964746 - Mapquest mobile site offers less features with gecko if compared to webkit

Status: RESOLVED FIXED

(Web Compatibility :: Site Reports, defect)

Description

[gabriele.vidali]

Attachment: mapquest.png

User Agent: Mozilla/5.0 (X11; Ubuntu; Linux x86_64; rv:29.0) Gecko/20100101 Firefox/29.0 (Beta/Release)
Build ID: 20140108093909

Steps to reproduce:

Open mobile website m.mapquest.com


Actual results:

Gps navigation system is not available.


Expected results:

The mobile site is a gps navigation software if opened with a webkit based browser. If opened with a gecko based browser it's just a map

Comment 1

[Karl Dubost💡 :karlcow]

This requires more analysis. 

First of all, on devices.
When opening with m.mapquest.com

* iOS Safari - a complex UI with a native feel.
* Opera Android (Presto) - similar UI than iOS Safari. 
* UCWeb - similar UI than iOS Safari.
* Firefox OS - a simple web form for directions (no style)
* Firefox Android - a simple web form for directions (with CSS applied)

So yes there is definitely a different experience depending on the user agent, but this needs to be analyzed a bit more, beforce being able to contact.

Comment 2

[Mike Taylor [:miketaylr]]

Assigning to me while I poke at its insides.

Comment 3

[Mike Taylor [:miketaylr]]

So, I don't actually get the lofi mobile site anymore--but am directed to the desktop site. Spoofing as an iPhone, the mobile site works surprisingly well. Basically everything works as expected except turning on the traffic view.

They're using iScroll 4.x, which is known to have compat issues with Gecko browsers. I would recommend they upgrade to 5.x (though I don't see any obvious places where that is breaking).

A few notes:

There are some inline MQs in the head which will only work for WebKit browsers:

	@media only screen and (-webkit-min-device-pixel-ratio: 1.5) and (-webkit-max-device-pixel-ratio: 1.99) {...}
	@media only screen and (-webkit-min-device-pixel-ratio: 2) {...}

In Firefox, you get the following error in the console, "error getting position" which comes from https://gist.github.com/miketaylr/8713468#file-mapquest-js-L32-L38... which is curious, given that getxCurrentPosition isn't an actual thing (just a typo?). Unsure right now why Chrome Mobile doesn't hit this codepath.

The site works surprisingly well, panning, zooming etc are functional--no obvious layout issues. So once we figure out this position bug (which may be why the traffic layers aren't appearing) I think we should recommend they send Firefox mobile browsers directly to the same m.mapquest.com site.

Comment 4

[Mike Taylor [:miketaylr]]

Further comments:

OK, so the "error getting position" error was an artifact of spoofing as iOS--See https://gist.github.com/miketaylr/8713468#file-mapquest-js-L46-L48 where "a" is defined as mqm.Util.isIOS5(). I have no idea if that's intentional or not, it really does seem like a typo.

https://gist.github.com/miketaylr/8717560#file-mqm-util-js-L449-L456 -- this should be updated to include a prefixless version.

Similar to the MQs in Comment 3, https://gist.github.com/miketaylr/8717560#file-mqm-util-js-L517-L528 should be updated to work in other browsers (plenty of resources on the web for this)

Traffic is working for me now (I think the previous iOS5 codepath/getPosition error was likely the culprit), so I think we can safely reach out and recommend they send us to the m dot site (and hopefully they can fix the small issues listed here).

Comment 5

[Mike Taylor [:miketaylr]]

Oh, one more thing--I just realized I was supposed to navigate directly to m.mapquest.com--instead I went directly to mapquest.com. But this doesn't really change anything.

Going directly to m.mapquest.com:

$ http --print hH http://m.mapquest.com User-Agent:"$IOS"
GET / HTTP/1.1
Accept: */*
Accept-Encoding: gzip, deflate, compress
Host: m.mapquest.com
User-Agent: Mozilla/5.0 (iPhone; CPU iPhone OS 6_0 like Mac OS X) AppleWebKit/536.26 (KHTML, like Gecko) Version/6.0 Mobile/10A5376e Safari/8536.25

HTTP/1.1 200 OK
Connection: close
Content-Encoding: gzip
Content-Language: en-US
Content-Type: text/html;charset=ISO-8859-1
Date: Thu, 30 Jan 2014 20:24:09 GMT
Server: Apache-Coyote/1.1
Server: Apache-Coyote/1.1
Set-Cookie: mqdid=0.1.0^!ADV!MQ4Mi^apple_iphone_ver2_sub525181; Expires=Sun, 02-Feb-2014 20:24:10 GMT; Path=/
Set-Cookie: mq_ep_v1=mobileweb_advanced; Path=/
Set-Cookie: JSESSIONID=98C58E9405AB48763E073F8937EC360F; Path=/; HttpOnly

Note the magic cookie: mq_ep_v1=mobileweb_advanced; set for iOS.

Comment 6

[Mike Taylor [:miketaylr]]

I don't know who submitted the mapquest app for Firefox OS, but it seems like it's an even more basic version than the screenshot attached. :/

Unsure who we need to reach out to about that.

Comment 7

[Karl Dubost💡 :karlcow]

just wow for the analysis. I guess it will be helpful for them too.

Comment 8

[Karl Dubost💡 :karlcow]

This might help https://github.com/orgs/MapQuest/members

Comment 9

[Karl Dubost💡 :karlcow]

Contacted the tech account of mapquest on twitter. Thanks to Mike analysis it should not be too hard to get a switch. 
http://twitter.com/MozWebCompat/status/429055525389479936

Comment 10

[gabriele.vidali]

Almost solved: the ui is now correct but I cant switch back to map view after I look for directions
http://imgur.com/V7ehi46

Comment 11

[Hallvord R. M. Steen [:hallvors]]

Tests say fixed, WFM!
The problems with the direction feature and any other remaining issues should be reported in separate bugs. Gabriele, would you mind doing that? If not, I will :)

Comment 12

[gabriele.vidali]

Please go on  and open the new bug.

Comment 13

[Hallvord R. M. Steen [:hallvors]]

No problem, we'll follow up in bug 1001491 - although I haven't gotten started on any analysis yet.