Re: [dev] off topic - awk versions performance comparison

From: Uriel <>
Date: Sun, 15 Aug 2010 05:28:53 +0200

On Sun, Aug 15, 2010 at 1:53 AM, Kris Maglione <> wrote:
> On Sat, Aug 14, 2010 at 10:38:32PM +0200, Uriel wrote:
>> On Sat, Aug 14, 2010 at 12:07 PM, Kris Maglione <>
>> wrote:
>>> nawk is one-true-awk from FreeBSD. I find the results strange, namely
>>> because Plan 9's awk is also one-true-awk. It also produces reasonable
>>> results with "^y" instead of "y", while gawk doesn't. I know that nawk
>>> uses
>>> a combination NFA/DFA, but I see that Plan 9's awk instead pre-process
>>> the
>>> expression and uses Plan 9's pure-NFA engine. My guess is that, because
>>> they're greedy algorithms, they both traverse the entire string looking
>>> for
>>> a possibly longer match, but my understanding of Plan 9's altorithm was
>>> otherwise.
>> I think one of the very few differences between Plan 9's awk and bwk's
>> awk is UTF-8 support, not sure why that would make a difference, but I
>> thought they were mostly identical otherwise.
> That's mostly true, but awk is still actively developed, and has had a lot
> of changes in the 11 years since it was last synced with Plan 9. It also
> uses an entirely different regular expression engine.

Interesting, I thought they had done a sync more recently *sigh* And
did't know about the completely different regexp engine.

I'm a bit surprised because in my experience I had found that if
anything Plan 9 awk's seemed to err more on the side of leaving things
the way they were upstream, for example system() uses ksh instead of


> --
> Kris Maglione
> The easy confidence with which I know another man's religion is folly
> teaches me to suspect that my own is also.
>        --Mark Twain
Received on Sun Aug 15 2010 - 05:28:53 CEST

This archive was generated by hypermail 2.2.0 : Sun Aug 15 2010 - 05:36:02 CEST