Hacker News Reader

Comment by pfdietz | Hacker News Reader

Something as simple as "is this string in this fixed set of strings" benefits from compile-time optimization of the matching algorithm. This is a place where Lisp's macros really shine, as they enable this kind of optimization to be performed at macro expansion time, without the need for preprocessing or matching algorithms built into the compiler itself.

View on HN

I'm not sure how efficiently it compiles, but Racket's match form is implemented using macros and is stunningly powerful!

https://docs.racket-lang.org/reference/match.html

The paper (IIRC) https://arxiv.org/abs/1106.2578 https://arxiv.org/pdf/1106.2578.pdf

I wrote a purely macro-based implementation of pattern matching for TXR Lisp.

It's documented over three consecutive chapters:

https://www.nongnu.org/txr/txr-manpage.html#N-5C9CED75 (Structural Pattern Matching)

https://www.nongnu.org/txr/txr-manpage.html#N-B5350C53 (Pattern-Matching Notation)

https://www.nongnu.org/txr/txr-manpage.html#N-CD6130ED (Pattern-Matching Macros)

It's implemented in one file: https://www.kylheku.com/cgit/txr/tree/stdlib/match.tl

Starting with the second definition of the function non-triv-pat-p (the real definition which replaces the temporary stub one), the module starts to eat its own dogfood by using its own pattern matching.

That function, for instance, decides whether a pattern is "trivial" or not, using pattern matching itself on the pattern matching notation.

The string quasiliteral pattern's expansion relies heavily on pattern matching, also.

Yep, trivia* for instance is very impressive in performance, especially considering you can switch optimizers.

* <https://github.com/guicho271828/trivia>

I've used it. :)

https://github.com/guicho271828/trivia/issues/108

Reply on news.ycombinator.com