260906 - crash w/https proxy that returns unexpected data

Status: VERIFIED DUPLICATE of bug 242393

(Core :: Networking, defect)

Description

[dean gaudet]

User-Agent:       Mozilla/5.0 (Windows; U; Windows NT 5.1; rv:1.7.3) Gecko/20040913 Firefox/0.10
Build Identifier: Mozilla/5.0 (Windows; U; Windows NT 5.1; rv:1.7.3) Gecko/20040913 Firefox/0.10

i've got privoxy installed (http://privoxy.org) and my configuration directs
http *and* https requests through privoxy (localhost:8118) ... when you visit
paypal.com and it references https://paypalssl.doubleclick.net/ that goes
through privoxy as a CONNECT request and privoxy denies it...

however privoxy's denial is plaintext ... it's a blank gif in my case -- it's a
plaintext response without the ssl negotiation.

but firefox crashes when given this response.

maybe you can reproduce by doing https://foo:80/ ?  i'd try right now but i
don't want to crash this firefox i'm typing into :)

-dean

Reproducible: Always
Steps to Reproduce:
1. install privoxy
2. configure firefox to proxy http and https through localhost:8118
3. visit paypal.com

Actual Results:  
firefox crashes

Expected Results:  
it shouldn't have crashed :)

Comment 1

[dean gaudet]

Attachment: proxy response

the attachment is the response from privoxy for the "CONNECT
paypalssl.doubleclick.net:443 HTTP/1.0" request ... maybe it will help.

Comment 2

[Joey Stanford]

I am having the same issue.  It impacts me on XP, Linux (SuSE), and Mac OS X.

Reference support thread: http://forums.mozillazine.org/viewtopic.php?t=132742

Comment 3

[Michael Blakeley]

I also see this problem with Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.7)
Gecko/20040917 Firefox/0.9.3. Workaround: edit Preferences so that firefox
doesn't use privoxy for SSL.

In case it's helpful, here's the stack trace from debian mozilla-firefox,
version 0.9.3-5 (testing).

Program received signal SIGSEGV, Segmentation fault.
[Switching to Thread 1075426240 (LWP 25165)]
0x40515009 in NSGetModule ()
   from /usr/lib/mozilla-firefox/components/libnecko.so
(gdb) bt
#0  0x40515009 in NSGetModule ()
   from /usr/lib/mozilla-firefox/components/libnecko.so
#1  0x40486fff in NSGetModule ()
   from /usr/lib/mozilla-firefox/components/libnecko.so
#2  0x40486e3f in NSGetModule ()
   from /usr/lib/mozilla-firefox/components/libnecko.so
#3  0x400f5d31 in nsInputStreamReadyEvent::EventHandler ()
   from /usr/lib/mozilla-firefox/libxpcom.so
#4  0x4010c397 in PL_HandleEvent () from /usr/lib/mozilla-firefox/libxpcom.so
#5  0x4010c2c4 in PL_ProcessPendingEvents ()
   from /usr/lib/mozilla-firefox/libxpcom.so
#6  0x4010df29 in nsEventQueueImpl::NotifyObservers ()
   from /usr/lib/mozilla-firefox/libxpcom.so
#7  0x4040a8d5 in nsBaseWidget::FreeNativeData ()
   from /usr/lib/mozilla-firefox/components/libwidget_gtk2.so
#8  0x4136f1df in g_vasprintf () from /usr/lib/libglib-2.0.so.0
#9  0x41349b92 in g_main_depth () from /usr/lib/libglib-2.0.so.0
#10 0x4134ac88 in g_main_context_dispatch () from /usr/lib/libglib-2.0.so.0
#11 0x4134afc0 in g_main_context_dispatch () from /usr/lib/libglib-2.0.so.0
#12 0x4134b603 in g_main_loop_run () from /usr/lib/libglib-2.0.so.0
#13 0x418165a3 in gtk_main () from /usr/lib/libgtk-x11-2.0.so.0
#14 0x4040ad06 in nsAppShell::ReleaseGlobals ()
   from /usr/lib/mozilla-firefox/components/libwidget_gtk2.so
#15 0x40356d64 in ?? ()
   from /usr/lib/mozilla-firefox/components/libnsappshell.so
#16 0x080b96f0 in ?? ()
#17 0x0805900c in data_start ()
#18 0xbffff968 in ?? ()
#19 0x080516e2 in xre_main ()
Previous frame identical to this frame (corrupt stack?)

Comment 4

[timeless]

[e-mail]: debian did you the favor of stripping symbols from
your build. please thank debian for its help and get a standard mozilla.org
build with talkback, or build mozilla w/o stripping symbols.

Comment 5

[Jeffrey Baker]

Timeless's response is rude and unhelpful.  CONFIRMED.

Comment 6

[:Gavin Sharp [email: [e-mail]]]

In an ideal world, bugs would be tested on a recent trunk build and moved to the
correct component before being confirmed. Jeffrey, if it was tested on a recent
build, please indicate that you did and provided the build ID in a comment.

Comment 7

[Jeffrey Baker]

You hit me midair :)  This was confirmed on the build contemporary with the
original report, but WFM on current Linux release and nightly.  Marking WFM.

Comment 8

[Christian :Biesinger (don't email me, ping me on IRC)]

*** This bug has been marked as a duplicate of 242393 ***

Comment 9

[Jeffrey Baker]

Verified.  Thanks Christian.