Vavoom 1.30 - Segfault occurs in Strife when fighting Spectre entities. Segfault consistently happens about one second after shooting Macil with Sigil and the Spectre appears. I think it may have something to do with the lightning the Spectre shoots. I've seen segfaults with all the previous Spectres and the Programmer so that seems to suggest a lightning relation as they are the only monsters to shoot lightning so far. The Macil Spectre is the first I can reproduce on demand and cannot get past in game play. From debug.txt:

Maximum texture size: 8192
Found GL_ARB_multitexture...
Multitexture extensions found.
Max anisotropy 16.000000
Clamp to edge extension found.
Segmentation fault (core dumped)

Stack trace from core file and GDB:

#0  0x08290125 in VMemberBase::GetFullName() const ()
#1  0x081a3c95 in PR_Traceback() ()
#2  0x080fd6c5 in Host_CoreDump(char const*, ...) ()
#3  0x081a7e60 in VObject::ExecuteFunction(VMethod*) ()
#4  0x08170f36 in VEntity::eventTouch(VEntity*) ()
#5  0x0816c3aa in VEntity::CheckRelThing(tmtrace_t&, VEntity*) ()
#6  0x0816bbd4 in VEntity::CheckRelPosition(tmtrace_t&, TVec) ()
#7  0x0816c877 in VEntity::TryMove(tmtrace_t&, TVec, bool) ()
#8  0x08170a6a in VEntity::execTryMoveEx() ()
#9  0x081a3d68 in RunFunction(VMethod*) ()
#10 0x081a3e13 in RunFunction(VMethod*) ()
#11 0x081a3e13 in RunFunction(VMethod*) ()
#12 0x081a3e13 in RunFunction(VMethod*) ()
#13 0x081a3e13 in RunFunction(VMethod*) ()
#14 0x081a3e13 in RunFunction(VMethod*) ()
#15 0x081a3fae in RunFunction(VMethod*) ()
#16 0x081a3e13 in RunFunction(VMethod*) ()
#17 0x081a7c9d in VObject::ExecuteFunction(VMethod*) ()
#18 0x08164892 in VEntity::SetState(VState*) ()
#19 0x081665c0 in VEntity::execSetState() ()
#20 0x081a3d68 in RunFunction(VMethod*) ()
#21 0x081a3e13 in RunFunction(VMethod*) ()
#22 0x081a7c9d in VObject::ExecuteFunction(VMethod*) ()
#23 0x08164892 in VEntity::SetState(VState*) ()
#24 0x08164cb7 in VEntity::AdvanceState(float) ()
#25 0x08166620 in VEntity::execAdvanceState() ()
#26 0x081a3d68 in RunFunction(VMethod*) ()
#27 0x081a3e13 in RunFunction(VMethod*) ()
#28 0x081a7c9d in VObject::ExecuteFunction(VMethod*) ()
#29 0x0819b4af in VThinker::Tick(float) ()
#30 0x08173130 in VLevel::TickWorld(float) ()
#31 0x082259e3 in SV_Ticker() ()
#32 0x08229660 in ServerFrame(int) ()
#33 0x080fc829 in Host_Frame() ()
#34 0x082cf5bb in main ()

Compiled with gcc 4.4.1 on Ubuntu 9.10, 32 bit. I can create and provide a core file and compiled binaries if needed. I'm trying to isolate the exact line where the segfault is occurring but haven't been successful yet. I think it may be in VObject::ExecuteFunction but that is just from working backwards from the stack while finding the signal processors and I'm not familiar enough with the code to know for sure.

This is fixed in the latest development version, I would recommend you get the source from the SVN repository, or if you are using Windows, you can get the latest compiled beta from here: [url:2woa2ihj]http://svn.drdteam.org/vavoom/

All right. I will give the SVN repository a try. Thanks.

I downloaded the SVN repository and built the trunk on Ubuntu 9.10/32 bit and have an entirely different problem when running the resulting build. I am not certain what the policy is for SVN builds. Since that is a work in progress, would a bug report be helpful or a nuisance? In the off chance it would be helpful, the new symptom is the game fails to load. Here's the bottom of debug.txt starting with what looks like the first error in case you are interested:

Sys_Error: Bad parent class, class Level, C++ GameObject, VavoomC Object)
- VClass::Define
- VPackage::Emit
- VPackage::LoadObject
- VMemberBase::StaticLoadPackage
- VPackage::Emit
- VPackage::LoadObject
- VMemberBase::StaticLoadPackage
- VPackage::Emit
- VPackage::LoadObject
- VMemberBase::StaticLoadPackage
- SV_Init
- Host_Init
Log: Doing C_Shutdown
Log: Doing CL_Shutdown
Log: Doing SV_Shutdown
Log: Doing delete GNet
Log: Doing V_Shutdown
Log: Doing T_Shutdown
Log: Doing Sys_Shutdown
Log: Doing delete GSoundManager
Log: Doing R_ShutdownTexture
Log: Doing R_ShutdownData
Log: Doing VCommand::Shutdown
Log: Doing VCvar::Shutdown
Log: Doing ShutdownMapInfo
Log: Doing FL_Shutdown
Log: Doing W_Shutdown
Log: Doing GLanguage.FreeData
Log: Doing ShutdownDecorate
Log: Doing VObject::StaticExit
Log: Doing VName::StaticExit
Uninitialised: Doing Z_Shutdown
STACK TRACE:

stack 0 0x82d689c frame 0 0xff829768
stack 1 0xf7193b56 frame 1 0xff8297a8
stack 2 0x80bd9e1 frame 2 0xff829828


ERROR: Bad parent class, class Level, C++ GameObject, VavoomC Object)
Z_Free after Z_Shutdown
... <277 Z_Free after Z_Shutdown lines removed for brevity>
Z_Free after Z_Shutdown

You are trying to run it with wrong data files.

Ah, I see. The contents of /usr/local/share/vavoom have changed between the SVN revision 4189 and Vavoom 1.30. Got it working, but my saved games are gone. No problem. I'll play up to the Macil battle and see if it still segfaults. Thanks.

Well now that I have SVN 4189 version of Vavoom working it looks like what I thought may be a bug related to lightning may be fixed. It is hard to say because even though I haven't seen any segfaults occur with the Programmer and Bishop, I cannot get past the Oracle Spectre so I haven't fought Macil. It looks like the Oracle's Spectre has become invulnerable in this build. I didn't have all that much trouble killing it in 1.30 with the Sigil but even with cheat codes and god mode I cannot take this particular boss down. Also I noticed a different behavior with flowing water and sewage. If you jump into the river in the village you cannot get out because the current is faster than how fast the player can run so you cannot reach the lift by the sewage entrance. Simple solution is to not jump in the river, but I don't know if this speed difference will make other parts of the game unplayable.

If you press the shift key and run does the river is still too fast?

[quote="Firebrand":3v51wedf]If you press the shift key and run does the river is still too fast? Yep. Jumping doesn't help either. The speed of the water varies depending on how close you are to the waterfall. Close to the waterfall the water is running faster than the player can run. But the water is running slow enough near the bridge that you can walk about 3/4 of the way to the lift before the current becomes too much. If the lift was near the bridge instead of the waterfall, you could easily expect that this was the way the game was supposed to work.

With some experimentation I found a way to reach the lift. It isn't possible to reach the lift using SVN 4191 (yes I do an update and rebuild before each try so my SVN number increases with each post). However if you walk backwards while in the water you can walk upstream to the lift. If you turn around while still in the water the current will push you away form the lift. I guess knowing that it would be possible to fight any strong current in the game, you just cannot see where you are going. <!-- s:lol: --><img src="{SMILIES_PATH}/icon_lol.gif" alt=":lol:" title="Laughing" /><!-- s:lol: -->

With SVN version 4195 I managed to play through Strife following the "not so good ending" path. Although I needed to use noclip to get into the space ship because the Sigil button wouldn't open the spaceship door like the walk through states. This might have been a side effect of my jumping on the button platform with an incomplete Sigil when it first appeared. If I would have retrieved the Oracle's Sigil piece first it might have worked. For some reason the Oracle Spectre is invincible in recent SVN versions of Vavoom if you fallow the best ending path. I can defeat the Oracle Spectre if I follow the Oracle's instructions and end up killing the Oracle after killing Macil and the Loremaster. In the not so good ending I end up skipping 1/4 of the game or more. No commons, mine, factory, or catacombs.

The invincible specter is fixed in the latest SVN revision, we are investigating the problem with the water currents.

I've just played through Strife as far as the Bishop fight, and also got lots of segfault crashes. I had vavoom 1.30 from the Fedora repositories. I've downloaded 1.33 from your downloads page and built it, and now find my save games don't show up, as a previous poster indicated. I used to run with vavoom -strife -iwaddir /opt/games/dos/STRIFE I now run with /usr/local/bin/vavoom -strife -iwaddir /opt/games/dos/STRIFE (i.e. the only change is the vavoom location following the cmake/make/install build steps). When I run with the new vavoom executable, my Load Game menu is empty. I can see my old save games, in ~/.vavoom/basev/strife/saves - . Are they stamped with the vavoom version number or something? How can I use them with vavoom 1.33? I'd really rather not play through the game again to reach the Bishop again, fun though it is. Any advice gratefully received <!-- s:) --><img src="{SMILIES_PATH}/icon_smile.gif" alt=":)" title="Smile" /><!-- s:) --> Nick

[quote="neek":1pjejzpi]I can see my old save games, in ~/.vavoom/basev/strife/saves - . Are they stamped with the vavoom version number or something? How can I use them with vavoom 1.33? I'd really rather not play through the game again to reach the Bishop again, fun though it is. NickAs far as I recall, there is no way to use save games between different versions of Vavoom. The saves store certain information that does not stay consistent when the program gets updated. At least, that's how it was last time I tried (which I think was way back with v1.15 or so?)