Would you know of a good set of regexps to benchmark?

I'm sure there are many I don't know about out there. It's just a fun thing to work on! I haven't looked, but I suspect that node.js must use the Google v8 regexp engine, in which case it would be bactracking and would not perform dominating regexp extraction, so rather slow on complex regexps.


I just benchmarked git-grep on the same regexps, and it performs super well. It takes generally slightly more time on 'sys' and 'user' results, but the 'real' times are generally better (multi-threading behaves superbly).

regexp   engine              sys  user real
foobar   git-grep            0.51 0.56 0.23
foobar   taskset 1 git-grep  0.32 0.48 0.81
foobar   jrep                0.23 0.08 0.32
foobar   grep-2.17           0.18 0.51 0.70

However it looks like it does not perform dominating regexps extraction, so the performance on complex regexps is only excellent (with a risk of getting worse).

regexp                                engine               sys  user real
^\s*(void|unsigned|int)\s+func\s+\(   git-grep            0.31 5.67 0.85
^\s*(void|unsigned|int)\s+func\s+\(   taskset 1 git-grep  0.28 4.10 4.38
^\s*(void|unsigned|int)\s+func\s+\(   jrep                0.25 0.12 0.35
^\s*(void|unsigned|int)\s+func\s+\(   grep-2.17           0.18 1.15 1.33

So out of the box it is really good. That tells me I should be looking more at the multi-threading!