167593 - error page crashes when middle-click opening "Try again" in background tab

Status: RESOLVED DUPLICATE of bug 157102

(Plugins Graveyard :: Java (Sun), defect)

Description

[haferfrost]

User-Agent:       Mozilla/5.0 (X11; U; Linux i586; en-US; rv:1.1) Gecko/20020830
Build Identifier: Mozilla/5.0 (X11; U; Linux i586; en-US; rv:1.1) Gecko/20020830

When error pages (see bug 28586) are enabled and I middle click (set to open the
link in a tab in the background) the "Try again" link on the error page, Mozilla
crashes with the following messages on the console:

INTERNAL ERROR on Browser End: Expected a version > 5! Version = 0

System error?:: No such file or directory



Reproducible: Always

Steps to Reproduce:
1.configure tab browsing, middle-click, open links in background (may not be
necessary but that's what I use), activate error pages pref
2.enter http://www.sjdfsjkfjasbdajshb.org  in address bar
3.on error page middle-click "Try again"

Actual Results:  
Zap!

Expected Results:  
Same error page loaded in background in a new tab.

Comment 1

[Matthias Versen [:Matti]]

You changed Mozilla UA string and you use Java
The error is from Java : Expected a version.... because Java gets also the wrong
UA and the calls exit().
Since Mozilla runs in the same process it crashes

*** This bug has been marked as a duplicate of 83376 ***

Comment 2

[Olivier Cahagne]

From the error, Java is guilty. There're similar Java issues (where Java is
totally unrelated) with mail attachments without extension, can't find the bug
for now.
I also have various issues involving Java with this on both Linux & Windows.

Comment 3

[haferfrost]

This is not a duplicate of bug 83376. I have Java disabled all the time (the
plugin is present, though) and I can start Mozilla just fine. There is no reason
why Java should suddendly get activated by clicking a link. But maybe this is a
separate bug (looks like a bad security issue to me). 
Anyway, if I change the user agent string back to normal, the bug persists, but
the message on the console changes to:

INTERNAL ERROR on Browser End: [0] Initialize. No docbase?
System error?:: Resource temporarily unavailable

Removing the Java plugin fixes the problem, although I see no good reason why
,because as I've said, I have Java disabled all the time (and JavaScript, too, btw).

Comment 4

[Olivier Cahagne]

related: bug 153930.
even more related: 157102.

Comment 5

[Olivier Cahagne]

-> docshell

Comment 6

[Olivier Cahagne]

*** This bug has been marked as a duplicate of 157102 ***

Comment 7

[haferfrost]

Come on guys. I know you want to get the bug count down, but marking as
duplicate without verification is not a solution. This bug still exists in
Mozilla/5.0 (X11; U; Linux i586; en-US; rv:1.3a) Gecko/20021117 
so it is obviously not a duplicate of bug 157102 which according to its Bugzilla
page has been fixed 3 months ago (before I even reported this bug here). I'm
reopening this bug (again).

Comment 8

[Olivier Cahagne]

Then we should reopen bug 157102, it's exactly the same report. Or comment
within bug 157004.

Comment 9

[haferfrost]

How is this the same report? The reported error message for bug 157102 is (see
comment #4 there)

INTERNAL ERROR on Browser End: Exec of "java_vm" failed: 2<
System error?:: No such file or directory
INTERNAL ERROR on Browser End: Could not read ack from browser
System error?:: Resource temporarily unavailable

Why is this the same as the "no docbase" error message for this bug here which
you can see in comment #3?

Comment 10

[Olivier Cahagne]

Marking NEW to get attention.
Adam, can you comment on this one ?

Comment 11

[haferfrost]

More observations on this bug: Before the crash, there are several seconds of
intense activity. On my K6-2/500 this is more than enough time to launch a
script that SIGSTOPs Mozilla. After that, it is easy to attach a debugger to
Mozilla. Unfortunately my build does not have debug information so all I could
get from gdb was "Program received signal SIG32, Real-time event 32." I'll see
if I get around to building a Mozilla with debug info so that I can provide a
stack trace. The key probably lies in finding out why on earth Mozilla wants to
launch the java_vm (I even have java disabled; not that it would make sense to
launch java when middle-clicking Try again in the first place).

Comment 12

[Andrew Hagen]

Is there a Talkback ID for this crash?

Comment 13

[Olivier Cahagne]

Andrew, no, because the crash is happening within Java plug-in.

Comment 14

[Matthias Versen [:Matti]]

Is this still an issue with recent builds ?

Comment 15

[timeless]

this is a duplicate, and this bug was kept alive way way way way too long because the reporter did not trust our triage team.

reporter: if you try to push this bug back to core, your account will be terminated.

https://bugzilla.mozilla.org/page.cgi?id=etiquette.html

2.2. Changing Fields # No whining about decisions.
If a respected project contributor has marked a bug as INVALID, then it is invalid. Someone filing another duplicate of it does not change this. Unless you have further important evidence, do not post a comment arguing that an INVALID or WONTFIX bug should be reopened. 

The same rule applies to the DUPLICATE resolution.