svn commit: r348007 - projects/fuse2/lib/libc/gen

[Konstantin Belousov]

On Mon, May 20, 2019 at 03:17:05PM -0700, Conrad Meyer wrote:
> On Mon, May 20, 2019 at 1:51 PM Konstantin Belousov <kostikbel at gmail.com> wrote:
> >
> > On Mon, May 20, 2019 at 02:43:36PM -0600, Alan Somers wrote:
> > > True, but I though that strlen would be more readable.  Clang is smart
> > > enough to realize that strlen's argument is a constant and omit the
> > > call to strlen, so that's what I went with.  Is there any other reason
> > > to prefer sizeof() - 1 ?
> >
> > The reason is that it would also work for not so smart compiler, or with
> > a smart compiler after we finally turn on -ffreestanding for libc, as it
> > ought to be.
> 
> If we did that, it might make sense to implement strlen() as a macro
> that did essentially the same thing using __builtin_strlen() (and
> __builtin_constant_p, if necessary) extensions.
It is too far-reaching to request to hack strlen() to use gcc extension in
response to a simple request to use sizeof() in appropriate situation.