Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!mnetor!seismo!ll-xn!ames!ucbcad!faustus
From: faustus@ucbcad.BERKELEY.EDU (Wayne A. Christopher)
Newsgroups: comp.lang.c
Subject: Re: draft ANSI standard: needs your tomatoes
Message-ID: <1155@ucbcad.BERKELEY.EDU>
Date: Wed, 3-Dec-86 18:59:02 EST
Article-I.D.: ucbcad.1155
Posted: Wed Dec  3 18:59:02 1986
Date-Received: Wed, 3-Dec-86 23:17:54 EST
References: <1384@hoptoad.uucp> <5421@brl-smoke.ARPA>
Organization: CAD Group, U.C. Berkeley
Lines: 18

I haven't seen the latest C draft, but I remember a few things from the
older one that bothered me, namely some of the things defined in
limits.h.  The constants EXTERNAL_NAME_LENGTH (6), INCLUDE_FILES_NEST
(4), and SOURCE_LINE_LENGTH (509) in particular seem to be very
inadequate.  How were they decided upon?  Did the committee find the
implementation with the lowest limits for each of these, and then use
the values for the minimum allowable values, thus giving the vendors an
excuse not to fix their compilers?  A compiler that won't accept a
600-line input file is clearly sub-standard and almost useless, and I
object to the ANSII committee approving of this limitation.  The
argument that the standard is only codifying existing practice and not
dictating to vendors how to write their compilers doesn't hold water --
how many compilers supported function prototyping before X3J11?
Anyway, maybe it is too late for me to make this objection...  If
anybody can enlighten me on the rationale behind this I would
appreciate it.

	Wayne