SeaMonkey stops responding on SMTP send + certain SSL/TLS sites

RESOLVED WORKSFORME

Status

RESOLVED WORKSFORME
7 years ago
7 years ago

People

(Reporter: slasha..........3+mozbug, Unassigned)

Tracking

SeaMonkey 2.3 Branch
x86
All

Firefox Tracking Flags

(Not tracked)

Details

Attachments

(2 attachments)

(Reporter)

Description

7 years ago
User Agent: Mozilla/5.0 (Windows NT 6.0; rv:8.0a2) Gecko/20110903 SeaMonkey/2.5a2
Build ID: 20110903013026

Steps to reproduce:

I am using Windows Vista.

Send an email using SMTP and browse certain SSL/TLS sites at the same time.

Websites that cause SeaMonkey to stop responding:
https://www.apache.org/
https://blekko.com/
https://www.grc.com/
https://github.com/
https://drupal.org/

Websites that do NOT cause SeaMonkey to stop responding:
https://encrypted.google.com/
https://secure.wikimedia.org/
https://www.youtube.com/
https://www.mozilla.org/

My SMTP settings are as follows:
Server name: pod51003.outlook.com
Port: 25
Connection security: STARTTLS
Authentication method: Normal password
User Name: myname@example.com

I can confirm that this behaviour is not recent because I had the same problems with 2.1, 2.2, 2.3 (I am now using 2.5 aurora).

I originally posted this on MozillaZine before here: http://forums.mozillazine.org/viewtopic.php?f=5&t=2295281


Actual results:

SeaMonkey stopped responding. Looking at the network traffic through Windows Task Manager revealed that when the program stopped responding the network traffic stopped altogether.
Please try running with disabled addons.
We need a stacktrace with debug symbols but i don't know the symbol server address for SM...
(Reporter)

Comment 2

7 years ago
I tried running with addons disabled and I still got the same problem. On a side note, ssl.gstatic.com also causes is to stop responding.

I also don't get a crash reporting window either; it just hangs until I end it no matter how long I leave it for.
(Reporter)

Comment 3

7 years ago
Actually, forget what I said about ssl.gstatic.com. There was something on https://encrypted.google.com/ that caused it to crash once, but it mustn't have been cached. Visiting https://ssl.gstatic.com/ does not cause it to crash during send.
(Reporter)

Comment 4

7 years ago
Created attachment 558362 [details]
debug file

I did a stacktrace using WinDbg (attached) using the instructions from
http://developer.mozilla.org/en/How_to_get_a_stacktrace_with_WinDbg

This was done on a new profile and the same results occurred.
There are a few more instructions that you missed under "After the crash or hang"
(Reporter)

Comment 6

7 years ago
Created attachment 558728 [details]
debug file (2)

^ see attached

I have done another debug using those instructions. Sorry, I didn't realise I'd missed those, thanks for letting me know. :-)
Is that a deadlock (0% CPU) or a loop (100% CPU for one CPI core) ?
(Reporter)

Comment 8

7 years ago
It is a deadlock.
(Reporter)

Comment 9

7 years ago
Just an update: somebody on the topic at MozillaZine has confirmed that the bug can be reproduced on Apple OSX.
http://forums.mozillazine.org/viewtopic.php?f=5&t=2295281

Also, I can confirm that it crashes on Debian GNU/Linux. So the bug is cross-platform.


In addition, it also appears to crash on ALL SSL websites now, including https://encrypted.google.com/, YouTube, Mozilla, Wikimedia.


I only just realised one detail that I forgot to add in the original post:
Sending mail using my other provider (ISP, no SSL) works fine, no crash. That was the reason why I suspected it was due to the SSL being used in multiple places, or something up with the implementation.

Updated

7 years ago
OS: Windows Vista → All

Comment 10

7 years ago
Is this still reproducible with 2.8?
(Reporter)

Comment 11

7 years ago
No, I could not reproduce this in SeaMonkey 2.8. I've marked the status as resolved. Thanks for everyone's help anyway. :)
Status: UNCONFIRMED → RESOLVED
Last Resolved: 7 years ago
Resolution: --- → INVALID
Resolution: INVALID → WORKSFORME
You need to log in before you can comment on or make changes to this bug.