454361 - "ASSERTION: mOriginalDisplay has not been properly initialized"

Status: RESOLVED FIXED

(Core :: Layout: Positioned, defect)

Description

[Jesse Ruderman]

Attachment: testcase

This fairly simple testcase triggers:

###!!! ASSERTION: mOriginalDisplay has not been properly initialized: 'mStyleDisplay->mOriginalDisplay != NS_STYLE_DISPLAY_NONE', file /Users/jruderman/central/layout/generic/nsHTMLReflowState.cpp, line 882

Regression from bug 243519, perhaps?

Comment 1

[Mats Palmgren (inactive)]

Attachment: Testcase #2

Comment 2

[Mats Palmgren (inactive)]

We create (and reflow) a frame for the root content even though
the style context has mDisplay == NS_STYLE_DISPLAY_NONE.

My gut feeling is that Opera and Safari is correct in not displaying
the text in the second testcase, but I haven't checked the CSS spec.
Does anyone know?

Comment 3

[Mats Palmgren (inactive)]

Attachment: Like so?

Comment 4

[Mats Palmgren (inactive)]

http://www.w3.org/TR/CSS21/visuren.html#propdef-display
"Descendant elements do not generate any boxes either;" and a little later:
"For the root element, the computed value is changed as described in the section on the relationships between 'display', 'position', and 'float'." (9.7)

http://www.w3.org/TR/CSS21/visuren.html#dis-pos-flo
"1. If 'display' has the value 'none', then 'position' and 'float' do not apply. In this case, the element generates no box"

So I think it's clear that we should not display the text.

Comment 5

[Robert O'Callahan (:roc) (email my personal email if necessary)]

Comment on attachment 338009 [details] [diff] [review]
Like so?

looks good to me

Comment 6

[Mats Palmgren (inactive)]

Attachment: reftest.diff

Comment 7

[Mats Palmgren (inactive)]

http://hg.mozilla.org/mozilla-central/rev/3cc51edbb1c9
http://hg.mozilla.org/mozilla-central/rev/9917bfc85075

-> FIXED

Comment 8

[David Baron :dbaron:]

Shouldn't this have added the root element to the undisplayed map so that if its 'display' property were changed dynamically, a frame would get created?  Or does that work some other way?

Comment 9

[Mats Palmgren (inactive)]

That triggers:
###!!! ASSERTION: undisplayed content must have a parent: 'parent', file /usr/moz/hg3/layout/base/nsFrameManager.cpp, line 581

The frame manager stores the undisplayed map using the parent pointer as a
hash key.  It's bug 283686 I think.

Comment 10

[Ria Klaassen (not reading all bugmail)]

Could this bug have caused a serious scroll delay on http://westseattleblog.com/blog/?page_id=432 (testcase from Bug 484572)?

Comment 11

[Robert O'Callahan (:roc) (email my personal email if necessary)]

Unlikely