News for hackers. Stuff that matters.
Re:Regular Expression Matching Can Be Simple and Fast (2007) by twic on Friday February 09, 2018 @04:50PM (#16344142)
There's probably someone who posts this comment every time this article comes up, but:

    #![feature(test)]
    
    extern crate regex;
    extern crate test;
    
    use regex::Regex;
    
    pub fn check(pattern: &Regex, input: &str) -> bool {
        pattern.is_match(input)
    }
    
    #[cfg(test)]
    mod tests {
        use super::*;
        use test::Bencher;
    
        #[bench]
        fn bench_check(b: &mut Bencher) {
            let n = 29;
            let pattern = Regex::new(&format!("{}{}", "a?".repeat(n), "a".repeat(n))).unwrap();
            let input = "a".repeat(n);
            b.iter(|| check(&pattern, &input));
        }
    }
And then:

    $ cat rust-toolchain
    nightly-2018-02-09
    $ cargo bench
       Compiling void v1.0.2
       Compiling lazy_static v1.0.0
       Compiling regex-syntax v0.4.2
       Compiling libc v0.2.36
       Compiling utf8-ranges v1.0.0
       Compiling unreachable v1.0.0
       Compiling thread_local v0.3.5
       Compiling memchr v2.0.1
       Compiling aho-corasick v0.6.4
       Compiling regex v0.2.6
       Compiling regexps v0.1.0 (file:///home/twic/Code/Regexps)
        Finished release [optimized] target(s) in 20.11 secs
         Running target/release/deps/regexps-9c084e63d7d31ac3
    
    running 1 test
    test tests::bench_check ... bench:         212 ns/iter (+/- 8)
    
    test result: ok. 0 passed; 0 failed; 0 ignored; 1 measured; 0 filtered out
Rust's standard regex library isn't the fastest in the world, and the lack of backtracking can be limiting, but it is highly optimised for Rust's main use case of posting showoff comments on Hacker News.

I also tried this in Java (OpenJDK 1.8.0_161-b14), and it took about 24 seconds per iteration at 29 characters.

Re:Regular Expression Matching Can Be Simple and Fast (2007) by burntsushi on Friday February 09, 2018 @05:17PM (#16344345)
For those that don't know, Rust's regex crate descends from RE2 and Go's regexp library, both of which were written by Russ Cox, the author of the OP.

The principle difference between Rust's regex crate and, say, RE2 is its focus on literal optimizations. This speeds up a very large number of regexes, and it is one of the reasons why it does tend to be quite fast.

Re:Regular Expression Matching Can Be Simple and Fast (2007) by kbenson on Friday February 09, 2018 @08:54PM (#16345386)
How good is ecosystem/crate support for more complex regex features? Speed is good, especially when you need it, but man, positive/negative lookahead/lookbehind and other advanced features are pretty damn wonderful once you know how to use them. The ability to choose what you want when you want it would be nice.
Re:Regular Expression Matching Can Be Simple and Fast (2007) by burntsushi on Friday February 09, 2018 @09:16PM (#16345464)
The regex crate doesn't support it, and it's not something I'm personally interested in working on. My goal at this point is to make the regex crate as good as possible at what it does, across the board. There's still a long way to go.

Other than that though, there is a WIP crate that actually builds on the `regex` crate and provides fancier features: https://crates.io/crates/fancy-regex (Created by the same author as Xi.)

You can also use bindings to a more established library. The Onigurama bindings seem well developed: https://github.com/rust-onig/rust-onig

There are also bindings to PCRE, but they are of varying quality. The regex crate's benchmark suite binds to several other libraries (including C++ and D libraries), but they are very minimal bindings sufficient for rough microbenchmarking: https://github.com/rust-lang/regex/tree/master/bench/src/ffi

Re:Regular Expression Matching Can Be Simple and Fast (2007) by kbenson on Friday February 09, 2018 @09:23PM (#16345477)
Thanks. The info on the other crates is what I was looking for, I wouldn't expect this one to do more than what it aims for, which seems to be a specific type of regex engine which isn't amenable to all features but gets extra speed in that trade-off.
Re:Regular Expression Matching Can Be Simple and Fast (2007) by aplorbust on Friday February 09, 2018 @10:43PM (#16345746)
"... and the lack of backtracking can be limiting..."

As someone who "gets by" on BRE and ERE, I would consider that a feature.

Re:Regular Expression Matching Can Be Simple and Fast (2007) by thedirt0115 on Friday February 09, 2018 @02:26PM (#16343111)
This is an excellent read. I would suggest reading this also (maybe even before), as it provides one of the simplest (incomplete, but still useful) regex implementations you'll ever see: https://www.cs.princeton.edu/courses/archive/spr09/cos333/be... It suffers from the problems remedied by the original submission (backtracking instead of DFA), but the code is shorter and simpler. They're great to compare.
Re:Regular Expression Matching Can Be Simple and Fast (2007) by glangdale on Friday February 09, 2018 @11:19PM (#16345855)
The biggest gap in the references/history section is the failure to mention V. Glushkov, whose NFA construction has much appeal and certainly predates Thompsons'. Reading the references, one might also draw the conclusion that regex implementation almost ceased between 1968 and 2007.

In our experience, regex matching can be simple and fast, but only sometimes. RE2's approach is straightforward but naive; we built a considerably faster (on some metrics) and more scalable system in Hyperscan. The cost of this turned out to considerable complexity - and there are still many regexes and inputs that don't perform particularly well in either system.

Re:Regular Expression Matching Can Be Simple and Fast (2007) by patrickmay on Friday February 09, 2018 @02:40PM (#16343220)
I thought this was going to be about compiling regular expressions, which is how CL-PPCRE (https://edicl.github.io/cl-ppcre/) manages to be faster than C. Better algorithms beat compilation, though.
Re:Regular Expression Matching Can Be Simple and Fast (2007) by carapace on Friday February 09, 2018 @04:27PM (#16343967)
Slightly off-topic, derivatives of regular expressions is fun and interesting:

https://en.wikipedia.org/wiki/Brzozowski_derivative

http://lambda-the-ultimate.org/node/2293/

Re:Regular Expression Matching Can Be Simple and Fast (2007) by harpocrates on Friday February 09, 2018 @02:19PM (#16343055)
The comments from the original post[1] of this are still quite relevant. I feel like I've seen this link posted here several times (much more recently than 2009), but I can't find the links.

  [1]: https://news.ycombinator.com/item?id=466845
Re:Regular Expression Matching Can Be Simple and Fast (2007) by kbenson on Friday February 09, 2018 @02:45PM (#16343254)
There's been good discussion quite a few of the times this has been posted. Check the "past" link below the submission info at the top for a quick search of all the submissions.
Re:Regular Expression Matching Can Be Simple and Fast (2007) by username223 on Friday February 09, 2018 @02:46PM (#16343274)
Yeah, Russ Cox ranted about pattern matching libraries a decade ago, ignoring the fact that they were already doing things far more complex than formal regular expression matching. In the decade since, that hasn't changed.
Re:Regular Expression Matching Can Be Simple and Fast (2007) by anilshanbhag on Friday February 09, 2018 @04:36PM (#16344039)
There are modules that implement the Thompson's algorithm For Python: https://github.com/xysun/regex
Re:Regular Expression Matching Can Be Simple and Fast (2007) by u801e on Friday February 09, 2018 @03:51PM (#16343718)
The vim text editor actually can use the NFA regular expression engine [1] for some searches. I'm not sure if that's an option for some of the languages mentioned in the article.

[1] http://vimhelp.appspot.com/options.txt.html#%27regexpengine%...

Re:Regular Expression Matching Can Be Simple and Fast (2007) by gumby on Friday February 09, 2018 @02:20PM (#16343059)
This should be labeled (2007) and has been on HN several times before. It's a good article.

The article (and the wikipedia) say Thomson put regexps into a version of QED for CTSS, but CTSS was only used at MIT. He must have traveled between MIT and Bell labs as part of the Multics project. The CTSS reference is pretty obscure.

Re:Regular Expression Matching Can Be Simple and Fast (2007) by Toast_25 on Friday February 09, 2018 @04:05PM (#16343823)
Thanks, I'll keep that in mind for next time. Is there a way to edit the title now?
Re:Regular Expression Matching Can Be Simple and Fast (2007) by gumby on Friday February 09, 2018 @06:43PM (#16344817)
The HN admins do it sometimes.
Re:Regular Expression Matching Can Be Simple and Fast (2007) by lerax on Friday February 09, 2018 @02:33PM (#16343176)
Yes, I know. But regex for everything is bad. What about Parsec? Anyone? http://book.realworldhaskell.org/read/using-parsec.html
Re:Regular Expression Matching Can Be Simple and Fast (2007) by kccqzy on Friday February 09, 2018 @03:22PM (#16343523)
Without the syntactic sugar provided by Haskell most parser combinators in other languages tend to be ugly, difficult to read and difficult to write.