Markdown parser, done right. 100% CommonMark support, extensions, syntax plugins & high speed
https://markdown-it.github.io/
You can not select more than 25 topics
Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
110 lines
2.5 KiB
110 lines
2.5 KiB
#!/usr/bin/env node
|
|
/*eslint no-console:0*/
|
|
|
|
// Fixtures generator from commonmark specs. Split spec to working / not working
|
|
// examples, or show total stat.
|
|
|
|
'use strict';
|
|
|
|
|
|
var fs = require('fs');
|
|
var util = require('util');
|
|
var argparse = require('argparse');
|
|
|
|
|
|
var cli = new argparse.ArgumentParser({
|
|
prog: 'specsplit',
|
|
version: require('../package.json').version,
|
|
addHelp: true
|
|
});
|
|
|
|
cli.addArgument([ 'type' ], {
|
|
help: 'type of examples to filter',
|
|
nargs: '?',
|
|
choices: [ 'good', 'bad' ]
|
|
});
|
|
|
|
cli.addArgument([ 'spec' ], {
|
|
help: 'spec file to read'
|
|
});
|
|
|
|
var options = cli.parseArgs();
|
|
|
|
////////////////////////////////////////////////////////////////////////////////
|
|
|
|
function readFile(filename, encoding, callback) {
|
|
if (options.file === '-') {
|
|
// read from stdin
|
|
|
|
var chunks = [];
|
|
|
|
process.stdin.on('data', function(chunk) {
|
|
chunks.push(chunk);
|
|
});
|
|
|
|
process.stdin.on('end', function() {
|
|
return callback(null, Buffer.concat(chunks).toString(encoding));
|
|
});
|
|
} else {
|
|
fs.readFile(filename, encoding, callback);
|
|
}
|
|
}
|
|
|
|
|
|
////////////////////////////////////////////////////////////////////////////////
|
|
|
|
readFile(options.spec, 'utf8', function (error, input) {
|
|
var good = [], bad = [],
|
|
markdown = require('..')('commonmark');
|
|
|
|
if (error) {
|
|
if (error.code === 'ENOENT') {
|
|
process.stderr.write('File not found: ' + options.spec);
|
|
process.exit(2);
|
|
}
|
|
|
|
process.stderr.write(error.stack || error.message || String(error));
|
|
process.exit(1);
|
|
}
|
|
|
|
input = input.replace(/→/g, '\t');
|
|
|
|
input.replace(/^\.\n([\s\S]*?)^\.\n([\s\S]*?)^\.$/gm, function(__, md, html, offset, orig) {
|
|
|
|
var result = {
|
|
md: md,
|
|
html: html,
|
|
line: orig.slice(0, offset).split(/\r?\n/g).length,
|
|
err: ''
|
|
};
|
|
|
|
try {
|
|
if (markdown.render(md) === html) {
|
|
good.push(result);
|
|
} else {
|
|
result.err = markdown.render(md);
|
|
bad.push(result);
|
|
}
|
|
} catch (___) {
|
|
bad.push(result);
|
|
}
|
|
});
|
|
|
|
if (!options.type) {
|
|
console.log(util.format('passed samples - %s, failed samples - %s', good.length, bad.length));
|
|
} else {
|
|
var data = options.type === 'good' ? good : bad;
|
|
|
|
data.forEach(function (sample) {
|
|
console.log(util.format(
|
|
'~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~\n' +
|
|
'src line: %s\n\n.\n%s.\n%s.\n',
|
|
sample.line, sample.md, sample.html));
|
|
if (sample.err) {
|
|
console.log(util.format('error:\n\n%s\n', sample.err));
|
|
}
|
|
});
|
|
}
|
|
|
|
process.exit(0);
|
|
});
|
|
|