MantisBT - Zandronum |
View Issue Details |
|
ID | Project | Category | View Status | Date Submitted | Last Update |
0000995 | Zandronum | [All Projects] Bug | public | 2012-08-28 04:01 | 2024-03-03 01:51 |
|
Reporter | NotJenova | |
Assigned To | Torr Samaho | |
Priority | normal | Severity | crash | Reproducibility | always |
Status | closed | Resolution | no change required | |
Platform | Linux | OS | Ubuntu | OS Version | 10.04 x86-64 |
Product Version | 1.0 | |
Target Version | | Fixed in Version | | |
|
Summary | 0000995: Server crash with Brutal Doom |
Description | Zandronum-server on 64 bit linux seems to crash when performing an execution with the berserk pack. We were unable to reproduce the bug on Windows, as it seems that the bug does not affect windows servers.
Here is the crash message:
zandronum-server: /home/blzut3/Code/Public/Zandronum/src/p_sight.cpp:646: bool P_CheckSight(const AActor*, const AActor*, int): Assertion `t1 != __null' failed.
Aborted
Here are the wads used:
skulltag_actors.pk3
cchest4.wad (or any wad)
brutaldoomv016b.pk3
If for some reason you need it, here is the command that was executed to start the server (in cooperative)
./zandronum-server -iwad iwads/doom2.wad -file skulltag_actors.pk3 -file cchest4.wad -file brutaldoomv016b.pk3 +sv_hostname "Test server" |
Steps To Reproduce | Grab the berserk pack and perform an execution on a monster. |
Additional Information | |
Tags | No tags attached. |
Relationships | related to | 0000849 | closed | Dusk | Stronghold STR20 triggers an assertion failure on server |
|
Attached Files | |
|
Issue History |
Date Modified | Username | Field | Change |
2012-08-28 04:01 | NotJenova | New Issue | |
2012-08-28 06:43 | NotJenova | Note Added: 0004492 | |
2012-08-28 07:49 | NotJenova | Note Deleted: 0004492 | |
2012-08-28 09:57 | Edward-san | Note Added: 0004493 | |
2012-08-28 11:13 | Dusk | Relationship added | related to 0000849 |
2012-08-28 11:13 | Dusk | Note Added: 0004495 | |
2012-08-28 18:43 | Torr Samaho | Note Added: 0004499 | |
2012-08-28 18:43 | Torr Samaho | Assigned To | => Blzut3 |
2012-08-28 18:43 | Torr Samaho | Status | new => feedback |
2012-08-28 19:02 | Blzut3 | Note Added: 0004500 | |
2012-08-28 19:05 | Blzut3 | Note Edited: 0004500 | bug_revision_view_page.php?bugnote_id=4500#r2459 |
2012-08-28 20:05 | Torr Samaho | Note Added: 0004501 | |
2012-08-29 01:19 | Blzut3 | Note Added: 0004503 | |
2012-08-29 01:19 | Blzut3 | Assigned To | Blzut3 => Torr Samaho |
2012-08-29 01:19 | Blzut3 | Status | feedback => assigned |
2012-08-29 12:22 | ZzZombo | Note Added: 0004504 | |
2012-08-29 19:38 | Torr Samaho | Note Added: 0004508 | |
2024-03-03 01:51 | Ru5tK1ng | Note Added: 0023284 | |
2024-03-03 01:51 | Ru5tK1ng | Status | assigned => closed |
2024-03-03 01:51 | Ru5tK1ng | Resolution | open => no change required |
Notes |
|
|
We need more informations.
Can you build zandronum in debug mode and reproduce this in gdb? We need its backtrace.
It would be more convenient if you try to reproduce this with a little example wad.
Also, since it aborts on zdoom code, the example wad should be readable by gzdoom 323 (obviously built with debug mode) in order to see if it's possible to reproduce this in the above program. |
|
|
(0004495)
|
Dusk
|
2012-08-28 11:13
|
|
|
|
|
That's really weird. Assertions are only triggered in debug mode, you shouldn't see something like in the official binaries ever. Blzut3, you are compiling the binaries in release mode, i.e. with -DCMAKE_BUILD_TYPE=Release, aren't you? |
|
|
(0004500)
|
Blzut3
|
2012-08-28 19:02
(edited on: 2012-08-28 19:05) |
|
Hmm that's weird. I was apparently building with the default value for CMAKE_BUILD_TYPE which does not appear to equate to any of proper values. (Produces binaries slightly larger than Release, and a lot smaller than RelWithDebInfo.) Do you want me to re-upload the binaries with that build variable properly set?
Edit: I just now noticed there's a 5th value which just uses CMAKE_CXX/C_FLAGS so that means the binaries were built with default optimization level. Question still remains though. Looks like the Mac binaries might also be affected.
|
|
|
|
Would be nice if you re-upload with that build variable properly set. This will fix the issue here: P_CheckSight explicitly checks that t1 and t2 are not NULL before using them, so even with the assertion disabled, the function will not crash. |
|
|
(0004503)
|
Blzut3
|
2012-08-29 01:19
|
|
All right, I've replaced the old builds with CMAKE_BUILD_TYPE=Release builds. I take it you still need to remove the assertion? |
|
|
|
Do I need to redownload my binary? |
|
|
|
Quote from Blzut3 I take it you still need to remove the assertion? Since CMAKE_BUILD_TYPE=Release eliminates the assertions from the binaries, there is no need to remove them. And since ZDoom put them in there for some reason, I rather keep them around.
Quote from ZzZombo Do I need to redownload my binary? If you are using the Linux x64 or the OS X binaries, then yes. |
|
|
|
Since this is so old, please make a new ticket if you have issues with 3.1 |
|