[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Bug-gnulib] why no strchr or strrchr?
From: |
Bruno Haible |
Subject: |
Re: [Bug-gnulib] why no strchr or strrchr? |
Date: |
Mon, 23 Aug 2004 17:58:12 +0200 |
User-agent: |
KMail/1.5 |
Oskar Liljeblad wrote:
> Why does gnulib implement memcpy but not strchr?
strchr() was supported a little earlier in SunOS than memcpy().
IIRC, in SunOS 4.1.3, strchr() was present but memcpy() was not.
It was added in SunOS 4.1.3_U1, around 1993.
> Why is autoscan telling me I should check for strchr before using it?
autoconf apparently recommends tricks for maximum portability.
"maximum" is more than nowadays reasonable.
> It would be extremely useful if there was a consensus between the
> autoconf/automake and gnulib crew on what standard to support.
Agreed.
> It would also be useful to know what
> gnulib modules are unnecessary (i.e. already implemented in C89
> C libraries, and there's no known systems implementing it
> incorrectly).
In gnulib's MODULES.html look at "Support for systems lacking ANSI C 89".
gettext assumes that the following are supported by the system (and I got
no bug reports because of it):
c-bs-a
memchr
memcmp
memcpy
memset
For the rest of the list, it depends whether you want to support systems
from 1991-1995 or not.
Bruno