175302 - some sort of form thing in OOM situation crashed

Status: VERIFIED DUPLICATE of bug 184202

(Firefox :: Address Bar, defect)

Description

[timeless]

Mozilla/5.0 (Windows; U; Win98; en-US; rv:1.2b) Gecko/20021014 Phoenix/0.3

PHOENIX caused an invalid page fault in
module TKITCMPS.DLL at 016f:0177b502.
Registers:
EAX=00000000 CS=016f EIP=0177b502 EFLGS=00010206
EBX=00000000 SS=0177 ESP=0065fb8c EBP=0065fc58
ECX=0177daf0 DS=0177 ESI=02221e14 FS=21c7
EDX=0065fbb4 ES=0177 EDI=0065fbb4 GS=0000
Bytes at CS:EIP:
8b 08 ff 91 24 01 00 00 8b 07 8b cf ff 50 14 85
Stack dump:
00000000 0065fbb4 100491ef 02221dc0 01771dc8 02221e14 0065fbb4 00869330 02221dc8
01950d30 1004c798 0065fbc4 00000000 8000003f 00000000 00000001

--
I might have been out of memory, EAX=0 usually means dereferencing a null pointer.

dbradley: could you possibly convert this into a useful line number?
(unfortunately I don't think this library is built for seamonkey, but i doubt it
has changed much since phoenix0.3 was released -- this is the official 0.3 win32
release, installed in a fresh directory.)

If not, invalidate it.

Comment 1

[David Bradley]

I'll have to get a phoenix build. I tried pulling down toolkit and building but
no go. Don't count on this happening anytime soon. Maybe later next week.

Comment 2

[Brendan Eich [:brendan]]

Get a stack, eh?

/be

Comment 3

[Asa Dotzler [:asa]]

Unable to reproduce with win2K or RH8.

Comment 4

[timeless]

brendan: no fair, i provided a stack, what dbradley has done in the past is
reconstruct the data based on the information you see there.
(here's another)
PHOENIX caused an invalid page fault in
module TKITCMPS.DLL at 016f:022db502.
Registers:
EAX=00000000 CS=016f EIP=022db502 EFLGS=00010206
EBX=00000000 SS=0177 ESP=0065fb8c EBP=0065fc58
ECX=022ddaf0 DS=0177 ESI=02315bd4 FS=2ecf
EDX=0065fbb4 ES=0177 EDI=0065fbb4 GS=0000
Bytes at CS:EIP:
8b 08 ff 91 24 01 00 00 8b 07 8b cf ff 50 14 85
Stack dump:
00000000 0065fbb4 100491ef 02315b80 022d1dc8 02315bd4 0065fbb4 02f91020 02315b88
01950d30 1004c798 0065fbc4 00000000
8000003f 00000000 00000001

Asa, steps to reproduce: run phoenix on a system with very little memory (32mb)
and very little free disk space, then grow phoenix until the os runs out of disk
space and then phoenix will crash when it fails a malloc.

Note that I wouldn't need to use these useless stacks if phoenix had talkback...

Comment 5

[timeless]

brendan: no fair, i provided a stack. Note that I wouldn't need to use these
hard to use stacks if phoenix had talkback.... What dbradley has done in the
past is reconstruct the data based on the information you see there.
(here's another)
PHOENIX caused an invalid page fault in
module TKITCMPS.DLL at 016f:022db502.
Registers:
EAX=00000000 CS=016f EIP=022db502 EFLGS=00010206
EBX=00000000 SS=0177 ESP=0065fb8c EBP=0065fc58
ECX=022ddaf0 DS=0177 ESI=02315bd4 FS=2ecf
EDX=0065fbb4 ES=0177 EDI=0065fbb4 GS=0000
Bytes at CS:EIP:
8b 08 ff 91 24 01 00 00 8b 07 8b cf ff 50 14 85
Stack dump:
00000000 0065fbb4 100491ef 02315b80 022d1dc8 02315bd4 0065fbb4 02f91020 02315b88
01950d30 1004c798 0065fbc4 00000000
8000003f 00000000 00000001

Asa, steps to reproduce: run phoenix on a (windows) system with very little
memory (32mb) and a small swap file on a disk with very little free space. grow
phoenix (by browsing) until the os runs out of disk space and then phoenix will
crash when it fails a malloc.

more details based on my most recent crash: i was browsing a bug (jgaunt's
reinvent the world for shared profiles) and some of the links from it (well not
really, the link was bad so i entered it manually). then i went back to the bug
and started typing, this action presumably opened autocomplete and because the
os was out of memory (all those big docs) phoenix crashed in some unchecked
allocation of hewitt's (yes, it's not fair of me to blame hewitt, but I have
case history).

Comment 6

[Asa Dotzler [:asa]]

"we don't work on systems with 32MB RAM and any browsing you can accomplish on
such a system is pure luck" sounds like a reasonable resolution to me.  

Comment 7

[Erik]

I wonder, if the problem still occurs on >=64MB systems that run out of memory
for whatever reasons...? If not, we should resolve this bug as invalid to get
rid of it from the list of unconfirmed bugs...

btw: perhaps this problem is related to bug #184202?

Comment 8

[timeless]

*** This bug has been marked as a duplicate of 184202 ***

Comment 9

[Simon Paquet [:sipaq]]

Mass verifiying of old duplicates.