This directory contains a Rust runner program for benchmarking Rust's
regex-lite crate. The regex-lite crate is authored by the
same person who created this barometer.
This runner program makes the following decisions:
- Only one pattern is supported.
- Currently, only haystacks that are valid UTF-8 are supported. At the time of
writing,
regex-liteonly supports searching&str.
The regex-lite crate specifically gives up functionality and performance in
exchange for faster compilation times and smaller binary size. To this end, the
extent to which regex-lite supports Unicode is that its fundamental atom of
matching is a single codepoint. That is, it knows about UTF-8. But it doesn't
have any other Unicode support. Case insensitive searching, \w, \d, \s
and \b are all ASCII-only. There is no support for Unicode character classes
such as \pL. Note though that [^a] matches any codepoint other than a by
necessity of being aware of UTF-8.