[Diaz]

So I'm guessing the fix was for sprites only (check if floor sprites look right without being powers of two, because they didn't in the past)

[Helixhorned]

It is only for sprites. I'm concerned about performance when it would be used with floors/ceilings, and so didn't implement it.

[fgsfds]

Is this really affects the performance that much?

[fgsfds]

How to make auto returning elevator? For example you lower the lift and it rises back after a while. I can make it with switch-activating elevators but what about simple ones?

[Paul B]

empy, on 17 May 2012 - 08:36 AM, said:

How to make auto returning elevator? For example you lower the lift and it rises back after a while. I can make it with switch-activating elevators but what about simple ones?

To do that you would need to use two, two way trains so that the train would activate the elevator to go back down to the main floor after so many seconds which would be determined from the SE 10 on the two way train which would activate the elevator to go back down again after it has been called.

[TerminX]

Can the few of you who had problems with Mapster32 locking up after using alt+tab in windowed mode test r2657 and higher to see if the issue is fixed?

[Paul B]

TerminX, on 17 May 2012 - 09:36 AM, said:

Can the few of you who had problems with Mapster32 locking up after using alt+tab in windowed mode test r2657 and higher to see if the issue is fixed?

Personally, I don't know how anyone can use Mapster or Eduke in Windows 7. It really doesn't work very well for me especially from a performance perspective. I did try running Mapster r2659 in Windowed Mode on my Windows 7 64Bit Pro PC. My video card in this test system was an: Nvidia 7600GS and I was able to jump back and forth from Windows to Mapster. At some point I did lose my windows cursor in Windows and it didn't come back until I quit Mapster.

However, my mouse was present in the Mapster window just didn't seem to carry over to windows consistently. Like the mouse focus was lost when jumping from Mapster to Windows and back and forth a few times. The mouse focus only seemed to stay in Mapster after that. I also noticed that if I "ALT" + "TAB" out of mapster I was unable to drag or move the Mapster window at all. It was stuck on my screen in a certain spot and I was unable to move the Mapster window around because the mouse would immediately jump into the Mapster editor instead of allowing me to drag the window around on my desktop.

I am also running a dual display with my windows 7 pro desktop extended on to a second monitor. The last time I ran windows 7 and Mapster in fullscreen with dual monitor support it hung my windows 7 PC up solid. I had to give my PC the 1 finger salute to restore my computer as it was frozen up solid.

I hope this information contributes in some way.

[fgsfds]

I suffered because of this problem. Now it works fine, thanks.

Quote

I was able to jump back and forth from windows to Mapster. But at some point I lost my windows cursor in Windows and it didn't come back until I quit Mapster.

This happens only when you swich to mapster by clicking directly on mapster window. Use alt+tab or tasbar to switch between mapster and other windows.

[TerminX]

It sounds like you're talking about how Mapster32 grabs the mouse cursor and doesn't let it move outside of its window until you lower the console, which pretty much has nothing to do with the issue I was asking for people to test. You should also lower the console before using alt+tab in Mapster.

As for hanging your computer, you should probably look at your hardware and drivers for answers to that.

empy, on 17 May 2012 - 10:07 AM, said:

I suffered because of this problem. Now it works fine, thanks.

Yay, that's what I wanted to hear.

[Paul B]

TerminX, on 17 May 2012 - 10:13 AM, said:

It sounds like you're talking about how Mapster32 grabs the mouse cursor and doesn't let it move outside of its window until you lower the console, which pretty much has nothing to do with the issue I was asking for people to test. You should also lower the console before using alt+tab in Mapster.

Thank you TerminX! Lowering the console in mapster before "ALT" tabbing totally helps and does exactly what it should do with the window! That's good to know!

Looks like a solid update! Now how about that netcode? =)

[Diaz]

Paul B, on 17 May 2012 - 09:57 AM, said:

Personally, I don't know how anyone can use Mapster or Eduke in Windows 7.

My new laptop comes with Windows 7 and I don't have any problems at all. Performance isn't great due to the fact the laptop comes with a GeForce GT315M, but other than that, it works just fine.

[Hank]

TerminX, on 17 May 2012 - 09:36 AM, said:

Can the few of you who had problems with Mapster32 locking up after using alt+tab in windowed mode test r2657 and higher to see if the issue is fixed?

I had no real issue, yet since the fix I don't need to call the console to ALT+TAB in Mapster. I tested this now for 30 minutes, works very well.
Cheers

[fgsfds]

A couple of bugs that I've reported recently:
http://forums.duke4....233#entry123233
http://forums.duke4....542#entry124542
http://forums.duke4....492#entry123492

[Paul B]

I have come across an interesting situation. I was play testing one of my maps tonight and I realized that if I kill a stay put alien trooper if he fakes his own death he comes back again but not as stay put. He is free to roam. I'm not sure this should happen this way as enemies labelled as "stayput" should remain stayput. Typically stay put is a good way to control the AI so they don't go somewhere you don't want them to.

This also makes me wonder if playing on Damn I'm good difficulty if the stayput AI respawns after being killed do they still stayput? I'll check this out and post back when I know the answer.

Okay same thing with "Damn I'm Good" difficulty. If you kill an alien and the alien is a stayput alien when he respawns he respawns as free roaming alien. This also happens with the Enforcer as well as the Brown Trooper aliens. I'd imagine that the Commaders and anything else that is labelled stayput will also have this problem as well.

Whats your thoughts on this Helix? is this something that should be corrected?

[Helixhorned]

empy, on 09 May 2012 - 12:32 AM, said:

Is this really affects the performance that much?

Attached is a test map containing one room with the ceiling tiled with a regular power-of-two texture, and the floor with non-power-of-two floor sprites. On my setup (Linux x86_64, Core2Duo 3GHz, GCC 4.6 with optimizations, 1680x1050), I get these fps values when the looking all the way up or down:

power-of-2 texture (up): 90 fps
non-power-of-2 texture (down): 50 fps
sloped power-of-2 texture (up): 30 fps

These will very likely be different on an x86 system, and it's hard to predict the impact in a realistic setting. But besides not finding it very pressing to have non-pow2 floor/ceiling tiles, their sloped variants would need to be implemented too for completeness, and those would be terribly slow (extrapolating from the above values).

Attached File(s)

•  nonpow2tex.zip (290bytes)
  Number of downloads: 283

[Helixhorned]

empy, on 17 May 2012 - 08:54 PM, said:

http://forums.duke4....542#entry124542
http://forums.duke4....492#entry123492

Paul B, on 18 May 2012 - 09:28 PM, said:

Whats your thoughts on this Helix? is this something that should be corrected?

These all belong into the category of game tweaks or "enhancements" with the potential to break maps that rely on such subtleties (well not really for you, Paul). I think we shouldn't make gameplay changes to the original Duke3D, but rather leave this to the modders. Of course, SE and ST behavior is something you can't change easily from CON right now...

[Mark]

I ran the test map with vsync enabled and FPS stayed a constant 60.
With vsync off looking up= 1550 down=1441 ahead=1480 I couldn't see where the slopes were so I don't know the measurement of that. But looking down and moving 360 degrees it never dropped below the 1441.
same resolution and CPU as Helix running under Vista

[Helixhorned]

You're supposed to do that with the classic renderer, since Polymost/Polymer have no problems at all with non-pow2 textures.

[Mark]

One big DUH to me.

[Micky C]

I is there still going to be a mapster32 feature (or a.m32 script feature) that allows you to quickly orient floor and ceiling textures correctly together for textures that have their relativity bit set, and are in different sectors with parallel or perpendicular first walls? Alligning the textures manually is hard and tedious enough but you have to make sure the flipping is correct, which is very hard to do.

[Micky C]

Bug report. I was working on a set of child sectors with some sprites on red walls bounding two sectors, then sometimes when I draw a sector onto a wall with these sprites, the sprites will change their sectnum and disappear from 3D mode and 2D mode. However when I go into sideview mode, the sprites are still in their exact same position. Is there a feature that places sprites in the sector that they're physically in? Because I have a LOT of sprites affected by this and it would take forever to do it manually.

[Helixhorned]

Micky C, on 30 May 2012 - 03:10 AM, said:

I is there still going to be a mapster32 feature (or a.m32 script feature) that allows you to quickly orient floor and ceiling textures correctly together for textures that have their relativity bit set, and are in different sectors with parallel or perpendicular first walls? Alligning the textures manually is hard and tedious enough but you have to make sure the flipping is correct, which is very hard to do.

Hope you haven't missed r2715 and following. Still need to code up the actual alignment with relative-aligned reference ceilings/floors.

[Micky C]

Yeah, I was using 2743. Oh, nice work on the allignment thing, I've fixed up a bit of stuff already. Once panning is included I'll fix the rest

[Paul B]

I've encountered a minor bug in Mapster / Eduke with blockable sprites. If a sprite object is blockable and the sprite happens to extend into a extended TROR child sector that sprite extending into this adjacent sector is no longer blockable in the extended sector area. Can we please have sprites that are set with the blockable attribute be blockable when they happen to extend into a TROR sector?

Probably just a small over sight?

My current Mapster editor version is: 2745

[Diaz]

I think that's because sectnums are different, which means the sprite doesn't really "exist" in the extended sector.

[TerminX]

Diaz, on 10 June 2012 - 03:55 AM, said:

I think that's because sectnums are different, which means the sprite doesn't really "exist" in the extended sector.

Yes, this is correct.

[Paul B]

TerminX, on 10 June 2012 - 07:10 AM, said:

Yes, this is correct.

If this isn't worth fixing I realized that I could address this by placing invisible sprites in the upper extended sector and set them to blockable as a work around. So whatever you think is best, where there is a will there's away to make it work. =) I just wanted to report this observation incase it wasn't suppose to be this way.


Well, I decided to try and map edit on my laptop. I got into Mapster and then quickly realized that my laptop doesn't have a second enter key to switch between 2D/3D mode. So I was stuck in 2D mode. I tried many combinations with my main enter key like CTRL + Enter or Alt + Enter to see if that would make my LENTER key function as a R- Enter key. With no such luck, I went to the wiki article and I couldn't find anything listed under the keyboard commands that would allow me to switch modes. So does this mean any laptops without a second enter key on the keyboard are screwed or is it possible to bind another key combination to switching between 2D/3D Modes? You really don't realize how important that R-Enter key is until its just not there.

[Diaz]

Paul B, on 10 June 2012 - 09:01 AM, said:

Well, I decided to try and map edit on my laptop. I got into Mapster and then quickly realized that my laptop doesn't have a second enter key to switch between 2D/3D mode.

Your laptop doesn't have a FN key? FN+Enter would probably do the trick, if you have that key that is...

[Helixhorned]

There's the remap option in mapster32.cfg. I like to have the key that's a backlash on US keyboards as RENTER:

remap = 2B-9C

[Paul B]

Diaz, on 10 June 2012 - 02:17 PM, said:

Your laptop doesn't have a FN key? FN+Enter would probably do the trick, if you have that key that is...

Thank you Function + Enter worked!!! Should probably update the Wiki page.

Thank you both for the prompt reply! =)