Skip to content

Language Detection

The parser used for a document is chosen by its language ID (given by the editor). If the language ID is empty or plaintext, the file name is used instead. If the language ID isn't known, comment markers the editor knows for the language are used, so that just comments are wrapped.

language: "rewrap-test-unknown", path: "/a/b.xyz", lineComment: "#"

# a        ¦      ->      # a b      ¦
# b        ¦              x          ¦
x          ¦              y          ¦
y          ¦

Markdown

Editors give some kinds of Markdown file their own language ID. For example VS Code uses prompt for .prompt.md files and instructions for copilot-instructions.md (#46). Files with a Markdown extension and an unknown language ID are still treated as Markdown, even if the editor knows comment markers for the language.

language: "prompt", path: "/a/b.prompt.md", blockComment: ""

# Title    ¦      ->      # Title    ¦
a b c d e f g             a b c d e f¦
h          ¦              g h        ¦

language: "instructions", path: "/a/.github/copilot-instructions.md", blockComment: ""

# Title    ¦      ->      # Title    ¦
a b c d e f g             a b c d e f¦
h          ¦              g h        ¦

Quarto and R Markdown documents are Markdown too (#54, #40).

language: "quarto", path: "/a/b.qmd", blockComment: ""

# Title    ¦      ->      # Title    ¦
a b c d e f g             a b c d e f¦
h          ¦              g h        ¦

language: "rmd", path: "/a/b.Rmd", blockComment: ""

# Title    ¦      ->      # Title    ¦
a b c d e f g             a b c d e f¦
h          ¦              g h        ¦

This also works for new documents that don't have a file name yet.

language: "quarto", blockComment: ""

# Title    ¦      ->      # Title    ¦
a b c d e f g             a b c d e f¦
h          ¦              g h        ¦

language: "rmd", blockComment: ""

# Title    ¦      ->      # Title    ¦
a b c d e f g             a b c d e f¦
h          ¦              g h        ¦