1212607 - -moz-window-dragging:drag is no longer working as of Firefox 44

Status: RESOLVED FIXED

(Core :: Layout, defect)

Description

[Kevin Grandon :kgrandon]

In bug 1118169 window-dragging was implemented as a CSS hook to allow calling users to drag HTML windows. In recent builds of Firefox it appears to have stopped working. 

In the CSS console, I'm seeing: Unknown property '-moz-window-dragging'.  Declaration dropped.

Comment 1

[Kevin Grandon :kgrandon]

This is possibly missing a patch from the larch branch. Removing the regression keyword until we can be sure.

Comment 2

[Kevin Grandon :kgrandon]

Likely a regression from this commit range: https://github.com/mozilla/gecko-dev/compare/ca77b70a8dec768e6b1e09371ee666d523f9f4b1...399a23b6c1ec9441a9a1b3b9b2bcb4d78fcb0a40

Comment 3

[Mats Palmgren (inactive)]

An effect of bug 1211040 most likely.

Comment 4

[Kevin Grandon :kgrandon]

Yes, looks like this was caused by bug 1211040. 

It appears that bug 1211040 broke -moz-window-dragging usage for Graphene which was being used for b2g desktop like runtimes where the chrome is written in HTML. Can we possibly restore this property, or expose it based on a permission?

Comment 5

[Kevin Grandon :kgrandon]

We set a property here which is supposed to allow us to enable window dragging: http://mxr.mozilla.org/mozilla-central/source/b2g/chrome/content/shell.js#403

webNav.QueryInterface(Ci.nsIDocShell).windowDraggingAllowed = true;

Comment 6

[Xidorn Quan [:xidorn] UTC+10]

Could you provide a link to the code where you use this property?

Comment 7

[Kevin Grandon :kgrandon]

(In reply to Xidorn Quan [:xidorn] (UTC+10, no dst) from comment #6)
> Could you provide a link to the code where you use this property?

It will be assigned to HTML elements which are considered to be the chrome of the application. The application is generally a certified app, and in the past we enabled usage of this property via windowDraggingAllowed. It seems that setting this property no longer works.

This may not be too clear, but here is an example of where it can be set through JS: https://github.com/mozilla/browser.html/blob/287afcd0fc8765152d13a861d478ed8513bfc13a/src/browser/window-bar.js#L17

Comment 8

[Xidorn Quan [:xidorn] UTC+10]

This property should be usable in chrome code, so if the page which uses this is loaded from chrome:// scheme, it ought to work.

However, there are two additional limitations after that bug:
1. the property is not accessible via CSSOM, so you cannot set or get `element.style.mozWindowDragging`;
2. the property isn't parsed when it is used in inline style, so `<div style="-moz-window-dragging: drag">` doesn't work.

But it should work if you use that inside a <style> or a CSS file.

Since the code you provide interact with reflex, which I'm not quite familiar with, I'm not sure which limitation did you meet. From the comment 0, I guess it is the second one. We can probably try to remove the that limitation if that makes sense to your case.

Comment 9

[Kevin Grandon :kgrandon]

(In reply to Xidorn Quan [:xidorn] (UTC+10, no dst) from comment #8)
> This property should be usable in chrome code, so if the page which uses
> this is loaded from chrome:// scheme, it ought to work.
> 
> However, there are two additional limitations after that bug:
> 1. the property is not accessible via CSSOM, so you cannot set or get
> `element.style.mozWindowDragging`;
> 2. the property isn't parsed when it is used in inline style, so `<div
> style="-moz-window-dragging: drag">` doesn't work.
> 
> But it should work if you use that inside a <style> or a CSS file.
> 
> Since the code you provide interact with reflex, which I'm not quite
> familiar with, I'm not sure which limitation did you meet. From the comment
> 0, I guess it is the second one. We can probably try to remove the that
> limitation if that makes sense to your case.

With react we're running into the first limitation most likely. What was the reasoning for introducing it?

In addition, we need to be able to access this from non-chrome code. As I said, previously this was allowed from the windowDraggingAllowed docshell property in non-chrome code. Can we restore that behavior and allow window dragging in non-chrome code if windowDraggingAllowed is set? If not, can we allow it for non-chrome code in certified apps? I believe only chrome privilieged code should be able to set the windowDraggingAllowed property.

Comment 10

[Kevin Grandon :kgrandon]

Please see this thread for more context: https://bugzilla.mozilla.org/show_bug.cgi?id=1118169#c13

Comment 11

[Xidorn Quan [:xidorn] UTC+10]

If it is useless in content even if it is exposed... I guess it's ok to restore the behavior. dbaron, WDYT?

Comment 12

[David Baron :dbaron: (⌚️UTC-4, no longer working on Mozilla)]

(In reply to Xidorn Quan [:xidorn] (UTC+10, no dst) from comment #11)
> If it is useless in content even if it is exposed...

I don't understand.  In what sense is it useless in content?  (If so, how was there a regression here?)

Or do you need to revisit bug 1069192 comment 12 and conditionally expose the DOM properties when the property works?

Comment 13

[Kevin Grandon :kgrandon]

> I don't understand.  In what sense is it useless in content?  (If so, how was there a regression here?)

I think it's useless in content unless the docshell is flagged from privileged code. In most content that will never be the case, so the property will be useless in those cases.

Xidorn, please correct me if that's not the case.

Is backing bug 1211040 out possible? I hate to ask when there's no tests that caught this, but I will file a bug to write a test for moz-window-dragging and the Graphene runtime.

Comment 14

[David Baron :dbaron: (⌚️UTC-4, no longer working on Mozilla)]

Yeah, I think backing out is fine.  (Or should it be half backed out?)

Comment 15

[Xidorn Quan [:xidorn] UTC+10]

(In reply to David Baron [:dbaron] ⌚UTC-7 from comment #12)
> (In reply to Xidorn Quan [:xidorn] (UTC+10, no dst) from comment #11)
> > If it is useless in content even if it is exposed...
> 
> I don't understand.  In what sense is it useless in content?  (If so, how
> was there a regression here?)

As kgrandon said in comment 13, they use a flag in docshell to control the effectiveness of this property, and docshell is only accessible via the chrome code, so it is useless for general web content.

> Or do you need to revisit bug 1069192 comment 12 and conditionally expose
> the DOM properties when the property works?

They said they want to allow it in not only chrome code, but also certificated apps, so conditionally exposing the DOM property probably won't work for them either. (see last paragraph in comment 9)

(In reply to David Baron [:dbaron] ⌚UTC-7 from comment #14)
> Yeah, I think backing out is fine.  (Or should it be half backed out?)

I'm going to have a patch to re-expose this single property. I don't see any reason for re-exposing the other property in that patch.

Comment 16

[Xidorn Quan [:xidorn] UTC+10]

Attachment: MozReview Request: Bug 1212607 - Re-expose -moz-window-dragging property to the content.

Bug 1212607 - Re-expose -moz-window-dragging property to the content.

Comment 17

[David Baron :dbaron: (⌚️UTC-4, no longer working on Mozilla)]

Comment on attachment 8671691 [details]
MozReview Request: Bug 1212607 - Re-expose -moz-window-dragging property to the content.

https://reviewboard.mozilla.org/r/21631/#review19435

Comment 18

[Xidorn Quan [:xidorn] UTC+10]

https://hg.mozilla.org/integration/mozilla-inbound/rev/40f2d8c486bfdc6201a5bab9131a833d0c60269f
Bug 1212607 - Re-expose -moz-window-dragging property to the content. r=dbaron

Comment 19

[Kohei Yoshino]

Updated the site compatibility doc accordingly. https://www.fxsitecompat.com/en-US/docs/2015/several-internal-css-properties-have-been-removed/

Comment 20

[Carsten Book [:Tomcat]]

https://hg.mozilla.org/mozilla-central/rev/40f2d8c486bf