Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP Posting-Version: version B 2.10.2 9/18/84; site faron.UUCP Path: utzoo!decvax!linus!faron!wdr From: wdr@faron.UUCP (William D. Ricker) Newsgroups: net.lang Subject: Re: AWK Message-ID: <331@faron.UUCP> Date: Wed, 21-Aug-85 09:13:19 EDT Article-I.D.: faron.331 Posted: Wed Aug 21 09:13:19 1985 Date-Received: Thu, 22-Aug-85 02:58:20 EDT References: <323@faron.UUCP> <249@lzwi.UUCP> <19@seismo.CSS.GOV> Reply-To: wdr@faron.UUCP (William D. Ricker) Distribution: na Organization: The MITRE Coporation, Bedford, MA Lines: 30 Keywords: AWK, Programming Pearls, Bentley Summary: In article <19@seismo.CSS.GOV> tiberio@seismo.CSS.GOV (Mike Tiberio) writes: > >I don't read CACM, what functions are you talking about? We have added >all the siingle argument math functions from libm.a to awk. We also use >awk on double precision data and found it necessary to modify awk so that >it would never convert the double to single during handling. Our version >came with getline and exit, are these the functions you are talking about? > >seismo!tiberio No, the 'functions' I was referring to is the facility for user-defined functions. Apparently Kernighan's latest version of AWK, in use at AT&T Bell Labs by Jon Bentley, CACM's "Programming Pearls" columnist, allows definitions of recursive AWK functions.) Guy Harris suggests these features might be in SVR2, but I didn't quite follow his description of what was in which release. (Bentley gives, in AWK, a subroutine library of various sorting and searching routines, pointing to particular spots he had trouble implementing even though he had formally verified the specifications.) -- William Ricker wdr@faron.UUCP (UUCP) decvax!genrad!linus!faron!wdr (UUCP) {allegra,ihnp4,utzoo,philabs,uw-beaver}!linus!faron!wdr (UUCP) Opinions are my own and not necessarily anyone elses. No warranty, expressed or implied, is given about the veracity of any statements contained herein. Applicable law in your state may differ.