Closed Bug 905963 Opened 11 years ago Closed 11 years ago

[ga-IE] Test failure "A new window has been opened" in /testPrivateBrowsing/testOpenClosePBKeyboardShortcut.js

Categories

(Mozilla QA Graveyard :: Mozmill Tests, defect, P2)

defect

Tracking

(firefox23 wontfix, firefox24 wontfix, firefox25 disabled, firefox26 fixed, firefox27 fixed, firefox-esr17 unaffected, firefox-esr24 disabled)

RESOLVED FIXED
Tracking Status
firefox23 --- wontfix
firefox24 --- wontfix
firefox25 --- disabled
firefox26 --- fixed
firefox27 --- fixed
firefox-esr17 --- unaffected
firefox-esr24 --- disabled

People

(Reporter: cosmin-malutan, Assigned: andrei)

References

()

Details

(Keywords: reproducible, Whiteboard: [mozmill-test-failure])

Attachments

(1 file, 1 obsolete file)

Happened on Windows 7 x(86) with Release ga-IE
http://mozmill-release.blargon7.com/#/functional/report/b7ef1fb3d9703aeaf2c46e07d2e5b38f

I reproduced this locally, I could't reproduce it with other locales.
Does that only happen on Windows? Could that also be an access key issue?
Keywords: reproducible
Priority: -- → P2
Summary: Test failure "A new window has been opened" in /testPrivateBrowsing/testOpenClosePBKeyboardShortcut.js → [ga-IE] Test failure "A new window has been opened" in /testPrivateBrowsing/testOpenClosePBKeyboardShortcut.js
I see this also happened on OS X, with ku locale. Please see the url.
Clearly a localisation issue.

privateBrowsingCmd.commandkey is set to R, which on the same key combination Firefox does a no-cache refresh (Cmd+Shift+R)

http://mxr.mozilla.org/l10n-central/source/ga-IE/browser/chrome/browser/browser.dtd#253

Filing a bug against the localisation now.
Assignee: nobody → andrei.eftimie
Status: NEW → ASSIGNED
OS: Windows 7 → All
Hardware: x86 → All
Depends on: 906673
CCing Axel since this is a localization issue.
Andrei, are you sure this failure happens with a Nightly build? AFAIK we have stopped the localization on mozilla-central. Can you please test this?
Flags: needinfo?(andrei.eftimie)
ga-IE is not working on trunk but on mozilla-aurora
http://hg.mozilla.org/l10n-central/ga-IE/htmlpushlog

ku is not working at all (last sign-off is for Firefox 10), all changes were made from myself to fix outstanding problems. TBH I don't think you should be looking at ku (see also bug 895996)
http://hg.mozilla.org/l10n-central/ku/htmlpushlog
As of yesterday, when I tested this, there was no Nightly build available for ga-IE.
All other branched have been manually tested.

I've set up a tracking flag for Nightly as well since the faulty code is available in l10n-central: http://mxr.mozilla.org/l10n-central/source/ga-IE/browser/chrome/browser/browser.dtd#253

The assumption I made is that any Nightly build that is going to be made for ga-IE will use the same code and exhibit the same problem. I'm guessing that is not the case anymore :)

This is the used repository: http://hg.mozilla.org/releases/l10n/mozilla-aurora/ga-IE/
Flags: needinfo?(andrei.eftimie)
(In reply to Francesco Lodolo [:flod] from comment #6)
> ga-IE is not working on trunk but on mozilla-aurora
> http://hg.mozilla.org/l10n-central/ga-IE/htmlpushlog

At least we should get serious failures fixed which prevent us from running our tests. Otherwise we might have to implement a blacklist for locales to test. In this case the fix seems to be simple enough. Lets follow-up on bug 906673 with it.

> ku is not working at all (last sign-off is for Firefox 10), all changes were
> made from myself to fix outstanding problems. TBH I don't think you should
> be looking at ku (see also bug 895996)
> http://hg.mozilla.org/l10n-central/ku/htmlpushlog

Thank you Francesco for all that extra work. I wonder if we should also fix that quickly in ku so we can run our tests. On the other side why do we still produce Firefox builds if no signoff on the locale has happened since Firefox 10? That sounds awkward. :S
Flags: needinfo?(l10n)
Running tests against trunk: locales which work there are very few (a whitelist would probably be more useful than a blacklist). Even if we fix these errors, what's the point of testing something that is not used and never updated? It seems a waste of resources to me.

Running tests against aurora: I think you need a blacklist here, because some locales are clearly not maintained (e.g. ku), in order to exclude them from testings.
Flags: needinfo?(l10n)
Verified the fix in Aurora.
Release and ESR17 will not get it.

Kevin, could we get this backported (bug 906673) in Beta as well?
Flags: needinfo?(kscanne)
Andrei, is there anything else needed here or can we close this?
Flags: needinfo?(andrei.eftimie)
I'm still waiting for a Beta build that fixed the issue, so I can marked it fixed on Beta as well.

The one from 22nd August was still broken.
Will probably have a resolve with tomorrow's Beta build.
Flags: needinfo?(andrei.eftimie)
(In reply to Andrei Eftimie from comment #13)
> I'm still waiting for a Beta build that fixed the issue, so I can marked it
> fixed on Beta as well.
> 
> The one from 22nd August was still broken.
> Will probably have a resolve with tomorrow's Beta build.

Bhavana and Francesco, will the latest changes be automatically included or would we need a specific l10n signoff?
Flags: needinfo?(francesco.lodolo)
Flags: needinfo?(bbajaj)
It needs a sign-off (and review) that has not been requested yet
https://l10n.mozilla.org/shipping/signoffs/ga-IE/fx24
Flags: needinfo?(francesco.lodolo) → needinfo?(kscanne)
Flags: needinfo?(bbajaj)
Cleared needinfo on self as :Flod  has given the needed answer here.
Works for me. Signing off now.
Status: ASSIGNED → RESOLVED
Closed: 11 years ago
Flags: needinfo?(kscanne)
Resolution: --- → FIXED
I don't see this in the latest Beta build (20130826142034) yet.

Tested on https://ftp.mozilla.org/pub/mozilla.org/firefox/candidates/24.0b6-candidates/build1/mac/ga-IE/Firefox%2024.0b6.dmg

Leaving this bug opened until the fix is landed in a Beta build.
Status: RESOLVED → REOPENED
Resolution: FIXED → ---
Verified with the latest Beta 24.0b7
All good
Status: REOPENED → RESOLVED
Closed: 11 years ago11 years ago
Resolution: --- → FIXED
Please don't for get to adjust the flags.
We've had a recent fail on 24 :
http://mozmill-release.blargon7.com/#/functional/report/3c3cd991de01b704f3f65783a1edeafc

This should have been fixed, I will look into it.
Status: RESOLVED → REOPENED
Resolution: FIXED → ---
We have another conflict only on Linux and Windows.

It seems opening the Private Browsing Mode uses different keyboard shortcuts on OSX (Alt+Cmd+"key") than on Windows and Linux (Ctrl+Shift+"key").

More information in bug 906673 comment 6
If it's reproducing constantly, let's get this skipped for the specific branch/platform affected.
Attached patch skip.patch (obsolete) — Splinter Review
Skipping the test.

Patch applies cleanly to:

mozilla-aurora
mozilla-beta
mozilla-release
mozilla-esr24

Test does not exist on mozilla-esr17, there is nothing to skip there.
And we have no ga-IE builds on Nightly, so we might leave this on default.
Attachment #802889 - Flags: review?(andreea.matei)
Comment on attachment 802889 [details] [diff] [review]
skip.patch

Review of attachment 802889 [details] [diff] [review]:
-----------------------------------------------------------------

::: tests/functional/testPrivateBrowsing/testOpenClosePBKeyboardShortcut.js
@@ +38,5 @@
>    pbWindow.close();
>  }
> +
> +setupModule.__force_skip__ = "Bug 905963 - Skip due to failure on ga-IE";
> +teardownModule.__force_skip__ = "Bug 905963 - Skip due to failure on ga-IE";

We use to put the actual failure in the skip messages and have this on in the commit eventually.
Attachment #802889 - Flags: review?(andreea.matei) → review-
Attached patch skip_2.patchSplinter Review
I actually took the message format from the similar skip from [lg] from bug 909227 attachment 802842 [details] [diff] [review] 
Don't know why I didn't use the standard skip message.

Anyway, updated here.

This needs to land on the branches named in comment 26
Attachment #802889 - Attachment is obsolete: true
Attachment #803620 - Flags: review?(andreea.matei)
Whiteboard: [mozmill-test-failure] → [mozmill-test-failure][mozmill-test-skipped]
This issue is fixed in Aurora,
lets hope it will be pushed to Beta as well.

Backed out skip:
http://hg.mozilla.org/qa/mozmill-tests/rev/c8fa38aad950 (mozilla-aurora)
It needs an explicit signoff on beta. If that hasn't happened it will not be fixed in beta. So please try that we get one.
Unfortunately the fix didn't make it into 25 :(

I don't think there's anything else we can do here.
We'll have this issue fixed on 26 and upwards.
Status: ASSIGNED → RESOLVED
Closed: 11 years ago11 years ago
Resolution: --- → FIXED
Whiteboard: [mozmill-test-failure][mozmill-test-skipped] → [mozmill-test-failure]
Product: Mozilla QA → Mozilla QA Graveyard
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Creator:
Created:
Updated:
Size: