ZDoom

Thanks to you for your time

Very nasty indeed! Basically all the refactoring that has happened was needed to truly reveal what was going on. Thanks for your patience

Just recompiled and tested, no crash this time, Great ! This was a really nasty bug

Thanks for the save, its been very much appreciated. I know you gave some saves before but they'd just got a bit stale with all the refactoring.

We've repaired this issue. basically that map in the set isn't constructed right and has either an upper or lower sector link without an appropriate partner. On the original code, this would have been an -1 out of bound access, which could explain why you had it on one system and not another. We've switched everything out to use pointers and they will absolutely crash if a null access occurs, which would be why it happens on Windows.

This fix probably didn't make it into the most recent devbuild, but it sounds like you're building the code from Github anyway. Feel free to spin out another build and test it out

I've managed to get a save that seems to isolate the issue : https://www.mediafire.com/file/61yvs09c ... dsave/file. Just fire the napalm gun in the opened hole on the pipe and the game crashes immediately.
The issue now seems to happen on win10, too, here is the crash report. I used the latest drd team 1.4pre-28.
Edit : i also patched the maps to the latest version using this https://www.moddb.com/mods/blood-trauma ... -map-fixes, maybe you need these to load the savegame...

Is there anyway you could get me a save just before the crash, then explain what you're doing to cause the crash? I think it's the only way we're going to try and isolate this. I have a Windows and a Linux setup so I can test in both platforms.

Yeah, i just recompiled from git and tried again, the program dies exactly on the same point as always...

Just saw your reply on the GLES post. Is this still occurring as of 1.3.0?

Just for fun, i re-tried with 1.1.0pre-663 with lldb and got this, it is slightly different from before :

Process 12428 stopped
* thread #1, name = 'raze', stop reason = signal SIGSEGV: invalid address (fault address: 0x555560413000)
frame #0: 0x00005555558fc72c raze`getzrange(vec3_t const*, short, int*, int*, int*, int*, int, unsigned int) + 396
raze`getzrange:
-> 0x5555558fc72c <+396>: movswq (%rax,%rdi,2), %rcx
0x5555558fc731 <+401>: movq 0x28(%rsp), %rdi
0x5555558fc736 <+406>: movq %rcx, %rax
0x5555558fc739 <+409>: shlq $0x4, %rax
(lldb) bt
* thread #1, name = 'raze', stop reason = signal SIGSEGV: invalid address (fault address: 0x555560413000)
* frame #0: 0x00005555558fc72c raze`getzrange(vec3_t const*, short, int*, int*, int*, int*, int, unsigned int) + 396
frame #1: 0x0000555555cd9a44 raze`Blood::GetZRange(spritetype*, int*, int*, int*, int*, int, unsigned int, unsigned int) + 468
frame #2: 0x0000555555d63d86 raze`Blood::MoveDude(Blood::DBloodActor*) + 4742
frame #3: 0x0000555555d69fd1 raze`Blood::actProcessSprites() + 2561
frame #4: 0x0000555555d6d043 raze`Blood::GameInterface::Ticker() + 499
frame #5: 0x00005555559254f5 raze`TryRunTics() + 3461
frame #6: 0x000055555592556d raze`MainLoop() + 77
frame #7: 0x000055555592c35e raze`RunGame() + 7006
frame #8: 0x000055555592d600 raze`GameMain() + 336
frame #9: 0x000055555576e063 raze`main + 259
frame #10: 0x00007ffff68fbb25 libc.so.6`__libc_start_main + 213
frame #11: 0x0000555555782d0e raze`_start + 46

Sorry for the late reply, but i didn't play a lot recently Anyway, i just tried to play the mod again using the latest git, but i still get the same segfault ... i tried to ask for help on the manjaro forum too, but they said to report the bug to the devs, so i'm back to the beginning...
Anyway, you can close the topic, i think that it's probably some obscure bug related to some library, who knows... i don't think that it will ever be fixed.

@SaveTheDoomer, any chance you could test this against the latest devbuild to see if its still an ongoing issue? There's been lots of changes in the last two months

https://devbuilds.drdteam.org/raze/

Rick Astley would tell you otherwise, so i am glad you stuck around with new info

Unfortunately, the bug is still there...

sinisterseed wrote: Judging by their profile OP uses Linux.

Perhaps it was edited because i did read a Windows reference in one of the earlier posts.

SaveTheDoomer wrote: I tried to upload the coredump, but i was told it's useless. At that point, i got frustrated with the whole thing and i decided to give up.

Rick Astley would tell you otherwise, so i am glad you stuck around with new info

Yes, your two crashes are the same, but I got a different one. I think they are caused by the same thing but I cannot fix this one unless I get it reproduced on my system.

I don't know, it seems the same thing that i posted earlier... except that i forgot the first part and just posted the bt, sorry...