Skip to content

mneumann/dhttpd

Repository files navigation

dhttpd

A fast HTTP server for the D language

Design

Parsing the HTTP request is very fast by using a Ragel 1 state machine inherited from the Mongrel Ruby webserver but extended a bit to save more memory.

Parsing stores positions into the buffer currently being parsed. If parsing needs more input it returns to the caller. Then the caller has to read more data and append it at the end of the buffer (or construct a new buffer containing all the content), so that all currently stored positions into the buffer stay valid. Then continue parsing with the bigger buffer at the last position.

We do not copy data, nor modify the buffer itself. Custom HTTP headers can be allowed, or disabled. If they are disabled, custom HTTP headers are just ignored. If they are enabled, they will be stored in an array.

About

A fast HTTP server for the D language

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages