[Danukem]

supergoofy, on Jul 12 2009, 09:03 PM, said:

Every time, I delete the folder and then I unpack the new source in a new empty folder. So veryclean is not needed.

I still use gcc 4.3.3, because I don't trust 4.4.0


r1458 still crashes. I shot a rocket and boom SIGSEGV crash.

Could you at least try using 4.4.0? I think TerminX has been using it for a while now.

DanM is having the same crash on rocket fire (consistently) as you, so it's either something you are both doing wrong or it's a problem with the source.

[supergoofy]

Ok I'll move to gcc 4.4.0. I'll try tomorrow or later today.

[Stabs]

which file do i d/l for 4.4.0?

[supergoofy]

I used dukeplus 2.03 with July 12 fix and compiled eduke32 with gcc 4.4.0 but r1458 crashes. I shot a rocket and boom SIGSEGV crash:



Fatal Signal caught: SIGSEGV. Bailing out.
13 318847168
3621: 1335 sizeat 1 1
3624: 1336 ifnotmoving 318847228
3626: 1336 ifcount 3 318847228
3629: 31337 {
3630: 1336 resetcount
3631: 1336 move 318838136 8
3634: 1336 ifrnd 64 318847224
3637: 1336 killit
3638: 31337 }
3639: 1338 enda
3640: 1342 randvar 422 2
3643: 1343 addvar 422 1
3646: 1344 setvar 428 0
3649: 1345 getactor 253 4 415
3653: 1346 cactor 2360
3655: 1347 getactor 253 15 407current actor: 775 (6318)
g_errorLineNum: 1338, g_tw: 4



Normal game without Duke Plus works ok.

So there is an incompatibility of Duke Plus and r1458 ?

[TerminX]

Try r1459.

[supergoofy]

ok thanks. I will report back.


compiled r1459 with gcc 4.4.0. now dukeplus rocket launcher works ok. I also fired up all weapons and all working ok. I assume the problem is fixed.


Thank you TX

[Stabs]

i 2nd that

cheers tx

[Lord Misfit]

r1459: the new additions for qsprintf have appearantly borked it beyond belief.






And no, I did not change anything for any of these displays or all the other ones that are messed up, and they were working fine in the last build I was using [r1451]. It also tends to flicker between random tiles in the places of the vars, only occasionally showing the proper var displays for brief moments at random.

[supergoofy]

Yeah r1459 has problems with Naferia mod.

Maybe there are other problems with other mods/TCs.


For now I will use r1454. I compiled it with gcc 4.4.0 and I will test it later. The gcc 4.3.3 compiled binary of r1454 worked fine, thus I believe the gcc 4.4.0 compiled binary will also work fine.

[Danukem]

Hey I have a question that's actually on the thread topic!

How can one change the radius of a light SE after it has started working? It used to be (for example, in the movie I posted at the start of this thread) that I could change the radius by changing the hitag of the SE. But that doesn't work any more. Once the SE is spawned with a certain hitag, it is locked in at that intensity regardless of its hitag. I thought maybe the radius value had been moved to temp_data, so I checked all the htg_t members 0-9, but it's not there. I also checked whether there is another sprite spawned by the SE that is the actual light, but nope. I tried changing the shade of the SE, but that didn't work either. So I'm at a loss. How does the SE know what intensity it is supposed to be and how can I change it on the fly?

Is the light getting baked onto the surfaces? If so, cool, but then it needs to detect whether something has changed and update, because lights need to be able to change intensity and move around and stuff.

EDIT: Plagman restored CON control to light SEs today!! Yeah!!

[BlackDays]

Well, low and behold, Polymer works perfectly on my laptop. (With good FPS to boot.)

Nice work fellas. *Works on an awesome mod.*

*seriously does.*

[MusicallyInspired]

Pics or it didn't happen.

[BlackDays]

 MusicallyInspired, on Jul 13 2009, 06:23 PM, said:

Pics or it didn't happen.

When I finish this weapon model and a few more textures. So it didn't happen yet.

[Lord Misfit]

r1460: qsprintf stuff still broken [in the same manner as in r1459].

Again, did not change any of the displays, and my r1460 was not self-compiled [and given to me by another], so blame can't go to me for it's not working if it's somehow due to how it was compiled. :\

[supergoofy]

I also tested r1460 and still has problems with Naferia mod

[TerminX]

Something really strange is going on with the new versions of the quote commands. I'll fix it when I have time (soon!) The commands work fine in all of my test cases but fail in NR for some reason... I'm sure it'll be worked out sometime today or tomorrow though. Anyway, try and remember not to let bug reports on a blatantly unfinished product turn into actual complaining.

[supergoofy]

No complains here. We just report bugs thus when the time comes for the new official snapshot, eduke32 will have less bugs.

[Hendricks266]

TerminX, I have a request. Could you make a new structure member that stores the previous palette an actor had? The game already uses something similar for the getlastpal command. As an extension, could you make it so that when a model has a "pal" token in a frame block, it uses this new structure member for the "pal" in a skin block. Example:

// Frozen Beach Babe (3670)
model "vacation/sprites/ice/3670_beachbabe_ice.md3" {
   scale 1.5 shade 0 flags 1

   skin { surface 0 file "vacation/sprites/pickups/0059_glasses.png" }

   skin { surface 1 file "vacation/sprites/ice/3670_beachbabe_ice.png" }

   skin { pal 0 surface 2 file "vacation/sprites/ice/3670_beachbabe_ice.png" }
   skin { pal 10 surface 2 file "vacation/sprites/ice/3670_beachbabe_ice_10.png" }
   skin { pal 11 surface 2 file "vacation/sprites/ice/3670_beachbabe_ice_11.png" }
   skin { pal 12 surface 2 file "vacation/sprites/ice/3670_beachbabe_ice_12.png" }
   skin { pal 13 surface 2 file "vacation/sprites/ice/3670_beachbabe_ice_13.png" }
   skin { pal 14 surface 2 file "vacation/sprites/ice/3670_beachbabe_ice_14.png" }
   skin { pal 15 surface 2 file "vacation/sprites/ice/3670_beachbabe_ice_15.png" }
   skin { pal 16 surface 2 file "vacation/sprites/ice/3670_beachbabe_ice_16.png" }
   skin { pal 18 surface 2 file "vacation/sprites/ice/3670_beachbabe_ice_12.png" }
   skin { pal 21 surface 2 file "vacation/sprites/ice/3670_beachbabe_ice_21.png" }
   skin { pal 22 surface 2 file "vacation/sprites/ice/3670_beachbabe_ice_11.png" }
   skin { pal 23 surface 2 file "vacation/sprites/ice/3670_beachbabe_ice_23.png" }

   frame { [b][color="blue"]pal 1[/color][/b] name "FRAME1" tile 3670 }
}

As an alternative, "lastpal" be implemented alongside "pal" and "surface".

[Danukem]

nevermind delete

[High Treason]

I just had a thought, you know the Save Game screen? (Silly question, you will know it ) is there any plans to make the screenshot for saved games be a screenshot from polymer instead of old 8-Bit?

Silly idea i know, but it's something i always wanted to see, so i was wondering if you were planning on doing this?

[The Commander]

I don't think it's a silly idea as some mods use high res tiles etc without 8bit counter parts so the screen shot preview looks black or has black boxes in them. (Though it might bump the file size up a little of the save games)

[TerminX]

1461 fixes the qsprintf issue, and maybe adds a few new problems for you guinea pigs to discover.

[Stabs]

r1461 : Normal game runs fine but when dukeplus is added it gets to the start nuke symbol & crashes, tried it from several different fresh installs.

Not sure if its a glitch everyone else has but i have transparent squares around most sprites that turn other sprites see through when they are on top of each other.

[Danukem]

DanM, on Jul 14 2009, 07:53 PM, said:

r1461 : Normal game runs fine but when dukeplus is added it gets to the start nuke symbol & crashes, tried it from several different fresh installs.

I'm not using 1461 yet so I don't know about that.

DanM, on Jul 14 2009, 07:53 PM, said:

Not sure if its a glitch everyone else has but i have transparent squares around most sprites that turn other sprites see through when they are on top of each other.

I have noticed that in Polymer from before 1461.

[Stabs]

thought i'd mention that sprite bug, hasnt really been mentioned before

[BlackDays]

DanM, on Jul 14 2009, 08:17 PM, said:

thought i'd mention that sprite bug, hasnt really been mentioned before

Yeah, I seem to remember seeing that in Polymost, actually.

[supergoofy]

with r1461 dukeplus crashed once (the crash happened when I first shrunk the enemy and then shoot with freezer)

and then a tinny enemy killed me






not only that but I started now a new game and I saw that it can fly (look at the center of the screenshot)




Something wrong with the shrinker....

Tinny aliens shooting and flying? Wtf?


The thing is that I like DukePlus more that way. It's funnier

[Danukem]

And none of this happens with earlier builds, right?

[supergoofy]

I will test with r1454 or older

it doesn't happen with r1454. but when i shrink an enemy I cannot kill him unless I squash him with my foot.

also in r1454 when I used the freezer I got

Fatal Signal caught: SIGSEGV. Bailing out.
1670 318720444
53850: 31337 {
53851: 8271 state 318719528
53853: 8272 ifrnd 48 318720424
53856: 8273 spawn 2270
53858: 8274 debris 2408 12
53861: 8275 killit
53862: 31337 }
53863: 8277 else 318720460
53865: 8278 action 318720052
53867: 31337 }
53868: 8280 enda
53869: 8283 ifrnd 64 318720504
53872: 31337 {
53873: 8285 ifnosounds 318720500
53875: 8286 soundonce 336
53877: 31337 }
53878: 8288 else 318720548
53880: 8289 ifrnd 64 318720548
53883: 31337 {
53884: 8291 ifnosounds 318720544
53886: 8292 soundonce 337current actor: 162 (1062)
g_errorLineNum: 23911, g_tw: 124


It's probably a similar crash like the one that happened with r1461.

[Danukem]

I'll check it with 1460 (don't know why you would go back so far) but it really looks to me like a CON code problem. I'll bet there's some loophole with the new AI that allows them to resume normal behavior when shrunk. There was something like this a few months ago that I thought I had fixed...maybe I fixed one way it could happen but I missed another.

EDIT: I think I see the problem now, testing with 1460. AFAIK it only affects medic troopers with the advanced AI option on. The additional AI for that type of trooper has a bug which will wrongly transition them out of their shrunk behavior and keep them shrunk. I didn't get any crashes.