[Paul B]

Helixhorned, on 29 February 2012 - 07:36 AM, said:

Ah, it's good that you're bringing this up; I had noticed the laser-endpoint glitch with Hypercore but didn't give it much thought. The core issue is the "4294901808" (0xFFFF0030 in hex): that number is outside the range of the CON number type, namely the signed 32-bit integer. I fixed the parsing to what it (coincidentally) was before and added a warning (though it's more to remind me than the CON coder). Read the revision log if you're interested in the gory details. Also, CON supports literal hexadecimal constants with the "0x..." syntax, though the validation of the text leaves something to be desired.

You know, that's exactly what I thought too.

[Danukem]

Helixhorned, on 29 February 2012 - 07:36 AM, said:

Ah, it's good that you're bringing this up; I had noticed the laser-endpoint glitch with Hypercore but didn't give it much thought. The core issue is the "4294901808" (0xFFFF0030 in hex): that number is outside the range of the CON number type, namely the signed 32-bit integer. I fixed the parsing to what it (coincidentally) was before and added a warning (though it's more to remind me than the CON coder). Read the revision log if you're interested in the gory details. Also, CON supports literal hexadecimal constants with the "0x..." syntax, though the validation of the text leaves something to be desired.

That's great, thanks. But could you tell me what number I should be using for the clipmask? I don't know where the clipmask bits are documented.

[LeoD]

Building r2392 fails for Microsoft Visual Studio 2010.

1>gamedef.obj : error LNK2019: Verweis auf nicht aufgel÷stes externes Symbol "_strtoll" in Funktion "_parse_decimal_number".
1>eduke32.exe : fatal error LNK1120: 1 nicht aufgel÷ste externe Verweise.
1>NMAKE : fatal error U1077: ""C:\Programme\Microsoft Visual Studio 10.0\VC\bin\link.EXE"": R³ckgabe-Code "0x460"
1>  Stop.

[Helixhorned]

DeeperThought, on 29 February 2012 - 08:39 AM, said:

That's great, thanks. But could you tell me what number I should be using for the clipmask? I don't know where the clipmask bits are documented.

The clipmask used for the collision functions is really a combinations of two -- the lower 16 bits are checked against the cstat of walls, and the upper 16 against the sprite cstat. Also, for floors and ceilings, "blocking" and "hitting" is mapped from the walls' one.

LeoD, on 29 February 2012 - 10:24 AM, said:

Building r2392 fails for Microsoft Visual Studio 2010.

1>gamedef.obj : error LNK2019: Verweis auf nicht aufgel÷stes externes Symbol "_strtoll" in Funktion "_parse_decimal_number".
1>eduke32.exe : fatal error LNK1120: 1 nicht aufgel÷ste externe Verweise.
1>NMAKE : fatal error U1077: ""C:\Programme\Microsoft Visual Studio 10.0\VC\bin\link.EXE"": R³ckgabe-Code "0x460"
1>  Stop.

Sigh, strtoll is mandated by C99, ninety-nine! Have they really not implemented these, or are we missing a compiler switch or something?
edit: also it's curious that it passes the compilation stage then

[Plagman]

Heh; they're not planning on supporting C99 for the C compiler at all as far as I know. I think this came up before, but don't use C99-isms or it'll break the native Windows builds.

[Hendricks266]

We are in luck! http://www.enchantedage.com/node/231

#if defined(_MSC_VER)
#define strtoll _strtoi64
#endif

[Gambini]

Not a mapster bug but the problem was brought here before so:

I´m still using the build with dissapearing viewscreens so i dont know if this is related and or is fixed already but: if i use a viewscreen and watch a room with monsters, they seem to "see me through the camera´s eye" and start shooting at it. I´m playing some unreleased stuff not by me so i can´t provide right now the example map.

[Hank]

 Gambini, on 29 February 2012 - 06:20 PM, said:

Not a mapster bug but the problem was brought here before so:

I´m still using the build with dissapearing viewscreens so i dont know if this is related and or is fixed already but: if i use a viewscreen and watch a room with monsters, they seem to "see me through the camera´s eye" and start shooting at it. I´m playing some unreleased stuff not by me so i can´t provide right now the example map.

I am trying to recreate this phenomena. No luck. My actors are either 'sleeping' or simply running around, none even look at the camera. Is this a basic map or is there coding involved?

[Micky C]

I think the smart tagging system needs a bit of tweaking. I thought it was supposed to always choose the next highest tag, but in a lot of cases, it actually makes huge jumps. I looked at one of the CBP 8 maptag files and it's already at over 400000! So either it's a super dooper fantastically dynamic map with like a million effects, or there's a bug.

[fgsfds]

Noclip doesn't work in mapster. Error "noclip is not a valid command or cvar". It doesn't work since build 2028.

[Micky C]

I'm not sure if you're up to speed with the latest eduke32 but they come out fairly frequently and the latest one is here: http://dukeworld.duk.../20120301-2394/
2028 is quite outdated.

Anyway for no-clip, press (' n) once or twice. Sounds like you're typing it in the console or something which is completely unnecessary. Use the keyboard command instead.

[fgsfds]

First, I use the latest 2394 build right now. I just meant that noclip console command worked fine before build 2028, was broken in 2028 and still doesn't work in latest build.
And second, ' n button enables\disables sprite clipping. It's not what noclip command do.

[Micky C]

Pressing ' n a second time enables complete no-clipping.

[fgsfds]

Thanks.
But I still don't understand why console command no longer works.

[fgsfds]

About this

Quote

r1973 | helixhorned | 2011-08-20 04:54:16 -0700 (Sat, 20 Aug 2011) | 1 line disable undo/redo editor feature until I can debug the disappearing sprites

5 months have passed and it's still disabled?

[underTaker]

I'm encountering weird bug with Mapster32. I'm not sure if you already know about this, but i think it's worth to say.
If i leave mapster32 without focus - you know, alt + tab to another window for a few seconds the windows goes to Not Responding state, and all i can do is to turn it off.
It's happening on latest mapster from synthesis version on Dukeworld. And many previous ones. I'm using Windows 7 x64.

[fgsfds]

underTaker, on 04 March 2012 - 06:20 AM, said:

I'm encountering weird bug with Mapster32. I'm not sure if you already know about this, but i think it's worth to say.
If i leave mapster32 without focus - you know, alt + tab to another window for a few seconds the windows goes to Not Responding state, and all i can do is to turn it off.
It's happening on latest mapster from synthesis version on Dukeworld. And many previous ones. I'm using Windows 7 x64.

Have the same problem. Vista\7\8 x32\64, different PCs, whatever. And I wrote about it several times but it seems no one cares.

[Hank]

I disabled my IE browser and set another browser as the default, then the ALT TAB stopped functioning for Mapster32. I re-established IE as the default browser; and now ALT TAB works. Don't ask, I have no clue what the relation is here or even if there is one; this is just an observation on one computer.

OS is Windows 7 64bit, HP-SP1

[underTaker]

Also if this is helpful - i almost always (lets say 80% of times) have crash of the Mapster when closing it. It returns standard windows error message about application crash.
Alt + tab works perfectly fine for me, and i have changed browser.

[Kyanos]

underTaker, on 05 March 2012 - 09:49 AM, said:

Also if this is helpful - i almost always (lets say 80% of times) have crash of the Mapster when closing it. It returns standard windows error message about application crash

Can I guess that %80 is when you quit from 3D mode? I get that same error every time I quit Mapster while in 3D mode.

[Helixhorned]

I get the exit crash every time I quit either Mapster or Duke in Windows Vista or 7 when compiled with GCC (doesn't happen with MSVC). I tried catching it once, but I think it happens in the GCC process teardown code (i.e. below main). Anyway, I don't think it's related to the more severe Alt-TAB crashes/hangs, which I couldn't reproduce yet.

[underTaker]

Drek said:

Can I guess that %80 is when you quit from 3D mode? I get that same error every time I quit Mapster while in 3D mode.

Hmm, i've never actually payed atention to that. I'll look for that next times.

Helixhorned said:

I get the exit crash every time I quit either Mapster or Duke in Windows Vista or 7 when compiled with GCC (doesn't happen with MSVC). I tried catching it once, but I think it happens in the GCC process teardown code (i.e. below main). Anyway, I don't think it's related to the more severe Alt-TAB crashes/hangs, which I couldn't reproduce yet.

I suppose i can't really help with Visual Studio JIT since it's not .net ;/ If I could help any other way though let me know.

[Gambini]

I also have a crash message "send report" everytime i quit either Mapster32 or Eduke32.

[Paul B]

Helix, when you have time and would like to visually see some bugs with TROR please see my attached map file. I am trying to create a new map by implementing more TROR in the map because I am hoping to perform some effects with TROR that have not been possible with the previous engine. However upon playing with TROR I've come across some show stoppers.

Some of the glitches include:

1) Looking through a masked transparent wall which doesn't display the correct wall tile.

2) Looking through a masked transparent wall (Window) allows you to see through a TROR covered sectors. For example: In the map attached if you close the bridge using the switch to make a sub floor if you look at the Bridge floor through a masked wall you can see right through it. Also TROR floor tiles aren't shown properly through masked walls.

3) Another thing I noticed which I am sure you probably are aware of... Is the space between the TROR floors need to be pretty big before Duke can seemlessly pass through them. In my case where I make a bridge floor Duke is unable to properly pass through the floor because it requires having the TROR extensions really close together. Is it possible for you to close that gap so that Duke doesn't get stuck between invisible TROR sectors when he jumps through them? I find if you duck while floating on an invisible TROR floor you can sometimes pass through it.

Please see my example one room map file if you have any questions about what i am talking about. This one room exposes some of the problems encountered when using TROR.

Thanks for your time,
Paul

Attached File(s)

•  problem.zip (4.74K)
  Number of downloads: 246

[Danukem]

Decided to try out Polymer with the latest build. Used some modified CON code but nothing that should cause what is pictured. That shot is from E5L2 in the recently released Metropolitan episode. What is happening is that every sprite shows the skybox in the transparent part of the sprite. It didn't start happening until I got to that area. I don't know if that's a new bug, probably not.

[Micky C]

Muahaha Micky C to the rescue.

Paul B, on 10 March 2012 - 09:54 PM, said:

1) Looking through a masked transparent wall which doesn't display the correct wall tile.

2) Looking through a masked transparent wall (Window) allows you to see through a TROR covered sectors. For example: In the map attached if you close the bridge using the switch to make a sub floor if you look at the Bridge floor through a masked wall you can see right through it. Also TROR floor tiles aren't shown properly through masked walls.

3) Another thing I noticed which I am sure you probably are aware of... Is the space between the TROR floors need to be pretty big before Duke can seemlessly pass through them. In my case where I make a bridge floor Duke is unable to properly pass through the floor because it requires having the TROR extensions really close together. Is it possible for you to close that gap so that Duke doesn't get stuck between invisible TROR sectors when he jumps through them? I find if you duck while floating on an invisible TROR floor you can sometimes pass through it.

4) Also if it is possible to allow for more than 8 TROR sectors stacked if you can implement that as I feel i'll be needing more layers.
Please see my example one room map file if you have any questions about what i am talking about. This one room exposes some of the problems encountered when using TROR.

Thanks for your time,
Paul

1) Not sure what that's all about
2) That's a polymer issue which is up to plagman to fix. It appears that masked TROR surfaces are the last things drawn and aren't drawn at all if there's a transparent thing between the surface and the player such as a sprite or masked wall. Had to do a few annoying things to overcome this in parkade.map.
3)Yep that's known about.
4)I've done more than 8 stacks and I'm pretty sure I've seen Zaxtor do more than 8 stacks

Trooper Dan, on 10 March 2012 - 10:36 PM, said:

duke0002.jpg

Decided to try out Polymer with the latest build. Used some modified CON code but nothing that should cause what is pictured. That shot is from E5L2 in the recently released Metropolitan episode. What is happening is that every sprite shows the skybox in the transparent part of the sprite. It didn't start happening until I got to that area. I don't know if that's a new bug, probably not.

That is a problem with the map itself related to the second point I mentioned above. What the mappers have done in that particular sector is made the floor masked (there are 4 modes; normal, transparent 1, transparent 2, and masked, which looks like normal). The thing is those floors don't need to be masked because there's no TROR involved. Perhaps a mapster script can be used to reset all the floors' bits to normal if there's ever going to be an update.

[Helixhorned]

Paul B, on 10 March 2012 - 09:54 PM, said:

1) Looking through a masked transparent wall which doesn't display the correct wall tile.

I don't see this. Which renderer do you use and where should I look?

[Paul B]

Helixhorned, on 12 March 2012 - 11:58 AM, said:

I don't see this. Which renderer do you use and where should I look?

Sorry for being vague, I always tend to use the Polymer renderer. For example, if you go into the room that uses the default brick texture and look back through the masked glass wall towards the floor where the extended TROR sectors are you will see the floor tile will disappear and show up as a black texture.
If you trigger the bridge with the switch on the wall then look down at the bridge through a masked wall you will see that the bridge will disappear.
If you try and jump down onto the bridge because the TROR sectors are so close together it seems Duke gets stuck on each layer. The closer the layers are together the more of a problem it becomes.
Thank you for your time & don't worry I will renew my membership for this year!
Paul

[fgsfds]

Not sure it's a bug, but I don't think it should work this way.
Make a sector and place faced down ceiling or floor rise\fall effector.

This is how it looks in mapster and, I suppose, should look in game.
http://i.imgur.com/cxJr3.png

But this is how it looks in game after activation.
http://i.imgur.com/uEBYZ.png

And with gpspeed (lotag 5000) that sould only change speed.
http://i.imgur.com/LEcn7.png

The higher gpspeed's lotag, the more this distance from actual starting point.

Attached File(s)

•  map.zip (361bytes)
  Number of downloads: 213

[Helixhorned]

I think it was Micky who complained about too large automatically-assigned tags a while ago. I believe that this is happening because some walls' lotags have linking semantics in the game, and the tag-label system also picks up those who have a lotag assigned as the TROR upwall. Can somebody verify that this is the case? The wall picnums in question are TECHLIGHT2, TECHLIGHT4, WALLLIGHT4, WALLLIGHT3, WALLLIGHT1, WALLLIGHT2, BIGFORCE, and W_FORCEFIELD. Guess I'll make an exception for these.

edit: Oops, it's a couple more. No need to verify, I'm pretty certain that this is the cause.
Here's the additional picnums:

Spoiler

ACCESSSWITCH, SLOTDOOR, LIGHTSWITCH, SPACEDOORSWITCH, SPACELIGHTSWITCH, FRANKENSTINESWITCH, MULTISWITCH, DIPSWITCH, DIPSWITCH2, TECHSWITCH, DIPSWITCH3, ACCESSSWITCH2, LIGHTSWITCH2, POWERSWITCH1, LOCKSWITCH1, POWERSWITCH2, HANDSWITCH, PULLSWITCH, ALIENSWITCH, DOORTILE5, DOORTILE6, DOORTILE1, DOORTILE2, DOORTILE3, DOORTILE4, DOORTILE7, DOORTILE8, DOORTILE9, DOORTILE10, DOORTILE22, DOORTILE18, DOORTILE19, DOORTILE20, DOORTILE14, DOORTILE16, DOORTILE15, DOORTILE21, DOORTILE17, DOORTILE11, DOORTILE12, DOORTILE23