Date: Thu, 8 May 1997 08:38:38 -0400 (EDT) From: Thomas David Rivers <ponds!rivers@dg-rtp.dg.com> To: ponds!zeta.org.au!bde, ponds!time.cdrom.com!jkh, ponds!barcode.co.il!nadav, ponds!lakes.water.net!rivers Cc: ponds!freebsd.org!hackers Subject: Re: /usr/include/ftpio.h is not C++ safe Message-ID: <199705081238.IAA05623@lakes.water.net>
next in thread | raw e-mail | index | archive | help
> >> They all already support K&R except <dialog.h>, <ftpio.h>, <libdisk.h>, > >> <md2.h>, <md4.h>, <md5.h>, <ncurses.h>, <pcap.h>, <pcap-namedb.h>, > >> <strhash.h>, <unctrl.h>. I.e., all except 0 CSRG-derived headers, 2 > >> LBL-derived ones (pcap*), 3 Linux-derived ones (dialog, ncurses, unctrl) > >> and 6 jkh/phk headers (the rest) support K&R. > > > > Oops - what about netdb.h and uses of 'const', etc... (which should > >work, but apparently don't.) > > Compile them with a K&R compiler (not gcc -traditional) or use `-Dconst='. > Both of these work bogusly by defining const away. > > Bruce > > Ok - just how K&R do we want to get? Do we say "late" K&R or early K&R? o) no Structure returns? o) no "void" type? o) no "enums"? o) slightly different expression promotion rules? o) no "const"? o) no "volatile"? There's a good bit to consider here. I'd be happy with just handling "const" in the system headers, since that's what GCC complains about. [along with __P() for compilers that don't support prototypes.] - Dave Rivers -
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199705081238.IAA05623>