Test failure 'The correct SSL error title is shown - 'Unable to Connect Securely' should equal 'Secure Connection Failed'' in testSecurity/testSSLDisabledErrorPage.js

RESOLVED FIXED

Status

Mozilla QA
Mozmill Tests
P1
normal
RESOLVED FIXED
3 years ago
3 years ago

People

(Reporter: Teodor Druta, Assigned: cosmin)

Tracking

({regression})

Version 2
All
Mac OS X
regression

Firefox Tracking Flags

(firefox34 fixed, firefox35 fixed, firefox36 unaffected, firefox37 unaffected)

Details

(Whiteboard: [mozmill-test-failure], URL)

Attachments

(1 attachment, 2 obsolete attachments)

(Reporter)

Description

3 years ago
Module:    testSSLDisabledErrorPage
Test:      testSecurity/testSSLDisabledErrorPage.js 
Failure:   'The correct SSL error title is shown - 'Unable to Connect Securely' should equal 'Secure Connection Failed'
Branches:  mozilla-aurora, mozilla-beta
Platforms: All
Locale:    All
Report:    http://mozmill-daily.blargon7.com/#/remote/report/bdaefa581a1cff14362f7944d610df05

Reproducible on all platforms.
(Reporter)

Updated

3 years ago
status-firefox34: --- → affected
status-firefox35: --- → affected
status-firefox36: --- → unaffected
status-firefox37: --- → unaffected
Whiteboard: [mozmill-test-failure]
(Reporter)

Comment 1

3 years ago
Created attachment 8531168 [details] [diff] [review]
skipssldisablederrorpage.patch

We have multiple failures in this test we should skip this test on beta and aurora.

Added skip patch.
Attachment #8531168 - Flags: review?(andreea.matei)
(Reporter)

Comment 2

3 years ago
Created attachment 8531170 [details] [diff] [review]
skipssldisablederrorpage.patch

Sorry, wrong file.
Attachment #8531168 - Attachment is obsolete: true
Attachment #8531168 - Flags: review?(andreea.matei)
Attachment #8531170 - Flags: review?(andreea.matei)
(Reporter)

Updated

3 years ago
Comment on attachment 8531170 [details] [diff] [review]
skipssldisablederrorpage.patch

Review of attachment 8531170 [details] [diff] [review]:
-----------------------------------------------------------------

Patch doesn't apply anymore with the other skip, please update. Also the commit message.
Attachment #8531170 - Flags: review?(andreea.matei) → review-
(Reporter)

Comment 4

3 years ago
Created attachment 8531191 [details] [diff] [review]
skipssldisablederrorpage.patch

Fixed rejections.
Attachment #8531170 - Attachment is obsolete: true
Attachment #8531191 - Flags: review?(mihaela.velimiroviciu)
Attachment #8531191 - Flags: review?(andreea.matei)
Comment on attachment 8531191 [details] [diff] [review]
skipssldisablederrorpage.patch

Review of attachment 8531191 [details] [diff] [review]:
-----------------------------------------------------------------

This has 4 failures, all on OSX. Please see how often it fails, if you can reproduce. We shouldn't skip otherwise or not on all platforms. Just saw the reports now, sorry.
Attachment #8531191 - Flags: review?(mihaela.velimiroviciu)
Attachment #8531191 - Flags: review?(andreea.matei)
Attachment #8531191 - Flags: review-
(In reply to Teodor Druta from comment #0)
> Branches:  mozilla-aurora, mozilla-beta

The status flags do not match those branches. Can you please fix that? 

Also please see the other correlated failure here which is:

The SSL error message contains disabled information - 'Firefox cannot guarantee the safety of your data on tlsv1-0.mozqa.com because it uses SSLv3, a broken security protocol.' should contain 'ssl_error_no_cypher_overlap'

Seeing this I do not think it is a bug in our test but somehow the certificate still has SSLv3 enabled? This needs more investigation and totally not a skip patch. We only do that for P1 and partly P2.
(Reporter)

Updated

3 years ago
status-firefox34: affected → wontfix
status-firefox36: unaffected → affected
(Reporter)

Updated

3 years ago
(Reporter)

Updated

3 years ago
(In reply to Henrik Skupin (:whimboo) from comment #6)
> The SSL error message contains disabled information - 'Firefox cannot
> guarantee the safety of your data on tlsv1-0.mozqa.com because it uses
> SSLv3, a broken security protocol.' should contain
> 'ssl_error_no_cypher_overlap'
> 
> Seeing this I do not think it is a bug in our test but somehow the
> certificate still has SSLv3 enabled? This needs more investigation and
> totally not a skip patch. We only do that for P1 and partly P2.

Bug 1096695 changed the message from "An error occurred during a connection to write.skyword.com. Cannot communicate securely with peer: no common encryption algorithm(s). (Error code: ssl_error_no_cypher_overlap)" to "Firefox cannot
> guarantee the safety of your data on tlsv1-0.mozqa.com because it uses
> SSLv3, a broken security protocol.".
We should update the test rather than disabling it.
Thank you Masatoshi. Given that the mentioned patch landed for 34 and 35, I do not see why the flags have been set wrongly. So I can also only repeat, lets get this fixed. Thanks.

Oh, and given that it is affecting beta and release builds we want to get this fixed sooner than later.
Blocks: 1096695
status-firefox34: wontfix → affected
status-firefox36: affected → unaffected
Keywords: regression
Priority: -- → P1
(Assignee)

Comment 9

3 years ago
I just checked latest Aurora, Beta and Release
http://mozmill-crowd.blargon7.com/#/remote/report/946d3cd70aaf19c33a8878f8c4b13ea3
http://mozmill-crowd.blargon7.com/#/remote/report/946d3cd70aaf19c33a8878f8c4aee643
http://mozmill-crowd.blargon7.com/#/remote/report/946d3cd70aaf19c33a8878f8c4ade2b9

This bug has been fixed somewhere down the road.
Also this failed only on OSX before and only on en-US. This would come handy, as I checked first on linux until I noticed that and also other locales.

Andreea could you backout the skip?
OS: All → Mac OS X
(Assignee)

Updated

3 years ago
Flags: needinfo?(andreea.matei)
This is not skipped, it failed only a few times on OSX. We can close it then, I think the SSL error message went back.
Status: NEW → RESOLVED
Last Resolved: 3 years ago
status-firefox34: affected → fixed
status-firefox35: affected → fixed
Flags: needinfo?(andreea.matei)
Resolution: --- → FIXED
Assignee: nobody → cosmin.malutan
You need to log in before you can comment on or make changes to this bug.