1243326 - Java 8 and firefox 42+

Status: RESOLVED WONTFIX

(Core Graveyard :: Plug-ins, defect)

Description

[[e-mail]]

User Agent: Mozilla/5.0 (Windows NT 10.0; WOW64; rv:41.0) Gecko/20100101 Firefox/41.0
Build ID: 20150910171927

Steps to reproduce:

Loading the following page https://tp.seg-social.es/fs/indexframes.html with java 8 65+ version.


Actual results:

Results in a stop plugin warning since firefox 42.
The application becomes unresponsive until the plugin it's stopped.

It happened since the upgrade to firefox 42 and java 8 in a widows 10 machine.

Actually working with a downgraded v 41.9 same java version (latest 8.71)


Expected results:

The page would have to load a java applet to load a cetificate file.

The working of the java applet don't matter the problem it's that it don't load, you don't need to log correctly in the page the problem it's that the explore option works.

Comment 1

[Loic]

The security of this website is really low:
https://www.ssllabs.com/ssltest/analyze.html?d=tp.seg-social.es
Outdated RC4 suites, insecure SSL3 still enabled, no TLS 1.1 or 1.2...

Comment 2

[Masatoshi Kimura [:emk]]

Neither Firefox 42 nor 41 are supported now. Is it reproducible with 44?

Comment 3

[[e-mail]]

The page uses a java applet to load a certificate.
The java app are slow to load in that windows 10 machine i'll try to replicate in a windows 7 with the firefox 43 that it's the version in release at least in spanish.

I know the page has serious issues ... even worst as it's a government page ,from spain ...

Comment 4

[Masatoshi Kimura [:emk]]

Spanish Firefox 44 Release is already available:
https://www.mozilla.org/en-US/firefox/all/

Comment 5

[[e-mail]]

It also happen with the 44 version in a windows 7 machine, but it's not as severe. But all java apps load much slower than the version 41 was using before.

The unresponsive java apps in the windows 10 machine may be related to some windows 10 changes, I'll search further about it.

But as I've said since upgrade to firefox 42+ I get errors related to java loading apps, seem like the brower loads the slow and give an stop plugin warning.

Sorry to no being of more help. No more machines to try. I'll try to make test in other machines when posible.

Comment 6

[YF (Yang)]

I see the java plugin ui and firefox ui hang more than ten seconds every time in Fx44.0 & Java 8 Update 71. It seems that due to network I/O affect the UI (Java & Firefox). I don't tested other versions for Firefox and Java.

In addition, I met a plugin crash: bp-bb084b1a-0ae1-4fd7-844d-8192d2160214, although it may be irrelevant.

Comment 7

[Loic]

Bad in FF44, WFM with FF47.

Comment 8

[alex_mayorga]

¡Hola Robert!

Per https://crash-stats.mozilla.com/report/list?product=Firefox&signature=hang+%7C+WaitForSingleObjectEx+%7C+WaitForSingleObject+%7C+npjp2.dll%400xb966 there were 7530 of these crashes in the past week.

Who should be looking at this one?

¡Hola Rubén!

Anyone you know at Ministerio de Empleo that could help fixing the SSL mess at https://tp.seg-social.es/fs/indexframes.html ?

¡Gracias!
Alex

Ps: Dropping a breadcrumb for myself... https://support.mozilla.org/en-US/questions/1112548

Comment 9

[Liz Henry (:lizzard) (relman/hg->git project)]

This is in the top 20 for 44.0.2 crashes.  It's showing up in other channels, but not as strongly as on release. 

I'm not sure if this is actionable; I can ask the crash debug team for help here.

Comment 10

[Robert Kaiser]

(In reply to alex_mayorga from comment #8)
> Who should be looking at this one?

Oracle should be looking into those.

Comment 11

[Rubén Martín [:Nukeador] ❌ [away till Aug 31st]]

Rafa, can you help us to contact someone working on this site?

Comment 12

[(not currently active) Ted Mielczarek]

I'm not sure this is actionable on Mozilla's side--these crash reports are essentially "the Java plugin was unresponsive for too long and we killed it". Essentially the only thing we could do here would be to bump the timeout, but that offers a crappy user experience in general. (You get to sit around with a hung Firefox for longer in most cases.)

I guess if we have the infrastructure and this is limited to particular sites we could add a site-specific override for the plugin timeout?

Comment 13

[alex_mayorga]

¡Hola!

A few data points from https://support.mozilla.org/es/questions/1112548#answer-853489

- On Firefox 45 with JRE 8u73, the user reports the hang goes away by switching back and forth with <Alt>+>Tab>.

- On Firefox 48 with JRE 8u73, the issue is resolved.

Seems like e10s is working around this.

Could the original reporter test with Firefox Developer Edition and Firefox Nightly and see if the bug is resolved there?

¡Gracias!
Alex

Comment 14

[Loic]

The user should test with e10s disabled. Could you ask him, please?

Comment 15

[Calixte Denizet (:calixte)]

This signature is #5 in plugin topcrashes on 45.0.1.
The number of crashes rose by ~100% between 2016-03-27 and 2016-03-30.

Comment 16

[Calixte Denizet (:calixte)]

On beta, it's #4 in plugin topcrash and there is a spike the 2016-03-28 (as for release).

Comment 17

[Liz Henry (:lizzard) (relman/hg->git project)]

Noting previous discussions of Java plugin crashes, in bug 1140616  and bug 874167. 
We don't see this crash with e10s enabled. 

Is there anything we can do here to stop users from crashing, even if the page isn't useable?
Andrew, what do you think of Ted's suggestion in comment 12?

Benjamin, do you have any suggestions here?

Comment 18

[Andrew McCreight [:mccr8]]

I don't really know anything about plugins, sorry.

Comment 19

[Benjamin Smedberg]

As far as I know, Firefox is not crashing. This is a plugin hang where Java stops responding, Firefox hangs until Java recovers or we hit our timeout. We know this is happening when Java shows various kinds of modal dialog boxes without spinning the Firefox event loop.

We're not prepared to add site-specific overrides or variations at this time. I think this is probably WONTFIX.

Comment 20

[marcino245]

I had the Java crashes multiple times a day. After Java 8 Update 91 has been released yesterday, I have no crashes anymore (as for now). They probably fixed the crash-bug?

Comment 21

[BugBot [:suhaib / :marco/ :calixte]]

Crash volume for signature 'hang | WaitForSingleObjectEx | WaitForSingleObject | npjp2.dll@0xb966':
 - nightly (version 50): 2 crashes from 2016-06-06.
 - aurora  (version 49): 36 crashes from 2016-06-07.
 - beta    (version 48): 1081 crashes from 2016-06-06.
 - esr     (version 45): 2920 crashes from 2016-04-07.

Crash volume on the last weeks:
             Week N-1   Week N-2   Week N-3   Week N-4   Week N-5   Week N-6   Week N-7
 - nightly          0          0          0          2          0          0          0
 - aurora           9          7          8          5          4          1          0
 - beta           122        228         71        139        230        225         34
 - release        968        909       1012       1202       1265       1184        309
 - esr            201        276        278        358        489        523        262

Affected platform: Windows