wikimedia / parsoid / 3242 / 2

85%

master: 85%

Build Type

push

travis-ci

Committed by jenkins-bot

Commit Message

<a href="https://github.com/wikimedia/parsoid/commit/3ab058e47">Fix autolink url parsing code

* The original check imported from Parsoid/JS did not match the legacy
  parser, which used a &quot;unicode \w&quot; implicitly in the word-boundary
  checks.  This may not have been entirely intentional or desirable: see
  https://phabricator.wikimedia.org/T238022#5665580

* The Parsoid/JS version of ConstrainedText/AutoURLLinkText suffered the
  same issue.

* Added new helper methods to get the last unicode character (in both PHP
  and JS) and test it for membership in &quot;unicode \w&quot;.  This test is
  extremely awkward in node v6 (unless you use the --harmony_regexp_property
  flag), but since Parsoid/JS is going away anyway, just grit your teeth
  and skim past the huge regexp (generated by regexpu) for now.

* Added a new parser test that passes with both Parsoid/PHP and
  Parsoid/JS.  It tests 2-, 3-, and 4-byte UTF-8 sequences before
  the autolink.

There may be other remaining instances of \w with the &#39;u&#39; regexp modifier
where Parsoid/JS doesn&#39;t match Parsoid/PHP; see T238022 for details.

Bug: T238022
Co-authored-by: C. Scott Ananian
Co-authored-by: Subramanya Sastry
Change-Id: I</a><a class="double-link" href="https://github.com/wikimedia/parsoid/commit/705d2fda4">705d2fda4</a>

Run Details

9662 of 12132 branches covered (79.64%)

15149 of 17854 relevant lines covered (84.85%)

32862.57 hits per line