363637 - {inc}abs pos frame in rel pos RTL block has right:auto at left edge

Status: RESOLVED FIXED

(Core :: Layout: Positioned, defect)

Description

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

In bug 328181 comment 17, Uri wrote (with correction rolled in):

However, the reflow branch also introduced a new regression: in the RTL lines
(3,4) of the "Absolute" and "Absolute, margins" sections of all three
testcases, the inner div is rendered outside the outer div, to its left.
I tried tracking this down, but I don't have enough time to properly learn the
code involved and understand where exactly the problem is.
(I suspect it might be due to GetNearestContainingBlock() returning a bogus
aContentArea, with a negative |right|, but I'm not really sure).

Steps to reproduce: load attachment 232963 [details]

Expected results:  All 4 inner boxes visible in "Absolute" and "Absolute, margins" sections

Actual results: Two of the boxes are off the left edge of the screen.


I actually don't see how this ever worked unless we were doing extra reflows:  CalculateHypotheticalBox calls GetNearestContainingBlock, which calls aFrame->GetSize(), but aFrame hasn't had its reflow finished yet, so its size hasn't been set yet.

Comment 1

[Boris Zbarsky [:bzbarsky]]

Attachment: This fixes the problem for me

Comment 2

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

Comment on attachment 248688 [details] [diff] [review]
This fixes the problem for me

>+    /* XXXbz Could we assert that aFrame doesn't have NS_FRAME_IN_REFLOW set? */

Please do.


r+sr=dbaron

Comment 3

[Boris Zbarsky [:bzbarsky]]

Attachment: With comments addressed and reftest testcase.

Comment 4

[Boris Zbarsky [:bzbarsky]]

Fixed, and added to reftest.

Comment 5

[Boris Zbarsky [:bzbarsky]]

Attachment: Header patch

This is also needed for this to work.  ;)

Comment 6

[(no longer active)]

Mass-assigning the new rtl keyword to RTL-related (see bug 349193).