|
From: | Derek Robert Price |
Subject: | Re: [Bug-gnulib] 03-getndelim2-c89.diff |
Date: | Tue, 22 Jul 2003 14:11:27 -0400 |
User-agent: | Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.4) Gecko/20030624 Netscape/7.1 |
Bruno Haible wrote:
Derek Robert Price wrote:What about the pre-ANSI C function declarations in memmove.c and maybe elsewhere: void * memmove (dest, source, length) char *dest; const char *source; unsigned length; { Is that just lack of motivation to update them or is there a reason for leaving them that way?1. Lack of motivation - why bother with these old modules that are not used by most packages.
Well, CVS is still using them though I couldn't really tell you if they are strictly necessary any longer, but I might supply patches with the hope of keeping CVS consistent.
2. Lack of testing facilities - the systems on which these modules were necessary (SunOS, Ultrix etc.) are not available for testing to most of us.
Well, I've started checking in some changes to CVS along these lines, so I'll probably hear about it soon (at the very least probably with our next release) if there are many platforms that want the pre-ANSI headers, but aren't ANSI headers a part of even the freestanding C89 specification?
Incidentally, the CVS nightly test platforms, which include Solaris 8, HP-UX 11.00, Darwin 5.5, an AIX platform who's uname output I don't understand (`AIX rioscpu2 3 4 000030498200'), and IRIX 6.5 haven't reported problems with any of this yet. Ultrix isn't in the list though and Solaris 8 is fairly recent.
Derek -- *8^) Email: address@hidden Get CVS support at <http://ximbiot.com>! -- I will not spank others. I will not spank others. I will not spank others... - Bart Simpson on chalkboard, _The Simpsons_
[Prev in Thread] | Current Thread | [Next in Thread] |