Closed Bug 1055464 Opened 10 years ago Closed 8 years ago

"Warn me when websites try to redirect or reload the page" broken with e10s

Categories

(Firefox :: Disability Access, defect, P3)

defect

Tracking

()

RESOLVED FIXED
Firefox 47
Tracking Status
e10s m9+ ---
firefox46 - fixed
firefox47 - fixed

People

(Reporter: annevk, Assigned: mconley)

References

(Blocks 1 open bug)

Details

(Keywords: dogfood)

Attachments

(3 files)

For links that open in a new window, e.g. from Gmail, redirects are followed automatically after flipping browser.tabs.remote.autostart.
This is supposed to be handled by TabsProgressListener.onRefreshAttempted:

http://hg.mozilla.org/mozilla-central/annotate/111a1da2a95d/browser/base/content/browser.js#l4218
Blocks: fxe10s
Component: Preferences → General
OS: Mac OS X → All
Hardware: x86 → All
Blocks: old-e10s-m2
Priority: -- → P3
not blocking m2, sorry.
No longer blocks: old-e10s-m2
Keywords: dogfood
Component: General → Disability Access
Flags: needinfo?(twalker)
I can reproduce this on latest Nightly with test case from bug 1150311: https://bugzilla.mozilla.org/attachment.cgi?id=8590090

I'll provide a full report shortly, but I want to get his into today's triage as I feel this is higher priority than trackinge10s+.
Flags: needinfo?(twalker)
Assignee: nobody → mconley
Attachment #8716463 - Flags: review?(dtownsend) → review+
Comment on attachment 8716463 [details]
MozReview Request: Bug 1055464 - Add RefreshBlocker to tab-content, and a listener in tabbrowser.xml. r=Mossop

https://reviewboard.mozilla.org/r/33819/#review30641

::: browser/base/content/tab-content.js:758
(Diff revision 1)
> +  },

When you don't include a method xpconnect automatically just throws NS_ERROR_NOT_IMPLEMENTED when called anyway so I don't think these are useful.
Attachment #8716464 - Flags: review?(dtownsend) → review+
Comment on attachment 8716464 [details]
MozReview Request: Bug 1055464 - Add RefreshBlocker front-end. r=Mossop

https://reviewboard.mozilla.org/r/33821/#review30645

::: browser/base/content/global-scripts.inc:27
(Diff revision 1)
> +<script type="application/javascript" src="chrome://browser/content/browser-refreshblocker.js"/>

Please also update https://dxr.mozilla.org/mozilla-central/source/testing/eslint-plugin-mozilla/lib/rules/import-browserjs-globals.js
Comment on attachment 8716465 [details]
MozReview Request: Bug 1055464 - Regression tests. r=Mossop

https://reviewboard.mozilla.org/r/33823/#review30651
Attachment #8716465 - Flags: review?(dtownsend) → review+
Comment on attachment 8716463 [details]
MozReview Request: Bug 1055464 - Add RefreshBlocker to tab-content, and a listener in tabbrowser.xml. r=Mossop

Review request updated; see interdiff: https://reviewboard.mozilla.org/r/33819/diff/1-2/
Attachment #8716463 - Attachment description: MozReview Request: Bug 1055464 - Add RefreshBlocker to tab-content, and a listener in tabbrowser.xml. r?Mossop → MozReview Request: Bug 1055464 - Add RefreshBlocker to tab-content, and a listener in tabbrowser.xml. r=Mossop
Comment on attachment 8716464 [details]
MozReview Request: Bug 1055464 - Add RefreshBlocker front-end. r=Mossop

Review request updated; see interdiff: https://reviewboard.mozilla.org/r/33821/diff/1-2/
Attachment #8716464 - Attachment description: MozReview Request: Bug 1055464 - Add RefreshBlocker front-end. r?Mossop → MozReview Request: Bug 1055464 - Add RefreshBlocker front-end. r=Mossop
Attachment #8716465 - Attachment description: MozReview Request: Bug 1055464 - Regression tests. r?Mossop → MozReview Request: Bug 1055464 - Regression tests. r=Mossop
Comment on attachment 8716465 [details]
MozReview Request: Bug 1055464 - Regression tests. r=Mossop

Review request updated; see interdiff: https://reviewboard.mozilla.org/r/33823/diff/1-2/
Depends on: 1247100
Comment on attachment 8716463 [details]
MozReview Request: Bug 1055464 - Add RefreshBlocker to tab-content, and a listener in tabbrowser.xml. r=Mossop

Approval Request Comment
[Feature/regressing bug #]:

Electrolysis, and the accessibility / user control setting for protecting the user from automatic redirection / reload.

[User impact if declined]:

Users that have e10s enabled will not be able to prevent sites from redirecting or reloading despite having the preference set.

Note that this isn't just an about:config pref, but a preference that we expose in about:preferences under Advanced > General.

[Describe test coverage new/current, TreeHerder]:

I've added automated tests for this.

[Risks and why]: 

This bug has opened up a 5% session restore talos regression on Linux 64, being tracked in bug 1247100. There are no regressions being tracked on any of the other platforms.

[String/UUID change made/needed]:

None.
Attachment #8716463 - Flags: approval-mozilla-aurora?
Comment on attachment 8716464 [details]
MozReview Request: Bug 1055464 - Add RefreshBlocker front-end. r=Mossop

Approval Request Comment

See comment 15.
Attachment #8716464 - Flags: approval-mozilla-aurora?
Comment on attachment 8716465 [details]
MozReview Request: Bug 1055464 - Regression tests. r=Mossop

Approval Request Comment

See comment 15.
Attachment #8716465 - Flags: approval-mozilla-aurora?
Blocks: 1246291
This doesn't appear to be fixed.  Might even be worse depending on your perspective.  See bug 1246291.  There I reported that the redirect warming is not appearing, nor is the page redirecting.
[Tracking Requested - why for this release]:

Tracy, do you mean this isn't fixed in nightly?   If so, can you reopen the bug/set the flags accordingly?
Mike, I will hold off on uplift for now and wait on developments in bug 1246291 if that seems right to you.
Flags: needinfo?(twalker)
Flags: needinfo?(mconley)
Thanks Liz, I thought I had re-opened this when I made the connection in bug 1246291.  Note: they are probably dupes. Unless the work for displaying the warning and the work for actually not redirecting will be treated separately, in which case, this bug should be used for the former and bug 1246291 used for the latter.
Status: RESOLVED → REOPENED
Flags: needinfo?(twalker)
Resolution: FIXED → ---
I would argue that this bug should be closed, and that the remaining work should occur in bug 1246291.

The reason for that is because the patch in this bug, as landed, DOES fix the STR that are laid out in comment 0 and comment 3.

The issue in bug 1246291 is a special-case of redirecting (redirect via the HTTP header, instead of a <meta> tag in an HTML document body).

The patch in this bug fixes the latter case for e10s. It also unintentionally fixes the former case (HTTP header redirect) for non-e10s. It does not fix the former case for the e10s case. That's what bug 1246291 is for.

Tracy, does that gel with your observations? If so, we can re-close this and continue the work in bug 1246291?
Flags: needinfo?(mconley) → needinfo?(twalker)
Yes, sounds good.
Status: REOPENED → RESOLVED
Closed: 8 years ago8 years ago
Flags: needinfo?(twalker)
Resolution: --- → FIXED
Untracking since this is tracked in 1246291.
Comment on attachment 8716463 [details]
MozReview Request: Bug 1055464 - Add RefreshBlocker to tab-content, and a listener in tabbrowser.xml. r=Mossop

Fixes redirects with e10s along with patches in bug 1246291.
Wes - may be best to uplift these after 1246291.
Attachment #8716463 - Flags: approval-mozilla-aurora? → approval-mozilla-aurora+
Attachment #8716464 - Flags: approval-mozilla-aurora? → approval-mozilla-aurora+
Attachment #8716465 - Flags: approval-mozilla-aurora? → approval-mozilla-aurora+
[bugday-20160323]

Status: RESOLVED,FIXED -> VERIFIED

Component: 
Name 			Firefox
Version 		46.0b4
Build ID 		20160322075646
Update Channel 	        beta
User Agent 		Mozilla/5.0 (Windows NT 6.1; WOW64; rv:46.0) Gecko/20100101 Firefox/46.0
OS                      Windows 7 SP1 x86_64

Actual Results: 
As expected

Expected Results: 
No automatic redirections are there for sites like Gmail.
Not reproducible for FF 46.0b4.
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Created:
Updated:
Size: