md2html: A Toy Markdown-Subset Parser
Setup: npm i
Building: npm run-script build
Testing: npm test
Running: node md2html-cli <file>
Tests are located in spec/md2htmlSpec.js.
The basic outline is to tokenize the input, create a parse tree from the tokens, and then format the parse tree into the appropriate output. This is a simple abstracted flow but keeps the full parsed document in memory, limiting its ability to operate on very large documents. A quick check suggests that it will parse 4MB in about one second using 160MB of memory (40MB per 1MB of input).
Whitespace is not preserved, which results in a single space around elements
even when the input has none. For example (link)[url]. becomes
<a href="url">link</a> . (note the space after the <a> tag.)
Features
Only a few syntaxes are supported:
###-style headers- links
- multi-line paragraphs