Last Comment Bug 778408 - mail.aol.com defaults to basic web UI when accessed with a user agent without Gecko/20100101
: mail.aol.com defaults to basic web UI when accessed with a user agent without...
Status: RESOLVED FIXED
:
Product: Tech Evangelism Graveyard
Classification: Graveyard
Component: English US (show other bugs)
: unspecified
: All All
: -- normal
: ---
Assigned To: english-us
:
Mentors:
Depends on: 782453
Blocks: 588909 797363
  Show dependency treegraph
 
Reported: 2012-07-28 01:41 PDT by j.j.
Modified: 2015-04-19 23:39 PDT (History)
9 users (show)
See Also:
QA Whiteboard:
Iteration: ---
Points: ---


Attachments

Description j.j. 2012-07-28 01:41:09 PDT
See bug 776376 comment 19 and
FYI: the webmail of aol.com also seems to identify "Firefox the wrong way" and does show a old version of the interface only. Overrided the useragent string to ... Gecko/20120429 ... and found the new interface appearing.

Reported the user agent issue to AOL. Waiting for response now.
Comment 1 Dão Gottwald [:dao] 2012-07-30 16:27:20 PDT
This will be an easy fix for AOL. The question is whether we can reach them to get them fix it before it hits users. If they only react once it has hit users, then the question is how severe this bug is, i.e.: how much inferior is this old version of the interface?
Comment 2 Alex Keybl [:akeybl] 2012-07-30 16:56:30 PDT
(In reply to Dão Gottwald [:dao] from comment #1)
> This will be an easy fix for AOL. The question is whether we can reach them
> to get them fix it before it hits users.

I think the real question is whether they can get a fix out in the next couple of weeks, since after that, backing out bug 588909 to may cause other server-side web regressions later in the cycle.

> If they only react once it has hit
> users, then the question is how severe this bug is, i.e.: how much inferior
> is this old version of the interface?

Does this UA change have so much value in the short-term that we'd consider taking any visible regression like this on a major web property?
Comment 3 Dão Gottwald [:dao] 2012-07-30 17:04:15 PDT
(In reply to Alex Keybl [:akeybl] from comment #2)
> Does this UA change have so much value in the short-term that we'd consider
> taking any visible regression like this on a major web property?

Since it took months for this bug to be reported, I expect the regression to be negligible. However, backing out bug 588909 from aurora surely is a reasonable option as well.
Comment 4 Henri Sivonen (:hsivonen) (Not doing reviews or reading bugmail until 2016-08-01) 2012-08-13 23:20:24 PDT
(In reply to Dão Gottwald [:dao] from comment #3)
> (In reply to Alex Keybl [:akeybl] from comment #2)
> > Does this UA change have so much value in the short-term that we'd consider
> > taking any visible regression like this on a major web property?
> 
> Since it took months for this bug to be reported, I expect the regression to
> be negligible. However, backing out bug 588909 from aurora surely is a
> reasonable option as well.

Instead of keeping changing the UA string back and forth for all sites thereby giving all sites the opportunity to sniff for the parts we are trying to remove, I suggest we send a different UA string for mail.aol.com only.

Chrome has done that sort of thing e.g. for Yahoo! Mail, and experience shows that they've been able to remove such hacks after some time, so we shouldn't be afraid of such hacks becoming too much of a maintenance burden.
Comment 5 Dão Gottwald [:dao] 2012-08-14 00:32:02 PDT
(In reply to Henri Sivonen (:hsivonen) from comment #4)
> (In reply to Dão Gottwald [:dao] from comment #3)
> > (In reply to Alex Keybl [:akeybl] from comment #2)
> > > Does this UA change have so much value in the short-term that we'd consider
> > > taking any visible regression like this on a major web property?
> > 
> > Since it took months for this bug to be reported, I expect the regression to
> > be negligible. However, backing out bug 588909 from aurora surely is a
> > reasonable option as well.
> 
> Instead of keeping changing the UA string back and forth for all sites
> thereby giving all sites the opportunity to sniff for the parts we are
> trying to remove, I suggest we send a different UA string for mail.aol.com
> only.
> 
> Chrome has done that sort of thing e.g. for Yahoo! Mail, and experience
> shows that they've been able to remove such hacks after some time, so we
> shouldn't be afraid of such hacks becoming too much of a maintenance burden.

See bug 782453.
Comment 6 Dão Gottwald [:dao] 2012-09-04 22:56:30 PDT
Firefox 17 isn't affected by this anymore since bug 782453 landed there.
Comment 7 Gervase Markham [:gerv] 2012-10-02 04:35:48 PDT
j.j.: have you heard anything back from AOL? What contact methods did you use to notify them of this problem?

Gerv
Comment 8 Bruno Marsal (Bullja) 2012-10-02 05:31:21 PDT
I described the problem to AOL Mail Support and got the standard answer first telling me to reset browser settings (cookies, cache, history, ...) and use default browser.

Replied to that with a step by step explanation (Gerv, forwarded to you) to reproduce the problem. 

From: AOL Mail Team <AOLMailQuestions@aol.com>
[...]
We always appreciate customers who take the time to give us their feedback. I'll pass what you've said onto our management team.
[...]

Since then, did not hear anything back (not expecting any response).
Comment 9 j.j. 2012-10-02 16:52:57 PDT
(In reply to Gervase Markham [:gerv] from comment #7)
> j.j.: have you heard anything back from AOL?
(I just cited Bruno Marsal's comment)
Comment 10 Kev Needham [:kev] 2012-10-03 07:33:45 PDT
Really don't like modifying UA strings per site as a general rule, and think we should avoid things like bug 782453 that is caused by sniffing for something we change, as it defeats the point of the change we made.

I hadn't seen this issue before, and will raise it with Aol, as well. Going through their general support rarely gets much traction.
Comment 11 Bruno Marsal (Bullja) 2012-10-03 07:48:02 PDT
Just tried again setting general.useragent.override with "Mozilla/5.0 (Windows NT x.y; rv:10.0) Gecko/10.0 Firefox/10.0" and was successfully logging into the new AOL Mail UI. Seems they fixed that.

Can somebody confirm?
Comment 12 Dão Gottwald [:dao] 2012-10-03 08:01:31 PDT
(In reply to Kev [:kev] Needham from comment #10)
> Really don't like modifying UA strings per site as a general rule, and think
> we should avoid things like bug 782453 that is caused by sniffing for
> something we change, as it defeats the point of the change we made.

Undoing the change for a handful of sites out of billions certainly doesn't defeat the point of the change.
Comment 13 Dão Gottwald [:dao] 2012-10-03 08:06:51 PDT
(In reply to Bruno Marsal (Bullja) from comment #11)
> Just tried again setting general.useragent.override with "Mozilla/5.0
> (Windows NT x.y; rv:10.0) Gecko/10.0 Firefox/10.0" and was successfully
> logging into the new AOL Mail UI. Seems they fixed that.
> 
> Can somebody confirm?

general.useragent.override isn't going to make a difference, the preference in question is general.useragent.override.aol.com. That said, it indeed seems like it's fixed.
Comment 14 Kev Needham [:kev] 2012-10-03 08:21:44 PDT
That's your opinion. I'll shut up moving forward.

(In reply to Dão Gottwald [:dao] from comment #12)
 
> Undoing the change for a handful of sites out of billions certainly doesn't
> defeat the point of the change.
Comment 15 Gervase Markham [:gerv] 2012-10-04 05:54:35 PDT
Why is this bug resolved? The problem is still present. The bug can be resolved when AOL has fixed their site and we can remove the preference which mitigates the problem. Adding a site to an override list is not a "fix", it's a temporary mitigation.

Gerv
Comment 16 Dão Gottwald [:dao] 2012-10-04 07:04:27 PDT
(In reply to Gervase Markham [:gerv] from comment #15)
> Why is this bug resolved?

Because AOL apparently fixed the problem.
Comment 17 Gervase Markham [:gerv] 2012-10-04 07:47:15 PDT
Oh, awesome! Sorry, I didn't notice that. So can we remove the entry from the UA override list, then?

Gerv
Comment 18 Dão Gottwald [:dao] 2012-10-04 07:47:45 PDT
bug 797363

Note You need to log in before you can comment on or make changes to this bug.