How it works.md

Q: How does auto-formatting work internally?

A: A simple syntax parsing rule is written using pest.rs to divide a sentence into multiple block types:

Block = _{ 
    MultiSpace     // multiple consecutive spaces
    | Space        // single space
    | SpecialBlock // inline code, math block, URL, time etc.
    | InlineEm     // **<XXX>** block of markdown
    | Punct        // halfwidth punctuation
    | FWPunct      // fullwidth punctuation
    | Num          // number
    | Cn           // Chinese
    | Eng          // English
    | Other        // character blocks other than the above types
}

By default, a space is added between each type. However, there are several cases where spaces are not added:

1. MultiSpace and the surrounding blocks
2. FWPunct and the surrounding blocks
3. Other and the surrounding blocks
4. Punct and the block on the left

For example, the line of 秦时moon汉时关， 万里长征人no还 will be divided into the following blocks:

• Cn: "秦时"
• Eng: "moon" --- inserts a space before Eng
• Cn: "汉时关" --- inserts a space before Cn
• FWPunct: "，" --- DOES NOT add spaces before or after FWPunct
• MultiSpace: " " --- DOES NOT add spaces before or after MultiSpace
• Cn: "万里长征人" --- DOES NOT add spaces after MultiSpace
• Eng: "no" --- adds a space before Eng
• Cn: "还" --- adds a space before Cn
• EOF

After the processing, the result is 秦时 moon 汉时关， 万里长征人 no 还

Q: How do converting rules work internally?

A: A hand-written rule parser parses user rules and organizes them in a trie tree. The common pattern can be expressed as: initial string + a trigger character = result string