Segmentation fault

[tonni]

5.0.0.2 look like buffed char crash seq i can log any unbuffed char with no problem :) not fresh install cvs update over 5.0.0.1

thx in advance

[Cryonic]

OK, well must be something to do with the spell information parser as I'm not using the latest spell txt files from my EQ directory and I'm not seeing SEQ5 b2 crashing. It has been very stable for me running multiple chars with various buffs (including buffs will not stack messages).

[Dark]

I did a complete reinstall of my Red had 9 and indtalled beta 2 and the latest spell files and haveing no problems (apart from the windows docking where i dont want them too)

[Khallus]

Added the latest spells_us.txt and eqstr_us.txt (to /usr/local/share/showeq) and it fixed the seg fault I was getting.

[reaver]

Can't seem to pin it down on exactly when it happens.

But the seg fault always happens after a Spell will not take hold message. But it's not consistent - doesn't happen after every message, can't find a pattern on why certain messages seg fault and others are ok.

-reaver

[tamasine]

seq faults on startup - seeminly as soon as it sees a packet. both .txt files up todate.

[Dedpoet]

seq faults on startup - seeminly as soon as it sees a packet. both .txt files up todate.

tamasine, are you running any non-EQ traffic on your network? I know I get seg faults all the time if something like Kazaa is running.

[tamasine]

at present, nothing bar windows spew, ssh and web browsing. 5.0.0.1 worked fine, 5.0.0.2 segfaults. (and i disabled backup-my-binary so until i backout the cvs to the previous version [read catch the cvs servers up] i'm back on 4.*)

certainly no udp apart from windows - never did get around to running these funky filesharing thingies ;)

unfortunately, all the time since i compiled it i've been raiding - and haven't had time to gdb the binary (assuming this is still valid. been about 10 years since i last had to diagnose binaries beyond trace [and tracing it, it segfaults with no obvious system calls nearby apart from a malloc-brk])

[tamasine]

Program received signal SIGSEGV, Segmentation fault.
[Switching to Thread 16384 (LWP 496)]
0x404f7469 in QString::QString(QString const&) ()
from /usr/local/qt3/lib/libqt-mt.so.3
(gdb) bt
#0 0x404f7469 in QString::QString(QString const&) ()
from /usr/local/qt3/lib/libqt-mt.so.3
#1 0x08136569 in ZoneMgr::restoreZoneState() (this=0x82e7830)
at /usr/local/qt3/include/qstring.h:936
#2 0x0813625e in ZoneMgr (this=0x82e7830, parent=0xbffff7c0,
name=0x818b611 "zonemgr") at zonemgr.cpp:61
#3 0x080a894b in EQInterface (this=0xbffff7c0, dlm=0x28, parent=0x0,
name=0x8183088 "interface") at interface.cpp:238
#4 0x080679fc in main (argc=1, argv=0xbffff7c0) at main.cpp:682

[Zaphod]

Originally posted by tamasine
Program received signal SIGSEGV, Segmentation fault.
[Switching to Thread 16384 (LWP 496)]
0x404f7469 in QString::QString(QString const&) ()
from /usr/local/qt3/lib/libqt-mt.so.3
(gdb) bt
#0 0x404f7469 in QString::QString(QString const&) ()
from /usr/local/qt3/lib/libqt-mt.so.3
#1 0x08136569 in ZoneMgr::restoreZoneState() (this=0x82e7830)
at /usr/local/qt3/include/qstring.h:936
#2 0x0813625e in ZoneMgr (this=0x82e7830, parent=0xbffff7c0,
name=0x818b611 "zonemgr") at zonemgr.cpp:61
#3 0x080a894b in EQInterface (this=0xbffff7c0, dlm=0x28, parent=0x0,
name=0x8183088 "interface") at interface.cpp:238
#4 0x080679fc in main (argc=1, argv=0xbffff7c0) at main.cpp:682

Strange, try and see if you have a stray lastZone.dat file somewhere either in your /usr/local/share/showeq or ~/.showeq directories. If you find one, delete it. If that isn't it, make sure you did a clean build of the software. try doing a 'make -f maintainer-clean' and then going through the usual build process.

Enjoy,
Zaphod (dohpaZ)

[tamasine]

yes boss ;)

didn't know about those files. killed it and no startup segv

[Leetlewon]

I am getting segfaults too, but only when the spells window is used. Close spell window and no problems.

I assume that since I've updated from cvs that I have the most recent spells_us.txt.

[Leetlewon]

Oops nevermind - I see now that this file comes from my windoze machine in the everquest directory. I will copy over a fresh version and see if it helps.

[old_fart]

I am gettng a seg fault with only the spawn list open:

Spell: XXXX has begun casting 'Spear of Torment' - Casting time is 4.2 Seconds
General: Unknown: 2371: XXXX, A Temple Patroller, 688
Spell: XXXX has begun casting 'Focus of Soul' - Casting time is 8 Seconds
Spell: XXXX has begun casting 'Dragon Force' - Casting time is 0 Seconds
General: Unknown: 2371: XXX, A Temple Patroller, 10
Spell: You begin casting Forlorn Deeds. Current Target is A_Temple_Patroller02(633)
Spell: You begin casting 'Forlorn Deeds' - Casting time is 6 Seconds
Spell: XXXX has begun casting 'Spear of Torment' - Casting time is 4.2 Seconds
Info: SpellItem 'Forlorn Deeds' finished.
General: Unknown: 2371: XXXX, A Temple Patroller, 803
Spell: XXXX has begun casting 'Dragon Force' - Casting time is 0 Seconds
General: Unknown: 2371: XXXX, A Temple Patroller, 10
Spell: You finish casting 'Forlorn Deeds'.
Spell: Unknown: 107
/usr/bin/startseq: line 3: 12050 Segmentation fault showeq

Is there any other log or logging that I can enable that might be usefull?

-j

EDIT: had to take out character names and replace with XXXX.

[Zaphod]

Originally posted by old_fart
I am gettng a seg fault with only the spawn list open:

Spell: XXXX has begun casting 'Spear of Torment' - Casting time is 4.2 Seconds
General: Unknown: 2371: XXXX, A Temple Patroller, 688
Spell: XXXX has begun casting 'Focus of Soul' - Casting time is 8 Seconds
Spell: XXXX has begun casting 'Dragon Force' - Casting time is 0 Seconds
General: Unknown: 2371: XXX, A Temple Patroller, 10
Spell: You begin casting Forlorn Deeds. Current Target is A_Temple_Patroller02(633)
Spell: You begin casting 'Forlorn Deeds' - Casting time is 6 Seconds
Spell: XXXX has begun casting 'Spear of Torment' - Casting time is 4.2 Seconds
Info: SpellItem 'Forlorn Deeds' finished.
General: Unknown: 2371: XXXX, A Temple Patroller, 803
Spell: XXXX has begun casting 'Dragon Force' - Casting time is 0 Seconds
General: Unknown: 2371: XXXX, A Temple Patroller, 10
Spell: You finish casting 'Forlorn Deeds'.
Spell: Unknown: 107
/usr/bin/startseq: line 3: 12050 Segmentation fault showeq

Is there any other log or logging that I can enable that might be usefull?

-j

EDIT: had to take out character names and replace with XXXX.

You're using ShowEQ 5.0.0.3 correct? And did you copy over spells_us.txt and eqstr_us.txt from your EQ machine? From those messages it doesn't look like you did.

Enjoy,
Zaphod (dohpaZ)

[KaL]

When I switched to ShowEQ v5.0.0.3, I was segfaulting frequently until I updated my spells* and eqstr* files.

The fact that ShowEQ crashes because of an outdated text file external to the program itself indicates a lack of robustness in that portion of the code.

However, I appreciate the efforts and work that went into SEQ, and that goes into it every time a patch comes out. Good show.

[Hobo]

Well Kal, I do believe it is a beta release afterall.

[BlueAdept]

I was getting segs until I copied over the new files. Now it works great. Even though I havent been on a raid yet, I havent had any segs from grouping.

[Zaphod]

Kal,

First thing that you've got to remember is that this is a beta of major changes to code that relies on knowledge reverse engineered from the EQ packet data and files and also trying to figure out what the hairy old uncommented code was doing.

Also, how well do you think EQ will work with an outdated/corrupte/missing data file. The data in those files are referenced by data in the data stream coming from the EQ servers. We do try to be robust, but it is hard to account for all the variant [i]stupid user tricks[i] (tm) nor tool variants that one will encounter...

As far as fixes for the underlying cause of this particular problem, and a couple others that still existed in 5.0.0.3 apply the patch I attached to this post (http://seq.sourceforge.net/forums/showthread.php?s=&threadid=4475).

If you are truely concerned about the robustness of the code you are more then welcome to submit patches that improve it.

Enjoy,
Zaphod (dohpaZ)

[KaL]

I wasn't trying to knock it.. I know keeping up with patches is a lot of hard work.

That's why I added the kudos at the end of my post. Just saying, maybe relying on an external file transfer isn't such a great idea.

[KaL]

Oh, and unfortunately I have had one segfault since updating my .txt files. Just one, though. I have no correlating data to add concerning what may have happened, though. But, the .txt files don't fix 100% of the segfaults.

[Zaphod]

Originally posted by KaL
Oh, and unfortunately I have had one segfault since updating my .txt files. Just one, though. I have no correlating data to add concerning what may have happened, though. But, the .txt files don't fix 100% of the segfaults.

That's why I pointed to a patch in my post above... Did you apply the patch I attached to this post (http://seq.sourceforge.net/forums/showthread.php?s=&threadid=4475)?

Enjoy,
Zaphod (dohpaZ)

[Dark]

FYI Z,

Applied the patch and had no crashes since. online time about 24 hours over 2 days.

Cheers
Dark

[KaL]

I'll give it a shot, thanks.

And Merry Christmas :)

[lildr00d]

NM - Fixed the problem