79782 - Crash in nsHTTPPipelinedRequest::OnStopRequest

Status: VERIFIED INVALID

(Core :: Networking: HTTP, defect)

Description

[Phil Peterson]

Using 2001-05-08-12 commercial release build on NT

1. Start the app with no network connection
2. Once the app is up, it tries to load my home page
3. Since there's no network connection, I close the app
4. boom

I see this all the time when fighting my flaky PPPoE connection over DSL. 

Stack Trace
0x00000000
nsHTTPHandler::ReleaseTransport
[d:\builds\seamonkey\mozilla\netwerk\protocol\http\src\nsHTTPHandler.cpp, line 1424]
nsHTTPPipelinedRequest::OnStopRequest
[d:\builds\seamonkey\mozilla\netwerk\protocol\http\src\nsHTTPRequest.cpp, line 874]
nsSimpleStreamProvider::OnStopRequest
[d:\builds\seamonkey\mozilla\netwerk\base\src\nsSimpleStreamProvider.cpp, line 57]
nsOnStopRequestEvent::HandleEvent
[d:\builds\seamonkey\mozilla\netwerk\base\src\nsRequestObserverProxy.cpp, line 159]
PL_HandleEvent [d:\builds\seamonkey\mozilla\xpcom\threads\plevent.c, line 589]
PL_ProcessPendingEvents [d:\builds\seamonkey\mozilla\xpcom\threads\plevent.c,
line 522]
nsEventQueueImpl::ProcessPendingEvents
[d:\builds\seamonkey\mozilla\xpcom\threads\nsEventQueue.cpp, line 375]
0x00813a90
0x0088ce1c

Comment 1

[Gagan]

gordon should look into this after/as he finishes implementing the dns cache.

Comment 2

[gordon]

I can image there might be problems with quiting the app while there are 
outstanding DNS lookups (though I don't have any evidence), but the stack trace 
here doesn't show anything to do with DNS.

Do we have some reason to believe this has something to do with the nsDNSService 
other than the faulty status display?

Gagan, do you want me to look at this because  you think it has something to do 
with DNS, or are you just load balancing HTTP bugs?

Comment 3

[Gagan]

former

Comment 4

[gordon]

Gagan, this doesn't look like DNS.

Comment 5

[benc]

Are you saying this happens: 

You start you computer, fire up mozilla and it dies?

-OR-

You start your computer, fire up PPPOE, and it doesn't go green, but you fire up
mozilla anyhow and it crashes?

The PPPOE client I use also hooks up DHCP and configures DNS, so if you don't
have a valid connection (my corporately correct test is connecting to AOL IM via
standalone AIM...). If your PPPEO is not coming up succesfully and you start
mozilla, I don't think it is obvious where mozilla would when it crashes.

Comment 6

[Phil Peterson]

I have a LinkSys router/hub which does PPPoE in hardware, and acts as a DNS
server forwarding client requests to the DNS servers it got from PPPoE. So the
steps are these:

1. Plug laptop into home network
2. Boot the laptop
3. Boot the app, expecting to use the PPPoE and DNS connections maintained by
the router/hub
4. Sit for a while in "looking up host..."
5. Realize that the LinkSys box isn't happy for some unknown reason.
6. Exit the app
7. Boom

> If your PPPEO is not coming up succesfully and you start
> mozilla, I don't think it is obvious where mozilla would when it crashes.

Sorry, I can't parse that.

Comment 7

[Gagan]

It would help if someone can attach the crash stack... adding helpwanted.

Comment 8

[Gagan]

oops wrong comments in the wrong bug.

Comment 9

[Gagan]

->neeti

Comment 10

[Auke]

Sometimes when I quit Mozilla 0.9 (Windows NT sp6a) by double clicking the
mozilla icon on the left of the title bar, the mozilla window dissapears but the
process called mozilla remains active (It can be seen in the task manager under
processes; it has dissapeared under applications) and it uses 99% of the
processing power. Perhaps this occurs during DNS lookup; but I don't know how to
figure that out. 

Auke

Comment 11

[Manoj]

I have encountered the same problem as has Auke... Mozilla's DNS lookup takes
very long at times, instances to note are:

www.cs.cmu.edu
www.ebank.hsbc.com
www.citibankcards.com

I will keep adding to this list as time goes on. I have IE6.0 also installed on
my computer and keep checking the load times for a page with Mozilla and IE and
there are times when IE just outperforms Mozilla by miles... 

Comment 12

[Gagan]

->gordon

Comment 13

[chris hofmann]

shiva, does this match up with any top crash data?

Comment 14

[gordon]

Okay stop it!  This bug is morphing.  The original summary was just plain wrong,
so I'm changing it.  Besides, Phil left the company, so I don't think we have to
fix this anymore...

Okay, okay...geez, I was only kidding.

This bug has nothing to do with DNS crashing or being slow.  Phil only *thought*
it was related to DNS because of bug 112336, which darin will be fixing shortly.

chofmann's question is on the money. A lot has changed since last May.  Is
anybody seeing similar crashes anymore?

Comment 15

[Darin Fisher]

this code doesn't exist anymore.  marking INVALID

Comment 16

[Shiva Thirumazhusai]

FYI : This crash only happened in Netscape 6.00 and Netscape 6.01.

Comment 17

[John Unruh]

Verified. I don't see a crash when starting mozilla before having a network
connection.