Bug 95203 – Dotted borders not drawn with rounded dots

• Component: Layout and Rendering
• Status: ASSIGNED

[Martin Leutelt]

As per CSS spec here http://www.w3.org/TR/css3-background/#border-style dotted borders should be drawn
with rounded dots.

Websites for testing:
http://www.w3schools.com/cssref/tryit.asp?filename=trycss_border-style
http://dev.harnos.at/webkit/border-bug/

[Martin Leutelt]

Created attachment 160987 [details]
Proposed patch

[Martin Robinson]

I think when evaluating this patch it would be very useful to see how it affects pixel results. Many pixel results are out of date so here's my suggestion:

1. Run pixel tests and update all results.
2. Run pixel results again and detect which results change.
3. Attach only results that changed to your patch.

Be sure to run tests with the jhbuild environment by running

$ Tools/Scripts/update-webkitgtk-libs

before you build.

[Martin Leutelt]

Created attachment 161671 [details]
Patch with updated expected results

[Martin Robinson]

A couple things worry me about this patch:

1. It doesn't look like any other WebKit browsers that I tested (Chromium and Safari on Mac)  and Firefox use rounded dots.
2. This patch changes the behavior for content not rendered according to the CSS spec as well.

[Martin Leutelt]

Thanks for your comments.

on 1: For Chromium I already submitted a patch (https://bugs.webkit.org/show_bug.cgi?id=94917). I'm waiting on some comments there. 
The patch for the Qt port has already been landed (https://bugs.webkit.org/show_bug.cgi?id=92383).
I can't build the mac port so I won't be able to submit a patch for that. To track all those patches I guess we should use this one (https://bugs.webkit.org/show_bug.cgi?id=25737).
Firefox uses rounded dots but also not according to the css spec.

on 2: Could you give me some example for a case where the patch would lead to some unexpected outcome?

[Anders Carlsson]

Comment on attachment 161671 [details]
Patch with updated expected results

Clearing review flag on patches from before 2014. If this patch is still relevant, please reset the r? flag.

[Myles C. Maxfield]

<rdar://problem/3175114>

[Simon Fraser (smfr)]

I don't think this is particularly easy