Bug 109681 – Incorrectly applies column-gap between pages in overflow: paged

• Component: CSS
• Status: NEW

[Rune Lillesveen]

For element or root pagination with overflow:paged-*, WebKit applies the column-gap length between individual pages in the paged container. This looks like is a side-effect of using the multicol implementation to paginate. That is an implementation detail and column-gap should not apply in this case.

[Rune Lillesveen]

Created attachment 188055 [details]
Test case for paged-y with column gap

[Rune Lillesveen]

Created attachment 188056 [details]
Test case for paged-y on root with column gap

[Rune Lillesveen]

Created attachment 188300 [details]
Patch

[WebKit Review Bot]

Comment on attachment 188300 [details]
Patch

Attachment 188300 [details] did not pass chromium-ews (chromium-xvfb):
Output: http://queues.webkit.org/results/16514605

New failing tests:
fast/overflow/paged-y-with-column-gap.html
fast/overflow/paged-x-div-with-column-gap.html
fast/overflow/paged-y-on-root-with-column-gap.html
fast/overflow/paged-x-with-column-gap.html
fast/overflow/paged-y-div.html
fast/overflow/paged-x-div.html

[Rune Lillesveen]

Since the expected results changed with my path, I regenerated the "expected" files for gtk since that is the target I'm using. I don't know if I need to regenerate the "expected" files for every build target, if I should remove them, or what.

[Rune Lillesveen]

(In reply to comment #5)
> Since the expected results changed with my path

"patch"

[Beth Dakin]

Comment on attachment 188300 [details]
Patch

Hi Rune! Thanks for your contribution! In the email I sent you, I was worried that this patch would break the Page::setPagination() API, but I actually think it will be fine. Check out this wiki page for layout test failures on other platforms, and let m know if you still have questions about how best to proceed: http://trac.webkit.org/wiki/TestExpectations

[Build Bot]

Comment on attachment 188300 [details]
Patch

Attachment 188300 [details] did not pass mac-ews (mac):
Output: http://queues.webkit.org/results/16530985

New failing tests:
fast/overflow/paged-y-div.html
fast/overflow/paged-x-div-with-column-gap.html
fast/overflow/paged-x-div.html
fast/overflow/paged-x-with-column-gap.html

[Rune Lillesveen]

Created attachment 188569 [details]
Patch

[Eric Seidel (no email)]

Comment on attachment 188569 [details]
Patch

dumpAsText() tests don't generally need rebaselining by other ports.  I would not expect this to fail for others.

[Rune Lillesveen]

(In reply to comment #7)
> (From update of attachment 188300 [details])
> Hi Rune! Thanks for your contribution! In the email I sent you, I was worried that this patch would break the Page::setPagination() API, but I actually think it will be fine.

I've only modified the CSS parts of what affects pagination gap, so unless gap is set via -webkit-column-gap UA styles, it should be fine.

> Check out this wiki page for layout test failures on other platforms, and let m know if you still have questions about how best to proceed: http://trac.webkit.org/wiki/TestExpectations

Wasn't all clear to me if/when I was supposed to do rebaselining for various platforms. I've added "Failure"s in TestExpectations.

[Rune Lillesveen]

(In reply to comment #10)
> (From update of attachment 188569 [details])
> dumpAsText() tests don't generally need rebaselining by other ports.  I would not expect this to fail for others.

The tests I marked as Failures for rebaselining were the old tests which didn't use dumpAsText().

[Rune Lillesveen]

Created attachment 188810 [details]
Patch

[Brent Fulgham]

Comment on attachment 188810 [details]
Patch

Unfortunately, this patch has been neglected for so long that it doesn't apply anymore. Could you please rebase the patch against our current sources so we can properly evaluate it?

[Brent Fulgham]

This continues to be an issue in Safari 15.

[Radar WebKit Bug Importer]

<rdar://problem/96972348>