517797 - crash from application/pdf object with no value for data attribute, or syntax error [@ nppdf.so@0x13ae7]

Status: VERIFIED INVALID

(Plugins Graveyard :: PDF (Adobe), defect)

Description

[reisio]

User-Agent:       Mozilla/5.0 (X11; U; Linux x86_64; en-US; rv:1.9.1.3) Gecko/20090913 Gentoo Firefox/3.5.3
Build Identifier: 

object element with type="application/pdf" and data attribute with no value (=""), or missing equal sign (data"...) crashes Firefox with Adobe PDF plugin (tested with: Windows Vista + Adobe Reader 9, Windows XP + Acrobat Pro, GNU/Linux + acroread).

Reproducible: Always

Steps to Reproduce:
1. Visit http://reisio.com/temp/object-pdf-data-crash_1of2.html or http://reisio.com/temp/object-pdf-data-crash_2of2.html in Firefox with Adobe PDF reader plugin support
Actual Results:  
Browser crashes.

Expected Results:  
Not crash / not pass to Adobe plugin.

bp-d0d6b1af-e13b-4068-8805-8d4432090920

Comment 1

[reisio]

Attachment: code for crash, missing equal sign

Comment 2

[reisio]

Attachment: code for crash, no value given

Comment 3

[timeless]

Sorry, your expected results aren't quite reasonable. this is a bug in whomever provided your pdf plugin (please get that information from about:plugins and paste it here). We are working on out of process plugins which would prevent them from crashing us, but that shouldn't be the subject of your bug. If possible, you should report this crash to your PDF plugin vendor (it could be adobe or someone else, i can't tell given the information provided).

Signature	nppdf.so@0x13ae7
UUID	d0d6b1af-e13b-4068-8805-8d4432090920
Time 	2009-09-20 14:33:11.671563
Uptime	3
Last Crash	54 seconds before submission
Product	Firefox
Version	3.5.3
Build ID	20090824085743
Branch	1.9.1
OS	Linux
OS Version	0.0.0 Linux 2.6.31-gentoo-m5-kms-no-oprofile #12 Mon Sep 14 21:58:29 CEST 2009 i686 GNU/Linux
CPU	x86
CPU Info	AuthenticAMD family 1 model 8 stepping 1
Crash Reason	SIGSEGV
Crash Address	0xaeae7ae7
User Comments	another acroread-related crash
Processor Notes 	
Crashing Thread
Frame 	Module 	Signature [Expand] 	Source
0 	nppdf.so 	nppdf.so@0x13ae7 	
1 	nppdf.so 	nppdf.so@0x141c5 	
2 	nppdf.so 	nppdf.so@0x203c8 	
3 	libxul.so 	nsNPAPIPluginInstance::InitializePlugin 	modules/plugin/base/src/nsNPAPIPluginInstance.cpp:1030
4 	libxul.so 	nsNPAPIPluginInstance::Initialize 	modules/plugin/base/src/nsNPAPIPluginInstance.cpp:808
5 	libxul.so 	nsPluginHostImpl::TrySetUpPluginInstance 	modules/plugin/base/src/nsPluginHostImpl.cpp:3872
6 	libxul.so 	nsPluginHostImpl::SetUpPluginInstance 	modules/plugin/base/src/nsPluginHostImpl.cpp:3670
7 	libxul.so 	nsPluginHostImpl::InstantiateEmbeddedPlugin 	modules/plugin/base/src/nsPluginHostImpl.cpp:3361
8 	libxul.so 	nsPluginStreamListenerPeer::OnStartRequest 	modules/plugin/base/src/nsPluginHostImpl.cpp:2025
9 	libxul.so 	nsObjectLoadingContent::OnStartRequest 	content/base/src/nsObjectLoadingContent.cpp:608

nppdf.so  	 	61B1BE1FF7D02A9240AA381A0D8C198C0  	nppdf.so

Comment 4

[reisio]

> paste it here). We are working on out of process plugins which would prevent
> them from crashing us, but that shouldn't be the subject of your bug. If

That covers it, then.

Comment 5

[timeless]

for bookkeeping purposes, this bug is invalid, it's a bug in a third party plugin, if someone wants to find the vendor, and file a report including a link to this bug, they're free to do so. if the vendor wants to use our bugzilla to keep our common customers informed, the vendor is welcome to ask for ownership to be assigned for this (or similar) bugs. in the end such a bug would be resolved as worksforme when the vendor announces a plugin version which fixes the problem.

otherwise,... as the bug isn't in our codebase, it's invalid.

Comment 6

[reisio]

I agree this is most immediately an Adobe problem, but if "We are working on out of process plugins which would prevent them from crashing us", and that's valid, then this bug is also valid (though happily redundant).

Comment 7

[timeless]

well. i'm not interested in finding that bug and spamming your bug into it, and i doubt you really want to follow or would understand the progress of that bug. there will be some hype when some future released version of firefox includes support for running plugins out of process, i doubt you'll miss it.