1166771 - pushstate for reader mode should reset browser.isArticle to false if the new page isn't readerable

Status: VERIFIED FIXED

(Firefox :: General, defect)

Description

[:Gijs (he/him)]

This was originally: https://github.com/mozilla/readability/issues/227

Reproducible on:
Beta 38.0.5 beta 3 - build1 (20150518141916)
Affected platforms: Windows 7 x86, Windows 8.1 x64, Mac OS X 10.9.5 and Ubuntu 14.04 x64

Steps to reproduce:
1. Navigate to http://espn.go.com/
2. Click on any article
3. Hit Back button
Actual result: Enter Reader View button is visible in URL bar for espn.com homepage
4. Click on Enter Reader View button
Actual result: 'Failed to load article from page'

Sample pages showing this issue:

    http://www.usatoday.com/
    https://www.youtube.com/
    http://espn.go.com/



(this also affects nightly)

Comment 1

[:Gijs (he/him)]

Attachment: MozReview Request: bz://1166771/Gijs

/r/9127 - Bug 1166771 - force isArticle to false on pushstate on non-article pages, r?margaret

Pull down this commit:

hg pull -r dc54fbcffb8da01ac19f7204239d6ea3bebc947e https://reviewboard-hg.mozilla.org/gecko/

Comment 2

[:Margaret Leibovic]

Comment on attachment 8608161 [details]
MozReview Request: bz://1166771/Gijs

https://reviewboard.mozilla.org/r/9125/#review7761

Ship It!

Comment 3

[Pulsebot]

https://hg.mozilla.org/integration/fx-team/rev/bd4b92eb955d

Comment 4

[:Gijs (he/him)]

Comment on attachment 8608161 [details]
MozReview Request: bz://1166771/Gijs

Approval Request Comment
[Feature/regressing bug #]: reader mode perf adjustments
[User impact if declined]: reader mode icon sticks around when it shouldn't
[Describe test coverage new/current, TreeHerder]: nope
[Risks and why]: low. This should only adjust the currently missed case where a page change doesn't cause a pagehide event but only a pushstate event where the new state of the doc doesn't need reader mode, whereas the old state of the doc *did*, which then causes the button to mistakenly hang around.
[String/UUID change made/needed]: nope

Dolske/Margaret, would love opinions about whether you think taking this for 38RC (gtb tomorrow!) is too risky or not. Re-reading the patch and the risk description above (which I think is accurate!) I think it's non-risky enough, but I appreciate it's late in the cycle.

Comment 5

[:Margaret Leibovic]

(In reply to :Gijs Kruitbosch from comment #4)

> Dolske/Margaret, would love opinions about whether you think taking this for
> 38RC (gtb tomorrow!) is too risky or not. Re-reading the patch and the risk
> description above (which I think is accurate!) I think it's non-risky
> enough, but I appreciate it's late in the cycle.

I think it is low-risk. Looking at the logic, in the worst-case scenario, we just end up sending a message to hide the reader button when we otherwise wouldn't send that message.

That being said, I think we should make sure to bang on a build to make sure that nothing unexpected happens, because even low-risk patches can sometimes contain surprises.

Comment 6

[Wes Kocher (:KWierso) (Not reading bugmail; email directly if needed)]

https://hg.mozilla.org/mozilla-central/rev/bd4b92eb955d

Comment 7

[Sylvestre Ledru [:Sylvestre]]

Comment on attachment 8608161 [details]
MozReview Request: bz://1166771/Gijs

It is a new feature, affect high-profile sites, should be safe. Taking it. Hopefully, it won't break everything :)

Comment 8

[Ryan VanderMeulen [:RyanVM]]

https://hg.mozilla.org/releases/mozilla-aurora/rev/62af7310ed73

Comment 9

[Ryan VanderMeulen [:RyanVM]]

https://hg.mozilla.org/releases/mozilla-beta/rev/bc222d611a7c

Comment 10

[Ryan VanderMeulen [:RyanVM]]

https://hg.mozilla.org/releases/mozilla-release/rev/17169e355c59

Comment 11

[Andrei Vaida [:avaida]]

Verified fixed on RC 38.0.5-build3 (20150521175336), using the sample pages from comment 0 along with several other similar pages. Further Regression testing will be conducted on this area during the validation of 38.0.5-build3.

Comment 12

[Ada [:adalucinet]]

Verified fixed using 39.0b1 build 2 (20150523155636), under the following platforms: Windows 7 64-bit, Mac OS X 10.9.5 and Ubuntu 14.04 32-bit.

Comment 13

[Florin Mezei, QA (:FlorinMezei)]

Removing qe-verify+ since verification on 38.0.5 Beta and 39 Beta should suffice.

Comment 15

[:Gijs (he/him)]

Attachment: MozReview Request: Bug 1166771 - force isArticle to false on pushstate on non-article pages, r?margaret