wikimedia / parsoid / 2278

85%

push

travis-ci

<a href="https://github.com/wikimedia/parsoid/commit/b652e479a">PHP section numbers are assigned during tokenization

The PHP Parser assigns section numbers during the preprocessing phase,
so any valid heading in the input gets a number, even if it&#39;s inside
a template argument which may (or may not) get expanded.  Then it only
recognizes the subset of these sections which occur at top-level.
Match this logic by maintaining the section number index in the tokenizer.

Parser tests updated &amp; improved: `html/php+tidy` sections to document
and validate the behavior of the PHP parser, especially important as we
get further into corner cases here.  The `notoc` option was used so that
the PHP-side tests don&#39;t have unnecessary clutter and better match the
Parsoid output.  You&#39;re meant to check that the href attribute in
the &quot;edit section&quot; link has the same section= value as Parsoid&#39;s
data-mw-section attribute (unless Parsoid gives a negative value).

Empty data-parsoid attributes were also added in appropriate places to
the html/parsoid clauses in order to make html2wt tests pass and
remove them from the blacklist.

Finally, use explicit `1=` syntax for template arguments, to avoid
unexpected interactions of heading syntax and template argument
syntax.  In particular, the level-1 heading `=foo=` gets parsed
as a named template argument `=foo`, which tends to confuse things.
Since idiosyncracies in template argument parsing is not what these
tests are intended to target, use explicit syntax to avoid the
potential issue.

Bug: T213468
Change-Id: I</a><a class="double-link" href="https://github.com/wikimedia/parsoid/commit/f4a95d7e7">f4a95d7e7</a>

9403 of 11816 branches covered (79.58%)

14816 of 17566 relevant lines covered (84.34%)

81077.2 hits per line