Date: Thu, 7 May 1998 11:34:35 -0400 (EDT) From: woods@zeus.leitch.com (Greg A. Woods) To: Wolfram Schneider <wosch@cs.tu-berlin.de> Cc: "Jordan K. Hubbard" <jkh@time.cdrom.com>, Randall Hopper <rhh@ct.picker.com>, Poul-Henning Kamp <phk@critter.freebsd.dk>, freebsd-bugs@FreeBSD.ORG Subject: Re: bin/5296 Message-ID: <199805071534.LAA14018@brain.zeus.leitch.com> In-Reply-To: Wolfram Schneider's message of ", May 5, 1998 23:07:04 %2B0200" regarding "Re: bin/5296" id <p1id8dsxw93.fsf@panke.panke.de> References: <199805041546.LAA16661@brain.zeus.leitch.com> <6821.894313171@time.cdrom.com> <199805042253.SAA21201@brain.zeus.leitch.com> <p1id8dsxw93.fsf@panke.panke.de>
next in thread | previous in thread | raw e-mail | index | archive | help
[ On , May 5, 1998 at 23:07:04 (+0200), Wolfram Schneider wrote: ] > Subject: Re: bin/5296 > > There are currently ~6000 PRs. A linear full text search require ~30 > seconds on disk (~35MB) and 7 seconds if cached in memory. > > Glimpse would not help. Glimpse put the 6000 filenames into 256 > blocks, thats 24 files per block. A search for a word which exists > once require to open (in average) 24 files. A search for a word which > exists in 10 PRs require to open ~200 files ;-( Hmmm. I guess that means using a real full text search engine, which means writing a bit more interface glue code (to stuff new PRs into the full-text engine, and to access PRs given the index search output) and allocating disk space for whatever percentage more the full text database takes (usually at least 50%). This is probably work that should be done directly in GNATS -- it's certainly a feature GNATS could use in general (i.e. not specific to just FreeBSD's needs) Liam Quin's text retrieval package (lq-text) would be a good engine to work with (and it's freely available). BTW, Thanks for doing the analysis of the viability of gimpse... -- Greg A. Woods +1 416 443-1734 VE3TCP <gwoods@acm.org> <robohack!woods> Planix, Inc. <woods@planix.com>; Secrets of the Weird <woods@weird.com> To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-bugs" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199805071534.LAA14018>