getting back the quote_ignores_source_dir behavior of -I-

Glenn Fowler gsf@research.att.com
Mon Oct 9 19:29:00 GMT 2006


restoring -I- functionality, either by one option or two, is the main point

-- Glenn Fowler -- AT&T Research, Florham Park NJ --

On Mon, 9 Oct 2006 12:05:59 -0700 Zack Weinberg wrote:
> On 10/9/06, Mike Stump <mrs@apple.com> wrote:
> > My take is that the idea to remove the -I- semantic was wrong.  I
> > don't see the value in randomly respelling existing options.  Instead
> > of this patch, I think we should just remove the deprecation of -I-,
> > if people didn't want to do that, than I think this patch is
> > preferable over not doing anything.

> If you don't see the point of separating the two distinct operations
> that -I- performs, so that people who need just one can have it
> without tripping over the other, then I decline to argue with you.  It
> *was* a mistake to deprecate -I- without providing replacements for
> all of its functionality, but at the time we had good reasons to
> believe that *no one* used the ignore-directory-of-current-file
> feature [have you noticed that under relatively plausible conditions,
> disabling it breaks syslimits.h, for instance?]

> I think Daniel's patch is entirely sensible, and if I still had patch
> approval authority, it would have gone in already.



More information about the Gcc-patches mailing list