Syntax highlighting YAML blocks within Pandoc Markdown files

Syntaxes & Themes
,
1

I’m trying to write a syntax for Pandoc Markdown files. My articles in Pandoc Markdown always contain YAML blocks fenced by a pair of triple hyphens ---. Sublime Text syntax-highlights YAML keys and list hyphens nicely within the metadata block.

In the Drafts syntax file though, I haven’t been able to figure out a regex to match YAML keys and list hyphens only within metadata blocks. What I’m looking for is a regex like:

((\w+)(: )(.*))|((\w+)(:)(\s\- )(.*))

But what this regex does is to match any instance of key: value or key: \n- value anywhere in the document. I want it only to match instances within the YAML header. Is this even possible?

1 Like
2

A bit tricky, but possible. I would create a rule that matches the key: value, but also looks ahead to only only match if there is a --- out ahead of it.

These example example patterns:

[
  {
    "match": "(?s)(\\w+: )([^\\n]+)(?=.*^---)",
    "exclusive": false,
    "comment": "key: value pairs in front matter",
    "captures": {
      "1": {
        "scope": "color.red"
      },
      "2": {
        "scope": "color.green"
      }
    }
  },
  {
    "match": "(?s)(\\w+:\\n)(- )([^\\n]+)(?=.*^---)",
    "exclusive": false,
    "comment": "key: value pairs in front matter",
    "captures": {
      "1": {
        "scope": "color.red"
      },
      "2": {
        "scope": "markup"
      },
      "3": {
        "scope": "color.green"
      }
    }
  }
]

Create results like:

CleanShot 2022-03-18 at 09.40.37@2x

1 Like
3

Please share this when you have it. YAML metadata blocks are also supported in GitLab-flavored Markdown and Multimarkdown. I regularly include metadata and it would be very useful to have it highlighted.

1 Like
4

Thanks for this idea! When I plugged the first regex into https://regex101.com/, it threw a “catastrophic backtracking error” at me, so I made one of the quantifiers possessive. Specifically, I had to make

(?s)(\\w+: )([^\\n]+)(?=.*^---)
into
(?s)(\\w+: )([^\\n]++)(?=.*^---)

for it to work. That seemed to do the trick. Except that I then noticed that if I manipulated the string and then re-formed it, the syntax highlighting got messed up. A screen recording to demonstrate:

Screen Recording 2022-03-18 at 10.10.32 PM

Any idea what could be going on? Here are my patterns:


    {
    "match": "^(---\\n)[\\s\\S]+(---\\n)",
    "exclusive": false,
    "comment": "YAML front matter",
    "captures": {
      "0": {
        "scope": "code.block"
        },
      }
    },
    {
    "match": "(?s)(\\w+: )([^\\n]++)(?=.*^---)",
    "exclusive": false,
    "comment": "key: value pairs in YAML front matter",
    "captures": {
      "1": {
        "scope": "code.literal"
      },
      "2": {
        "scope": "code.block"
        }
      }
    },
    {
    "match": "(?s)(\\w+:\\n- )([^\\n]++)(?=.*^---)",
    "exclusive": false,
    "comment": "key: - value pairs in YAML front matter",
    "captures": {
      "1": {
        "scope": "code.literal"
      },
      "2": {
        "scope": "code.block"
        }
      }
    },

…keeping in mind that I have created a theme to go along with this syntax. Thanks!

5

Well, no replies to this, so I suppose that it must be an issue with Drafts itself. My workaround will be to put the YAML header into a form that’s stably highlightable by Drafts. I’ll have to write a script to normalise the syntax before saving it for processing by Pandoc.

6

What do you have the rangeExtensionType set as in your syntax? If it is anything other than fullText, maybe try changing it to that and giving it another try.

1 Like
7

Wow, @sylumer! Setting rangeExtensionType to fullText totally worked! I have no idea why, though. Anyway, I’m looking forward to developing this syntax more fully now. Thank you!!! :champagne:

8

Read the definition for it in the article on creating syntaxes. It sets out the reasonin that definition.

1 Like