Improve Gherkin support

[Hirse]

Resolves #3638 Resolves #3639

Changes

• Move keywords to mode to support keywords with spaces and those ending with a colon
• Update list of keywords (see below)
• Change * from symbol to keyword since that is how it's described in the reference
• Tighten variable definition to disallow whitespace
• Only recognize keywords, comments, tags, and docstrings at the beginning of a line
• Support docstrings starting with triple backticks
• Remove (non-docstring) string mode

Reference

• https://cucumber.io/docs/gherkin/reference
• https://cucumber.io/docs/gherkin/languages/#gherkin-dialect-en-content

Checklist

• [X] Added markup tests
• [X] Updated the changelog at CHANGES.md

[pkufranky]

thanks very much

[joshgoebel]

Technically all features (except variables) are only valid at the beginning of a line.

Can't we do this with multi-match and ^ (start line)? Just use the first part of the match to grab the whitespace...

[joshgoebel]

and those ending with a colon

Is the colon part of the keyword though?

[Hirse]

Updated with a new version that removes the custom keyword pattern and fixes false positives.

I couldn't figure out a good way to match for "beginning of line, skipping the whitespace" as /^[ \t]*/ triggers the "zero-width regex" error so there's currently a bunch of duplication.

[pkufranky]

Any progress for this pr?

[joshgoebel]

@Hirse is this ready to go? sorry for long wait. coud you update changes.md?