Hacker News Reader

You don't have to use a channel for coordination. Here is a lexer implementation (in C!) that very closely follows Rob Pike's talk and uses a ring buffer for coordination and is plenty fast.

https://github.com/lpereira/lwan/blob/master/src/lib/lwan-te...

If you watch the talk carefully, Rob Pike himself mentions this near the end of the talk.

View on HN

Even if you do this, the CPU still ultimately has to synchronize if you put the workflow across threads, so you have to get a contention hit somewhere. (Lockfree data structures still hit locks in the silicon, after all.) I could be wrong, but when it comes to performance it seems like your best option is moving the concurrency elsewhere. For design? I dunno. I find the design to be elegant for sure, but I also don’t mind handwriting parsers in general.

The lexer is entirely single-threaded, as in the original Rob Pike talk.

>I find the design to be elegant for sure, but I also don’t mind handwriting parsers in general.

But this is a hand-written parser! Just using function pointers to store the state instead of using an enum and switch statements.

Yep, fair enough. As a single-threaded design, I have no particular problems with it.

Reply on news.ycombinator.com