988653 - 'border-image' layout is wrong for fragmented boxes

Status: VERIFIED FIXED

(Core :: Layout: Block and Inline, defect, P4)

Description

[Mats Palmgren (inactive)]

Attachment: Testcase #1

It looks like the current layout would be correct for 'box-decoration-break:clone'
but the default should be 'slice'.

STR
Load the attached testcase.

ACTUAL RESULTS
Borders are drawn on all sides.

EXPECTED RESULTS
For the fragmented block at the top: there should be no border at the bottom edge
in the first column, nor at the top of the second column.  The left and right
borders of the second fragment should render as if continued below the first
fragment.
For the inline element at the bottom: there should be no right border on the
"a" fragment, and no left border on the "b" fragment.  The top/bottom border
of "b" should render as if it were part of "a", on the right side.

I think bug 613659 provides all supporting code for fixing this bug.

Comment 1

[Mats Palmgren (inactive)]

Attachment: Testcase #1, v2

Comment 2

[Mats Palmgren (inactive)]

Attachment: Desired rendering of Testcase #1

Comment 3

[Mats Palmgren (inactive)]

Attachment: WIP

https://tbpl.mozilla.org/?tree=Try&rev=dc64e822b399

Comment 4

[Mats Palmgren (inactive)]

(note to self: cyan/lime is hard to tell apart for small sizes)

Comment 5

[Mats Palmgren (inactive)]

Attachment: fix + reftest

Pretty straight-forward fix.  The hard part was writing the test; it seems
we have some pixel-rounding differences between platforms in the border-image
or GFX code.  The b-d-b Layout code is platform agnostic so I think
restricting the test to Linux is good enough to test that.

https://tbpl.mozilla.org/?tree=Try&rev=9d90ca00606b
https://tbpl.mozilla.org/?tree=Try&rev=594647d332f9

Comment 6

[Cameron McCormack (:heycam)]

Comment on attachment 8398842 [details] [diff] [review]
fix + reftest

Can you please clean up the inconsistent use of (and the few instances of trailing) white space in the <style> elements of the test/reference files.  It's rather messy. :)

Comment 7

[Mats Palmgren (inactive)]

Cleaned up the test and pushed:
https://hg.mozilla.org/integration/mozilla-inbound/rev/8a8c87fcd2d9

Comment 8

[Phil Ringnalda (:philor)]

https://hg.mozilla.org/mozilla-central/rev/8a8c87fcd2d9

Comment 9

[Mats Palmgren (inactive)]

Backed out on m-c:
remote:   https://hg.mozilla.org/mozilla-central/rev/581208970e99
remote:   https://hg.mozilla.org/mozilla-central/rev/ea214ef03c3f

Comment 10

[Mats Palmgren (inactive)]

https://hg.mozilla.org/integration/mozilla-inbound/rev/75cafc47ff36

Comment 11

[Wes Kocher (:KWierso) (Not reading bugmail; email directly if needed)]

https://hg.mozilla.org/mozilla-central/rev/75cafc47ff36

Comment 12

[Paul Silaghi, QA [:pauly]]

Verified fixed 32.0a1 (2014-05-08), Win 7 x64