Date: Mon, 23 Jun 2003 08:25:55 -0500 (CDT) From: Sean Farley <sean-freebsd@farley.org> To: freebsd-hackers@FreeBSD.ORG Subject: Re: Replacing GNU grep revisited Message-ID: <20030623080515.S43893@thor.farley.org> In-Reply-To: <20030623114005.GB12521@HAL9000.homeunix.com> References: <20030621103502.K18572@thor.farley.org> <20030622005852.GB59673@HAL9000.homeunix.com> <20030623114005.GB12521@HAL9000.homeunix.com>
next in thread | previous in thread | raw e-mail | index | archive | help
On Mon, 23 Jun 2003, David Schultz wrote: > On Sun, Jun 22, 2003, Sean Farley wrote: > > On Sat, 21 Jun 2003, David Schultz wrote: > > > > 2. GNU's grep is using libgnuregex. The speed-up by dds@ would not > > be felt? > > I was referring to freegrep, which I thought used the native libregex. Yes, it does. I was just confused. > In any case, if freegrep solves problems that GNU grep has, has the > features people care about, is competitive in terms of performance, > and has no known major bugs, then I would have no objections to it. > As long as there are good technical reasons to switch, licensing > advantages and performance are the least critical issues in my mind. > We're not talking about a factor of 2, right? It varies. Simpler searches are handled within FreeGrep such as expressions with "^", "$" or ".". This is how GNU grep searches so quickly; it "cheats" before it falls-back to the regex library. More complex searches can hit the regex library and slow FreeGrep down. In those cases, the ones I have found, it is still below a factor of two difference. I usually do tests on /usr/src. I have seen FreeGrep take 1:10 while GNU's grep took 0:58. Interesting. I found that GNU's grep actually finds a match for "grep -ail freebsd /usr/ports/distfiles/*": /usr/ports/distfiles/ezm3 ezm3 is a directory with a filename that contains FreeBSD in it. Sean ----------------------- sean-freebsd@farley.org
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20030623080515.S43893>