Bug dump

[Ritz]

I've reported most of these on IRC at some point, but that's a good way for information to get lost, so for posterity:

• Polyobject flats aren't affected by linedef type 491, PolyObject - Set Translucency.
• Conveyor belts work on slopes, but not FOF slopes.
• Polyobjects will trigger 300 - Continuous linedef executors, but not 301 - Each Time.
• A midtexture on a sector with 55 - Continuous Ceiling Mover won't move if the ceiling of that sector has F_SKY1 applied. This bug was introduced in 2.1.15.
• Running back and forth across a large number of dynamic FOF vertex slopes will eventually crash the game.
• Back textures on polyobjects no longer seem to render, so you can't do the original ACZ lift where you can see the inside walls through the missing flat. Or maybe I'm not building it right?
• Colormaps don't play well with dynamic slopes that have transparent flats.
• When polyobjects are parented, the child polyobject is unaffected by the rotation of the parent.
• Polyobjects can't rotate and follow a waypoint path simultaneously.

 

[Monster Iestyn]

• Polyobject flats aren't affected by linedef type 491, PolyObject - Set Translucency.

Actually they are last I recall, it's just that if your camera is underneath/inside a PolyObject sometimes the translucency sometimes the translucency stops working? Which of course is still a bug, but it's a known one at least.

• Conveyor belts work on slopes, but not FOF slopes.

What part of them doesn't work on FOF slopes exactly? The flat scrolling part or the carrying objects part?

• Polyobjects will trigger 300 - Continuous linedef executors, but not 301 - Each Time.

How are you triggering linedef executors with them exactly?

• A midtexture on a sector with 55 - Continuous Ceiling Mover won't move if the ceiling of that sector has F_SKY1 applied. This bug was introduced in 2.1.15.

Hm, sounds strange, provide a sample map?

• Running back and forth across a large number of dynamic FOF vertex slopes will eventually crash the game.

A crash log would help here, so I can check what exactly is causing the crash.

• Back textures on polyobjects no longer seem to render, so you can't do the original ACZ lift where you can see the inside walls through the missing flat. Or maybe I'm not building it right?

This bug has already been reported before and has since been fixed.

• Colormaps don't play well with dynamic slopes that have transparent flats.

I've already fixed the bugs with colormaps on sloped flats too, after seeing you talk about the issues with them in IRC.

• When polyobjects are parented, the child polyobject is unaffected by the rotation of the parent.

Unaffected in what way? I have an old test map where several child polyobjects definitely do rotate with the parent (and still do IIRC), if you mean actually rotating around their own anchors.

• Polyobjects can't rotate and follow a waypoint path simultaneously.

I also have a test map that defies this claim, I think you're supposed to use the "override" version of the Rotate special for combining the two.


Yeah I'll look around for those test maps and add them to this post later, so you can check for yourself.

 

[Ritz]

Actually they are last I recall, it's just that if your camera is underneath/inside a PolyObject sometimes the translucency sometimes the translucency stops working? Which of course is still a bug, but it's a known one at least.

Nope:

What part of them doesn't work on FOF slopes exactly? The flat scrolling part or the carrying objects part?

Carrying, scrolling is fine.

How are you triggering linedef executors with them exactly?

With 80 - Trigger floor touch applied to the polyobject sector. Redenchilada can confirm that this is busted.

Hm, sounds strange, provide a sample map?

I think I already brought this to your attention before? It might've been Redenchilada. Look at the skybox object in whatever versions of my map you've got, one of the four chains isn't moving. I'll resend it if I have to.

A crash log would help here, so I can check what exactly is causing the crash.

SRB2 v2.1.15 -ERROR LOG-


srb2win caused an Access Violation in module srb2win.exe at 0023:0052c65e.
Exception handler called in main thread.
Error occurred at 8/6/2016 12:46:14.
F:\Misc\Games\SRB2\srb2win.exe, run by Administrator.
8 processor(s), type 586 6.15363.
Program Memory from 0x00010000 to 0x7FFEFFFF
0 MBytes physical memory.
Write to location 00000272 caused an access violation.

Registers:
EAX=000000f4 CS=0023 EIP=0052c65e EFLGS=00010246
EBX=09142e00 SS=002b ESP=0028f4fc EBP=00000280
ECX=0000014c DS=002b ESI=0937aef8 FS=0053
EDX=00000028 ES=002b EDI=00000272 GS=002b
Command Line parameters: -nomusic -file GLZt10.wad Bytes at CS : EIP:
88 ?? 07 ?? 0f ?? b6 ?? 14 ?? 16 ?? 0f ?? b6 ?? 14 ?? 13 ?? 88 ?? 14 ?? 2f ?? 8d ?? 3c ?? 6f ?? 
Stack dump:

Unaffected in what way? I have an old test map where several child polyobjects definitely do rotate with the parent (and still do IIRC), if you mean actually rotating around their own anchors.

I also have a test map that defies this claim, I think you're supposed to use the "override" version of the Rotate special for combining the two.

Yeah I'll look around for those test maps and add them to this post later, so you can check for yourself.

"Unaffected" in the sense that if the parent and the child are rotating around their own anchors, I expect the child to rotate around the parent's anchor in addition to its own rotation. I can have the child rotate however it wants, but I can't combine the movement of the two objects. A test map on both counts would be swell, it's entirely likely I'm not setting this up correctly.

 

[Monster Iestyn]

"Unaffected" in the sense that if the parent and the child are rotating around their own anchors, I expect the child to rotate around the parent's anchor in addition to its own rotation. I can have the child rotate however it wants, but I can't combine the movement of the two objects.

Ah, well yeah that's different. That's not really a bug, just a feature that doesn't exist. One can make the child appear to rotate around the parent if you line their anchors up together, but otherwise ...yeah I don't think having it rotate about an axis of its own as well as around the parent is possible currently.