HTML parser: handle </br> and </p> in SVG

[zcorpan]

Fixes #5113.

• At least two implementers are interested (and none opposed):
  • Already implemented in Chromium
  • Already implemented in Gecko (fragment case tests should pass when https://bugzilla.mozilla.org/show_bug.cgi?id=1205631 is fixed)
  • WebKit: @gsnedders Unmatched </p> or </br> inside foreign context needs a special parser rule #5113 (comment)
• Tests are written and can be reviewed and commented upon at:
  • Test </p> and </br> in SVG html5lib/html5lib-tests#135
  • Update tests/testdata (Test </p> and </br> in SVG) html5lib/html5lib-python#534
• Implementation bugs are filed:
  • Chrome: https://bugs.chromium.org/p/chromium/issues/detail?id=1005713 (fixed Dec 2, 2019)
  • Firefox: N/A
  • Safari: https://bugs.webkit.org/show_bug.cgi?id=226915

(See WHATWG Working Mode: Changes for more details.)

---

/parsing.html ( diff )

[zcorpan]

Together with #6455, I believe this matches current Chromium. @mfreed7 can you confirm?

[mfreed7]

I think that's right. In Chromium, <table><math><p>foo gives me <math><p><table> in DOM. That's the correct check, right?

[zcorpan]

For #6455, yes. This PR is about these cases:

http://software.hixie.ch/utilities/js/live-dom-viewer/saved/9362

<svg></p><foo>

should produce:

• html html
  • html head
  • html body
    • svg svg
    • html p
    • html foo

and

<svg></br><foo>

should produce:

• html html
  • html head
  • html body
    • svg svg
    • html br
    • html foo

[mfreed7]

Thanks. I can confirm that behavior in Chromium also. 😄

[zcorpan]

This matches Gecko as well for the non-fragment case. (The fragment case is different because Gecko doesn't yet implement #6399 + #6455.) cc @hsivonen

@gsnedders does this comment #5113 (comment) indicate implementer interest for WebKit?

[domenic]

Simon, should we merge this? Aligning the spec with 2/3 implementations seems good!

[zcorpan]

I think it's good to merge, yeah