been considering remaking a "COOLEXPLOSION2" or something as a replacement so deflected mindblasts actually hurt non-players and such, but I'm trying to get an idea of the animation speed and fuse/lifetime of the original COOLEXPLOSION1 is in exact values. Does anyone know that information on this topic? It would be a big help for me.

[XThX2]

If nothing else works, you can always set an actorvar for it to point the actioncount by "ifactioncount 1 setactorvar[THISACTOR].frame 1" but I'm sure htg_t would work fine. Where are you using htg_t 0 and 2 ? You can try EVENT_GAME just to see if it works.

[Danukem]

Lord Misfit, on Oct 24 2009, 10:30 AM, said:

I should let you guys know I have gotten most of this all working now with deflecting stuff. It's not the exact method above, but I do have EVENT_KILLIT involved with a lot of it, and I still let the original projectile die and spawn a new one[but I can destroy the originals instantly too]. The only minor snafu with it is that when it deflect's a COOLEXPLOSION1, I try to get it's count and actorcount just before it's deleted through htg_t 0 & 2, but they always come back as 0 so when the new projectile is spawned, it spawns at it's default animation frames instead of part-way into the animation already.

Because of this I've been considering remaking a "COOLEXPLOSION2" or something as a replacement so deflected mindblasts actually hurt non-players and such, but I'm trying to get an idea of the animation speed and fuse/lifetime of the original COOLEXPLOSION1 is in exact values. Does anyone know that information on this topic? It would be a big help for me.

Why shouldn't the animation reset, and why would anyone care ? I appreciate attention to detail, but this crosses the line into being a weird fetish. You are spending time and adding code to deal with "problems" that no one cares about (and I'm not saying that based only on this feature).

[Hendricks266]

		case COOLEXPLOSION1__STATIC:
			t->shade = -127;
			t->cstat |= 8192;
			t->picnum += (s->shade>>1);
			break;

				if (s->picnum == COOLEXPLOSION1)
				{
					s->shade++;
					if (s->shade >= 40) KILLIT(i);
				}

Does that help?

[Fox]

How can I draw something at event DISPLAYROOMS and so that it will be draw BEHIND the transparent water ror sectors? Because whenever I draw something there using roratesprite, it appers in front of the ror sectors.

[CruX]

Getting arrows to stick in a wall (if they actually 'hit' a wall) is something I've been trying to hammer out since I started working on the bow for my mod, and I still haven't managed to get it working. I thought at first I could do it with the htg_t struct, but that only appears to work for bullet-type projectiles (come to think of it, I've only managed to get it to do anything for shotgun, SHOTSPARK1 and chaingun projectiles). Now I'm starting to think it has something to do with the hitscan command, but if returning what's under the crosshair is it's only function, then I'm having some trouble imagining how that could pull the desired effect either. Can anyone help me figure this out?

[XThX2]

You could possibly check for the velocity of the arrow, and make it hitscan the time it is spawned (once) to get the wall it hitscanned's x/y/z values. When the velocity is set to 0, you can make the arrow spawn another arrow (this time, no projectile code; possibly with different art) at the coordinates it left off. (That's the first thing that comes to my mind now)

[CruX]

XThX2, on Nov 4 2009, 10:56 AM, said:

You could possibly check for the velocity of the arrow, and make it hitscan the time it is spawned (once) to get the wall it hitscanned's x/y/z values. When the velocity is set to 0, you can make the arrow spawn another arrow (this time, no projectile code; possibly with different art) at the coordinates it left off. (That's the first thing that comes to my mind now)

Hmmm.. I did try something earlier where I coded the arrow to spawn a SHOTSPARK1 on impact, then coded the SHOTSPARK1 to use htg_t if it was spawned by an arrow. This way it'd know whether or not it was hitting a wall, and if it did I had it change it's picnum to the arrow (a different set, like you suggested). It's gotten me a bit closer to what I wanted, but now for some reason the arrow ends up spinning around in place when it spawns (even though I know I gave it the correct frames). Strange stuff...

[XThX2]

As far as my tests resulted, ifspawnedby when used by shotspark1 does not have any effect. Instead, I do this;

	  findnearsprite IONPARTICLEBOOM 500 returnvar1
	  ifvarn returnvar1 -1
	  {
		  getactor[returnvar1].htg_t 2 temp
		  ifvare temp 0
			  killit
		}

Where IONPARTICLEBOOM is a bullet based projectile that spawns an explosion when it hits a wall. Try it this way and see if it works. Of course, you would get the necessary information from it before killing it.

[CruX]

XThX2, on Nov 4 2009, 11:53 AM, said:

As far as my tests resulted, ifspawnedby when used by shotspark1 does not have any effect. Instead, I do this;

	  findnearsprite IONPARTICLEBOOM 500 returnvar1
	  ifvarn returnvar1 -1
	  {
		  getactor[returnvar1].htg_t 2 temp
		  ifvare temp 0
			  killit
		}

Where IONPARTICLEBOOM is a bullet based projectile that spawns an explosion when it hits a wall. Try it this way and see if it works. Of course, you would get the necessary information from it before killing it.

Well, I changed the code from changing picnums to just having it spawn the stuck arrow actor and that got rid of the crazy rotating problem. Last hurdle is trying to figure out how to get it closer to the wall to actually create the illusion that it's really stuck in the wall (right now it spawns a few inches away so, it's just sort of hanging in mid-air). Just to see what would happen, I coded the arrow to move further inwards towards the wall once it spawns then stop after a few tics, but it vanishes if it goes too far, and just hangs in the middle of the wall if it's actually 'within' it. Is this something that only a model can fix?

[XThX2]

Is that arrow rendered as vertically flat ? I would probably suggest you change the y-repeat of the arrow and cstat of it to make it vertically flat. That way, it would look like it's actually in the wall. (It can possibly make the arrow look longer than it actually is though )

[CruX]

XThX2, on Nov 5 2009, 11:37 AM, said:

Is that arrow rendered as vertically flat ? I would probably suggest you change the y-repeat of the arrow and cstat of it to make it vertically flat. That way, it would look like it's actually in the wall. (It can possibly make the arrow look longer than it actually is though )

At this point I should really know what vertically flat means, but if it's aligned with the floor, then no it's not. Also I found another way to do it that's pretty simple (so simple, i can't believe I didn't think of it beforehand ). I just made a bullet-type projectile that had no graphic or extra value, and spawned a custom decal. I then coded the arrow to spawn another invisible custom actor on impact that would shoot said projectile that would leave behind this decal and that decal, once being spawned, would spawn the arrow in the wall. Sort of has a jerry-rigged feel to it, but it works pretty nicely.

[Fox]

Whenever I use startlevel command in multiplayer, it seems like to be resulting in a votation beetween players... but I wanted it to go right through levels. How can I do it?

[XThX2]

EmericaSkater, on Nov 5 2009, 12:22 PM, said:

At this point I should really know what vertically flat means, but if it's aligned with the floor, then no it's not. Also I found another way to do it that's pretty simple (so simple, i can't believe I didn't think of it beforehand ). I just made a bullet-type projectile that had no graphic or extra value, and spawned a custom decal. I then coded the arrow to spawn another invisible custom actor on impact that would shoot said projectile that would leave behind this decal and that decal, once being spawned, would spawn the arrow in the wall. Sort of has a jerry-rigged feel to it, but it works pretty nicely.

Vertically flat means well, open mapster32, put a sprite and press R, the first time you do it; what you see is vertically flat. (Aligned with wall)

Can they be picked up, by the way ?

[CruX]

XThX2, on Nov 6 2009, 07:38 AM, said:

Vertically flat means well, open mapster32, put a sprite and press R, the first time you do it; what you see is vertically flat. (Aligned with wall)

Can they be picked up, by the way ?

Picked up AND climbed on. You'll get to see when I nail down a few more changes and post some more videos. (I got the full version of FRAPS just for you )

[Danukem]

Ilovefoxes, on Nov 6 2009, 06:10 AM, said:

Whenever I use startlevel command in multiplayer, it seems like to be resulting in a votation beetween players... but I wanted it to go right through levels. How can I do it?

It's not supposed to stop for voting, that sounds like a bug.