commit bc34c4e1b5de52bec36a83548810f0be3c72ed19
author Richard Townsend <richard.townsend@arm.com> Wed May 27 10:38:12 2020
committer Commit Bot <commit-bot@chromium.org> Wed May 27 10:38:12 2020
tree 657ed6875ed7b44e55699ec8938a95256b067963
parent 5cca9894cd09aa1b614849d262986b5c7568c523

Implement synchronous budgeted foreground HTML parser

Previously, there were two main ways to parse HTML:
* The HTMLDocumentParser's PumpTokenizer method, used for HTML
  fragment parsing (foreground parser)
* The BackgroundHTMLParser, used for page loading (background parser).

This CL changes HTMLDocumentParser::PumpTokenizer so that it can yield
to other tasks during page loading, adding a third way of parsing
(synchronous, budgeted foreground HTML parsing). It does this either
when a really important page loading task needs to run, or when a
(budget-controlled) chunk of HTML has been processed. Because it can
be interrupted, it can replace the BackgroundHTMLParser's
functionality whilst being less complicated and offering a faster
time-to-first paint.

Enable with --enable-blink-features=ForceSynchronousHTMLParsing

Design doc:
https://docs.google.com/document/d/1ecupQAUrvnAxpGmAitrudaLFd0jN7DAf5dgYSpXFoqI

Bug: 901056
Change-Id: I14b8c4bede28479911951181f85395eb617b8b16
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1333755
Reviewed-by: Kouhei Ueno <kouhei@chromium.org>
Reviewed-by: Mason Freed <masonfreed@chromium.org>
Commit-Queue: Richard Townsend <richard.townsend@arm.com>
Cr-Commit-Position: refs/heads/master@{#772182}