1  |   Order by index

May 5, 2003, 4:56am

Is there anything you guys can do with the software to put an end to the
crashes? Take your pick its either aworld.exe, AW_3d3.DLL or AW_3d8.DLL that
causes some errors. I know you guys want a log, but the crashes sometimes
take place after the log is 3 Gb big! Sorry guys my PC studders when
opening a notepad file that big!
Yesterday I was in A!!CTOY2 world and I was running 3d mode without T&L and
Defiance was running 3d mode with T&L enabled, we both crashed at the same
time, mine was that damned aw_3d3.dll again, his was aw_3d8.dll that
crashed. How can two people in different parts of the "REAL world" crash at
the same time?
When I too was using 3d mode with T&L enabled it was always aw_3d8.dll that
crashed.
Guys I have been with AW for 4 years now and never seen crashes like these!
3.3 was very stable, 3.2 never crashed!
I have tried 6 diffferent video cards all made by various manufactuers from
Voodoo to ATI to Diamond to Gforce4 cards, what gives with all the crashes?
Is it world server issues? browser issues? What gives?
Its not just my PC I know people running 3 Gb dual processors with 2Gb of
ram and they are crashing worse than me!
Here's all the info I can provide (sorry the log from AW is always to big to
open)

Date 05/05/2003 Time 02:24
AWORLD caused an invalid page fault in
module AWORLD.EXE at 0197:00459e2e.
Registers:
EAX=00000000 CS=0197 EIP=00459e2e EFLGS=00010206
EBX=055e0620 SS=019f ESP=0076fbd0 EBP=055e7eb0
ECX=055e2af0 DS=019f ESI=004be9be FS=1d67
EDX=00000000 ES=019f EDI=004dd7ed GS=0000
Bytes at CS:EIP:
8b 48 08 51 e8 29 70 fd ff 53 89 44 24 20 e8 4f
Stack dump:
055e2af0 004dd7e4 004be9be 055e7eb0 00000000 00000001 004be9c8 00457633
00000000 055e2ef0 055e065c 00000000 bff45801 0043cbe0 055e7eb0 05e18310
**********************************************************************
Date 05/05/2003 Time 02:26
AWORLD caused an invalid page fault in
module AWORLD.EXE at 0197:00459e2e.
Registers:
EAX=00000000 CS=0197 EIP=00459e2e EFLGS=00010202
EBX=058416b0 SS=019f ESP=0076fa4c EBP=0585c240
ECX=05841670 DS=019f ESI=004be9be FS=1d67
EDX=00000000 ES=019f EDI=004dd7ed GS=0000
Bytes at CS:EIP:
8b 48 08 51 e8 29 70 fd ff 53 89 44 24 20 e8 4f
Stack dump:

MS Windows ME (4.90. build 3000)
DirectX 8.1 (4.08.01.0881)
Video card: ATI Radeon 7000 PCI
Video Memory: 64 Mb DDR
Driver: ATIDRAG.DRV
Version: 4.13.01.9009
Sound card: Creative SBLive!
Driver: Emu10k1.VxD 4.06.00.0711
Total system memory:SDram PC-100 256 Mb
AW Browser mode Direct 3D mode with T&L enabled.

May 5, 2003, 9:13am

Bruce, there's a reason they call it a "Beta": It's not done yet!
Therefore, it's not going to be as stable as a finished product! :P

As for your log file, you can use Ananas' Tail program to grab the last
50 lines from your crash log.

http://oct31.de/tmp/tail.zip

Put it in your AW directory and open a DOS prompt to that directory and
type this:

tail -50 aworld.log > 50lines.txt

That'll take the last 50 lines of your log file and put it in a file
called 50lines.txt

[View Quote]
--
Goober King
Oh no.... Beta!
gooberking at utn.cjb.net

May 5, 2003, 12:19pm

I analyzed this world at length this morning, and found that it is packed
with mega high-poly objects. There are single objects that weigh in at 3
meg, or about 30 avatars worth of geometry. At 2n 6e, there are several
mega-poly objects right next to each other, so that even with low visibility
the user will still render them all. I estimate the user would be rendering
about 100 avatars worth of polygons if they were standing in this spot, even
with visibility set to minimum. This place pushes the limits of what AW (or
any 3d app) can do, and crashes should be expected under extreme conditions
like this.

Some of the larger objects caused AW to freeze for 30 seconds while they
were parsed. On slower systems, this would probably take much longer and
cause most users to give up and assume the program had locked up completely.

3.4 takes a little more memory and a little more CPU time than 3.3, so it
should be expected that very taxing worlds will push some users over the
edge and cause them to crash.

--
Shamus Young
Email: shamus at activeworlds.com
Home Page: www.shamusyoung.com
[View Quote]

May 5, 2003, 1:52pm

4 things.

1. His email is bill at activeworlds.com and I know you know that.
2. Why cross post? You just like attention right? As a beta tester, you're
suppose to keep things seperate.
3. Nice Subject title, I like the way you made it personal and so URGENT!
4. I love the idea of a personal and urgent message made public to ensure
Bill and AW are held accountable.

Geez

May 5, 2003, 3:40pm

A linker map before the file has been compressed
would help here, I hope they kept it


[View Quote]

May 5, 2003, 4:40pm

Briefly, I only crossposted into general.discussion because as you well
know, beta is also "open beta" and none of the open beta testers can post in
the beta newsgroups, it is an issue we are all facing, that is why I
cross-posted.
2nd yes I could email him personally, but then it comes across as ONE issue,
ONE case, that is not the case, this involves serveral citizens.
[View Quote]

May 5, 2003, 5:23pm

No YS we were out at 0n 250e the object limit stops at 0n 40E, we were
no-where near any objects to render.
[View Quote]

May 5, 2003, 6:36pm

"mrbruce" <A1CTWorld at aol.com> wrote in
news:3eb60b25 at server1.Activeworlds.com:

><<snipped>>

Tip: reset the log file every time you start AW. Personally I use a simple
batch program:

==startaw.bat==
at echo off
del aworld.log
start aworld.exe
==end-of-file==

Just make your shortcuts to AW point to the startaw.bat file instead of the
aworld.exe (must be in the main AW directory or it won't work).

KAH

May 6, 2003, 12:00pm

the log is deleted automaticly lol :)
-SWE

[View Quote]

May 6, 2003, 1:08pm

"swe" <swe at swe-e.com> wrote in news:3eb7c001 at server1.Activeworlds.com:

> the log is deleted automaticly lol :)

Not by AW.

KAH

May 6, 2003, 1:57pm

ya it is.

[View Quote]

May 6, 2003, 2:28pm

Yes it is. everytime you restart, so does the log.

[View Quote]

May 6, 2003, 4:23pm

Have you been able to reproduce this since? Has it happened in any other
worlds? Can you post the last 50 lines or so from the log? PLEASE do not
cross post. If you have the ability to post in the Beta newsgroup still then
you should post it to this group ONLY. If you cannot post to the beta
newsgroup then please send a bug report via email to
support at activeworlds.com with the word "Beta" in the subject.

Bill

[View Quote]

May 6, 2003, 5:50pm

oh, that was new to me, my logs keep holds it, at the end it gets 14 GB big,
and a pain to get away.
hehe

[View Quote]

May 7, 2003, 12:31pm

[View Quote] Oh, I thought it should be cleared when AW is restarted?
"Before, if the "flush" option was enabled, the log file would be appended
each time, thus growing bigger and bigger. Now it will start with a blank
log each time."

--
Mauz
http://mauz.info

May 7, 2003, 6:17pm

"mauz" <mauz at my.activeworlds.com> wrote in
news:3eb918b3 at server1.Activeworlds.com:

> Oh, I thought it should be cleared when AW is restarted?
> "Before, if the "flush" option was enabled, the log file would be
> appended each time, thus growing bigger and bigger. Now it will start
> with a blank log each time."

In any case, I have to delete it myself every time because it's not purged
at every start.

KAH

Jul 7, 2003, 7:20pm

About a month ago, a world I am CT in was having repeated crashes.

I tracked it down to builders linking to to images on th groups.msn.com site
and using **extremely** long URLS of the form:

http://groups.msn.com?big_long_string_of_indecipherable_gobbledygook

The nature of the crashes was such that once one of these 'poison' URLS
was in the world, you could not even run the browser anymore without
crashing again.

Trial and error investigation determined that deleting just the cached files
with the long ugly names would allow the browser to run long enough to
go into the world and delete the URL from the object in question. Of course
going back into the world would recreate the long filename again, so in
order to delete the offending object you had to suffer through another
crash.

The browser only seemed to crash when you left the world, so you might
enter the world and happily work in it for hours. When you went to leave
though your browser would crash and from then on you could not run it
without deleting your cache.

I don't know if it has anything to do with your current issue, but it's
something to bear in mind. It may even be fixed with 3.4, but I'm not
that eager to test it out.

It's also unfortunatly a way for troublemakers to cause more trouble,
but at least you know what to look for if you get crashes... very, very
long filenames in the cache folders.

Generally the long names will end up in the 'misc' folder BTW

[View Quote] <SNIP>

> we both crashed at the same
> time, mine was that damned aw_3d3.dll again, his was aw_3d8.dll that
> crashed. How can two people in different parts of the "REAL world" crash
at
> the same time?

Jul 8, 2003, 1:33am

Actually, when you have identified such URLs, the best way is to delete
them from a propdump and then reload the cleaned propdump to the world.



[View Quote]

1  |   Order by index