1483499 - <legend> not child of <fieldset> should have 2px padding

Status: RESOLVED FIXED

(Core :: Layout, defect, P3)

Description

[Simon Pieters [:zcorpan]]

WebKit/Chromium/EdgeHTML always have 2px padding. For Gecko html.css says `fieldset > legend` instead of just `legend`.

See https://github.com/whatwg/html/issues/2729

Test: 
http://web-platform-tests.live/html/rendering/non-replaced-elements/the-fieldset-element-0/legend.html

Comment 1

[Daniel Holbert [:dholbert]]

Simon makes a reasonably compelling argument -- the current "fieldset >" qualifier means that we get different amounts of padding depending on whether the legend is inside of "display:contents" or not.

Kind of a strained scenario to design for, but there's some logic to the idea that there should be consistency there -- and given that plus the consensus of other browsers, it seems reasonable to change here.

Comment 2

[Daniel Holbert [:dholbert]]

Here's a testcase with the display:contents scenario: https://jsfiddle.net/xmcpsozh/
If you view that in Firefox and look closely [or use devtools to inspect computed style], you can see that the display:contents-nested legend has less padding than the normal legend.

Comment 3

[Daniel Holbert [:dholbert]]

Mihir might take this -- tentatively assigning this to him.

All we need to do here is:
 (1) Merge this all into a single "legend { ... }" rule (no fieldset > legend):
https://dxr.mozilla.org/mozilla-central/rev/4e56a2f51ad739ca52046723448f3129a58f1666/layout/style/res/forms.css#58-66

 (2) Write a reftest based on my jsfiddle from comment 2, with a display:contents wrapper around <legend> in the testcase and without it in the reference case.

Comment 4

[Daniel Holbert [:dholbert]]

I'll go ahead and close this out. Note that I'll be leaving behind one "fieldset > legend" declaration here -- that one's covered by bug 1485502.

Comment 5

[Daniel Holbert [:dholbert]]

Attachment: Bug 1483499: Give <legend> consistent default padding, regardless of whether it's a child of fieldset. r=mats

Comment 6

[Daniel Holbert [:dholbert]]

Try run w/ reftests, all platforms (to validate that new tests pass everywhere & to catch possible changes in number of fuzzy pixels on platform-specific fuzzy annotations):
  https://treeherder.mozilla.org/#/jobs?repo=try&revision=c739c00aa6d36384cfa2bf66b1c048b242853c88

Try run w/ all layout-relevant tests, linux-only (to catch any other test dependencies on old behavior):
  https://treeherder.mozilla.org/#/jobs?repo=try&revision=4b47c71fd59f319dd46a2d897c8b366a3cb5668f

Comment 7

[Daniel Holbert [:dholbert]]

Hmm, looks like I've got some test expectations to update.

Comment 8

[Mats Palmgren (inactive)]

Comment on attachment 9003318 [details]
Bug 1483499: Give <legend> consistent default padding, regardless of whether it's a child of fieldset. r=mats

Mats Palmgren (:mats) has approved the revision.

Comment 9

[Daniel Holbert [:dholbert]]

https://treeherder.mozilla.org/#/jobs?repo=try&revision=0344ccf32f2808a61fc997f6dbf4880d50283393&selectedJob=195827965

Comment 10

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/fc18abb2ea52
Give <legend> consistent default padding, regardless of whether it's a child of fieldset. r=mats

Comment 11

[Web Platform Test Sync Bot (Matrix: #interop:mozilla.org)]

Created web-platform-tests PR https://github.com/web-platform-tests/wpt/pull/12682 for changes under testing/web-platform/tests

Comment 12

[Web Platform Test Sync Bot (Matrix: #interop:mozilla.org)]

Upstream web-platform-tests status checks passed, PR will merge once commit reaches central.

Comment 13

[Natalia Csoregi [:nataliaCs]]

https://hg.mozilla.org/mozilla-central/rev/fc18abb2ea52

Comment 14

[Web Platform Test Sync Bot (Matrix: #interop:mozilla.org)]

Upstream PR merged

Comment 15

[Cristian Comorasu, QA [:ccomorasu], Releae Desktop QA - Please contact [e-mail]]

Marking this as qe-verify -, per comment 12.