Re: How to handle *_unlocked when they're wrapped by _REENTRANT

[Marc Espie <Marc dot Espie at liafa dot jussieu dot fr>]

On Mon, Jan 11, 1999 at 12:59:46PM -0800, Joe Buck wrote:

> Ulrich wrote:
> > >Maybe it's even desirable to enable the _unlocked function only for
> > >glibc 2.  I.e., protect the definitions of the HAVE_* macros in
> > >system.h with #if __GLIBC__ >= 2.

> Marc Espie writes:
> > No, some other systems are moving in the direction of thread-safe code,
> > while not using GLIBC at all (FreeBSD, OpenBSD...)

> > I think it's better to disable it for broken systems, rather than 
> > special-code the right behavior for Linux and fuck the rest of the free
> > world...

> I'm sure Ulrich didn't intend favoritism or "fucking" anyone, however, his
> proposal was a step down the path we got off of with autoconf: we never
> detect OS versions, we detect features, because it's unsafe otherwise.  If
> glibc has a nice feature the BSD folks might copy it or vice versa,
> therefore we must never do #ifdef Linux or #ifdef GLIBC or #ifdef BSD.

Sorry about the way it came out. I didn't intend to sound harsh.

I've been wading thru binutils recently, which is shock-full of
special-casing and where it's sometimes hard to find out whether the  patch
applies to only one specific system (guess which one...) or whether it
should be a generic fix, hence my fuse is a bit short where such issues
are concerned...

apologies to Ulrich.
-- 
	Marc Espie		
|anime, sf, juggling, unicycle, acrobatics, comics...
|AmigaOS, OpenBSD, C++, perl, Icon, PostScript...
| `real programmers don't die, they just get out of beta'