Last Comment Bug 319219 - ChatZilla appears connected but nothing sends (commands or messages) - 100% CPU usage
: ChatZilla appears connected but nothing sends (commands or messages) - 100% C...
Status: RESOLVED FIXED
Caused by bugs listed in "depends on"...
: hang
Product: Other Applications
Classification: Client Software
Component: ChatZilla (show other bugs)
: Trunk
: x86 All
: -- critical (vote)
: ---
Assigned To: Robert Ginda
: Samuel Sieb
:
Mentors:
: 320160 320267 (view as bug list)
Depends on: 318419
Blocks:
  Show dependency treegraph
 
Reported: 2005-12-05 14:42 PST by Wayne Mery (:wsmwk, NI for questions)
Modified: 2006-05-20 19:17 PDT (History)
7 users (show)
See Also:
Crash Signature:
(edit)
QA Whiteboard:
Iteration: ---
Points: ---


Attachments
[checked in] Catch bug 291386 in the act and explain to the user (4.18 KB, patch)
2005-12-06 09:11 PST, James Ross
samuel: review+
gijskruitbosch+bugs: review+
Details | Diff | Splinter Review
workaround (7.37 KB, patch)
2006-01-01 23:25 PST, Olli Pettay [:smaug]
no flags Details | Diff | Splinter Review

Description Wayne Mery (:wsmwk, NI for questions) 2005-12-05 14:42:12 PST
Chatzilla 0.9.67+ [SeaMonkey 1.1a/2005102013]

Channel appears to be operational - messages are received, and messages typed in input area are put in the displayed in the conversation log, and therefore appear to have been sent to the channel.   (open moznet channels were: seamonkey, chatzilla, developers, bugs)

But 
a) no one else in the channel sees my messages and 
b) no commands work (/join, /leave, /motd, /list, etc) - they are accepted without error and there is no response
c) private chat requests don't connect

My nickname=wsm. Possibly related events:
* different machine running IRC with similar nickname wsmwk crashed earlier in the day
* jsconsole entry, don't know if this occured at same time as my trouble started, or perhaps long before --  Error: uncaught exception: [Exception... "Component returned failure code: 0x80004005 (NS_ERROR_FAILURE) [nsIControllers.removeController]"  nsresult: "0x80004005 (NS_ERROR_FAILURE)"  location: "JS frame :: chrome://navigator/content/navigator.js :: Shutdown :: line 808" data: no]  (similar to bug 2438893)
Comment 1 James Ross 2005-12-05 14:52:50 PST
I wouldn't be surprised if this is just the Mozilla timers screwing up (which they've been doing rather too frequently of late). Check this:

  /eval this.primServ.sendQueue.length

On any connected server. If it's >100 you have got the timer problem, which we can't do anything about, and you really should restart Mozilla or Firefox to fix.

And PLEASE don't file bugs like this in the NEW state, it is definately UNCO for now.
Comment 2 Wayne Mery (:wsmwk, NI for questions) 2005-12-05 15:04:47 PST
(In reply to comment #1)
> I wouldn't be surprised if this is just the Mozilla timers screwing up (which
> they've been doing rather too frequently of late). Check this:
> 
>   /eval this.primServ.sendQueue.length
> 
> On any connected server. If it's >100 you have got the timer problem, which we
> can't do anything about, and you really should restart Mozilla or Firefox to
> fix.

will be able to test this in the future I'm sure - as I've had this problem multiple times.  what should be "normal" or nominal value?
 

> And PLEASE don't file bugs like this in the NEW state, it is definately UNCO
> for now.

agree, realized my mistake as soon as I clicked commit.
Comment 3 Felix Miata 2005-12-05 17:18:31 PST
For several weeks this happened to me in the OS/2 trunk, but the problem seems to have disappeared recently.
Comment 4 Felix Miata 2005-12-05 21:01:37 PST
Void comment 3. It just did it again in 2005120512 OS/2 trunk.
Comment 5 Wayne Mery (:wsmwk, NI for questions) 2005-12-06 08:49:18 PST
Bug 291386 is implicated
Comment 6 James Ross 2005-12-06 09:11:21 PST
Created attachment 205147 [details] [diff] [review]
[checked in] Catch bug 291386 in the act and explain to the user

This patch will alert the user when the queue reaches 1000 items (very unlikely normally), and then with an ERROR at 10,000 and every 1000 after that.
Comment 7 Wayne Mery (:wsmwk, NI for questions) 2005-12-06 09:20:54 PST
mod up to OS=ALL (known to happen on OS/2 and windows, no report yet of linux or mac)

keyword=hang (chatzilla is essentially hung)
Comment 8 :Gijs Kruitbosch 2005-12-06 10:21:23 PST
Comment on attachment 205147 [details] [diff] [review]
[checked in] Catch bug 291386 in the act and explain to the user

Yes, please :)
Comment 9 :Gijs Kruitbosch 2005-12-13 13:08:26 PST
*** Bug 320160 has been marked as a duplicate of this bug. ***
Comment 10 James Ross 2005-12-14 09:08:01 PST
*** Bug 320160 has been marked as a duplicate of this bug. ***
Comment 11 :Gijs Kruitbosch 2005-12-14 09:58:31 PST
*** Bug 320267 has been marked as a duplicate of this bug. ***
Comment 12 Olli Pettay [:smaug] 2006-01-01 23:25:52 PST
Created attachment 207327 [details] [diff] [review]
workaround

With this workaround I haven't seen the Excess Flood problems anymore.
*ErrorCounters are there just for debugging. (And I can see that the
counters have been increased quite much during the weekend while chatzilla was running.)
Comment 13 :Gijs Kruitbosch 2006-02-16 03:36:43 PST
I think this should be fixed with trunk ChatZilla on trunk Firefox/SeaMonkey/XULRunner now?

Per:
https://bugzilla.mozilla.org/show_bug.cgi?id=318419#c23 (which I am assuming is about bug 307527, but I'm not sure about this).

There is also work done on bug 318419, but no checkins there yet.

Can someone confirm whether this problem has gone away?
Comment 14 Wayne Mery (:wsmwk, NI for questions) 2006-02-16 05:25:12 PST
I have not upgraded - waiting for these guys to settle down with their patch ideas.  But I can test it over the next few days.  

Does rebooting the PC help induce the wrap problem?
Comment 15 James Ross 2006-02-16 05:50:22 PST
(In reply to comment #14)
> Does rebooting the PC help induce the wrap problem?

No, it resets it, in fact. The problem occurs at multiples of the wrap point since the start of Windows.
Comment 16 James Ross 2006-03-25 12:41:50 PST
Ok, since bug 318419 is fixed and a release made with it in (Firefox 1.5.0.1) I propose that this bug is FIXED.

Note You need to log in before you can comment on or make changes to this bug.