wikimedia / parsoid / 2160

85%

push

travis-ci

Tweak QuoteTransformer code + add edgy test specing prevToken arg

* String context preceding a bold token is used in some cases
  to convert it to a "'" and an italic.

* In the scenario where the first quote token on a line is a bold token,
  and the prevToken is a string, the prevToken arg passed in by
  the SyncTTM plays a role in deciding which bold token on the line
  is converted.

  Added a test that specs this behavior. The test fails if prevToken
  arg is removed from the signature. The test is decidedly edgy and
  it is unclear how important this support is, but we can make that
  decision at a later time.

* There might be other ways to examine surrounding context (for example,
  by looking at the tsr value of the bold token and extracing a 2-char
  substring preceding it from env.page.src, but this strategy runs into
  difficult for content from templates whose tsr values are stripped.

  But, this can be solved in the PEG tokenizer by storing the preceding
  two-char string for bold tokens.

  But, this for later. For this patch, the test captures the
  expectation.

* Also simplified the code that sets up the string context - the loop
  is unnecessary.

Change-Id: If0108e329

9321 of 11728 branches covered (79.48%)

15019 of 17730 relevant lines covered (84.71%)

79924.65 hits per line