saliences.com : Webln

Over the last two days, I have been improving my implementation of Mustache.js using something that I have been calling fat-strides. The basic premise is that the existing tokenizer in the parser is not very smart in that it creates tokens that are too small, which results in a more complicated finite state machine than is strictly necessary. With fat-strides, I increased the size of the tokens by making the lexer much more intelligent. As a result, the parser is now smaller (17 kb versus 23 kb uncompressed), produces smaller code (the compiled mustache programs have far fewer function blocks in them), and is generally 10% faster then the predecessor which was already hella fast. In addition to all of those goodies, the parser will now give you detailed error information about the templates (partial names, line, and character numbers).

All of this is great, but the end goal of all this is to make this implementation of the Mustache syntax be complaint with the Mustache specification, which is the next step and work is already well underway.

Try it out, you will not be disappointed.