1379834 - Window context bug in target="_blank" and window.open()

Status: NEW

(DevTools :: Responsive Design Mode, defect, P3)

Description

[Cristian Adrian Giagante]

Attachment: firefoxBug.html

User Agent: Mozilla/5.0 (Windows NT 10.0; WOW64; rv:54.0) Gecko/20100101 Firefox/54.0
Build ID: 20170628075643

Steps to reproduce:

ONLINE FILE EXAMPLE: http://muwlauncher.com/firefoxBug.html

STEPS TO REPRODUCE THE BUG:

1-PRESS THE SUBMIT BUTTON IN DESKTOP MODE AND ENABLE POPUP WINDOWS IF IT'S NEEDED, WHEN THE
NEW WINDOW WILL BE OPEN YOU CAN SEE THE POST VARS SENDED myData WITH VALUE important_data.

2-NOW OPEN DEVTOOLS AND SELECT RESPONSIVE DESIGN MODE (CTRL+SHIFT+M) AND PRESS THE
SUBMIT BUTTON, NOW YOU CAN SEE THAT THE POST VARIABLES DONT WAS SENDED (THE FORM METHOD ACTION POST WAS IGNORED), THE REQUEST WAS SENDED IN GET MODE.

3-NOW TRY TO RUN THIS IN CONSOLE IN DESKTOP MODE (NO RESPONSIVE DESIGN MODE)
var firefoxBug = window.open("","testBug") 
(enabled popup window if it's needed when firefox ask to it), a new windows is created,
back to your original window and run in console: firefoxBug (you can see the window object)

4-NOW TRY THE 3 STEP IN MOBILE MODE, NULL WILL BE RETURNED ALWAYS.




Actual results:

You lost the window context info, lost action method info from forms, and data variables, always "null" returned trying to use window.open() (with popup blocker disabled)

THIS BUG AFFECT DEKSTOP FIREFOX AND MOBILE FIREFOX IN REAL DEVICES


Expected results:

The mobile mode and real mobile devices must respect the normal behavior on desktop, the action method must be the original post not changed to get, and javascript window.open() must return a window object to control a new window generated.

Comment 1

[Richard Newman [:rnewman]]

I can reproduce this on current Nightly. window.open always returns null in Responsive Design Mode.

Comment 2

[J. Ryan Stinnett [:jryans] (Use needinfo, replies may be slow)]

Has someone checked this on a real mobile device? (Comment 0 says "MOBILE FIREFOX IN REAL DEVICES", but I am not sure if that was assumed because it happened in Responsive Design Mode on desktop.)

It's entirely possible that Responsive Design Mode is the issue here, assuming it only happens in there.

Comment 3

[Cristian Adrian Giagante]

Firefox IOS version has this bug (tested on iphone 6S plus)
Firefox for Android dont has this bug (tested on samsung s7)

Comment 4

[Cristian Adrian Giagante]

A few hours ago the responsive design editor changed on firefox desktop (hotfix?) to a black style and the issue is not being replied here.
But the issue on browser stack in firefox for ios is still happen, someone knows if the hotfix affect firefox for ios too?.

Comment 5

[J. Ryan Stinnett [:jryans] (Use needinfo, replies may be slow)]

(In reply to Cristian Adrian Giagante from comment #4)
> A few hours ago the responsive design editor changed on firefox desktop
> (hotfix?) to a black style and the issue is not being replied here.
> But the issue on browser stack in firefox for ios is still happen, someone
> knows if the hotfix affect firefox for ios too?.

If the RDM changed to a black appearance, it likely means you are seeing the old RDM because e10s is disabled for some reason.
Check about about:support's "Multiprocess Windows" entry to see why that might have happened.

Comment 6

[Cristian Adrian Giagante]

I see that the environment is changing all time, Multiprocess Windows is being enable and disable but the new RDM is still bugged.
If someone need help to test can mail me or answer here. Greetings and thanks community

Comment 7

[J. Ryan Stinnett [:jryans] (Use needinfo, replies may be slow)]

(In reply to Cristian Adrian Giagante from comment #4)
> But the issue on browser stack in firefox for ios is still happen

What test case are you using to check Firefox for iOS?  The steps in comment 0 are talking about RDM on desktop.

Comment 8

[Richard Newman [:rnewman]]

(In reply to Cristian Adrian Giagante from comment #3)
> Firefox IOS version has this bug (tested on iphone 6S plus)
> Firefox for Android dont has this bug (tested on samsung s7)

Firefox for iOS doesn't use Gecko; if you see this behavior there, that warrants a new bug in that product.

Comment 9

[Cristian Adrian Giagante]

(In reply to J. Ryan Stinnett [:jryans] (use ni?) from comment #7)
> (In reply to Cristian Adrian Giagante from comment #4)
> > But the issue on browser stack in firefox for ios is still happen
> 
> What test case are you using to check Firefox for iOS?  The steps in comment
> 0 are talking about RDM on desktop.

I'm testing on browserstack in real device. Using my example on http://muwlauncher.com/firefoxBug when you load the site, and click on the submit button you dont get the post vars in the response. Same issue than in desktop when RDM is being used.

Comment 10

[Cristian Adrian Giagante]

(In reply to Richard Newman [:rnewman] from comment #8)
> (In reply to Cristian Adrian Giagante from comment #3)
> > Firefox IOS version has this bug (tested on iphone 6S plus)
> > Firefox for Android dont has this bug (tested on samsung s7)
> 
> Firefox for iOS doesn't use Gecko; if you see this behavior there, that
> warrants a new bug in that product.

i thinks is a little weird that the same issue is present on the real device iphone on firefox ios, could be possible than the RMA issue is related with firefox ios issue? is exactly the same.

Comment 11

[Richard Newman [:rnewman]]

(In reply to Cristian Adrian Giagante from comment #10)

> i thinks is a little weird that the same issue is present on the real device
> iphone on firefox ios, could be possible than the RMA issue is related with
> firefox ios issue? is exactly the same.

There is essentially no shared code between Firefox on desktop and Firefox for iOS. The latter uses Apple's WKWebView.

It's possible that there are two bugs with similar root causes -- most likely that somewhere we're mis-handling a handoff to a new tab -- but it's vanishingly unlikely that there is a single code bug that is causing both the behavior in RDM and in Firefox for iOS.


(In reply to J. Ryan Stinnett [:jryans] (use ni?) from comment #7)

> What test case are you using to check Firefox for iOS?  The steps in comment
> 0 are talking about RDM on desktop.

I can also reproduce this issue manually in Firefox for iOS on a real device.

Comment 12

[Richard Newman [:rnewman]]

I filed Bug 1380202 for iOS. We can scope this down to just RDM on desktop.