[cfe-commits] [PATCH] Warn if -isysroot is given an invalid path

Eric Christopher echristo at gmail.com
Mon Dec 17 21:11:39 PST 2012


On Mon, Dec 17, 2012 at 9:01 PM, John McCall <rjmccall at apple.com> wrote:

> On Dec 17, 2012, at 8:42 PM, Eric Christopher <echristo at gmail.com> wrote:
> > On Mon, Dec 17, 2012 at 7:14 PM, John McCall <rjmccall at apple.com> wrote:
> > On Dec 14, 2012, at 10:41 AM, Eric Christopher <echristo at gmail.com>
> wrote:
> >> Sure, but I just don't see that providing much value.  The point of
> adding a warning was to help catch things like typos in the sysroot path.
>  What are the chances that someone is going to mis-type the sysroot and
> also specify -Wmissing-sysroot?  I'm assuming that typos are most likely to
> come up when someone is running the compiler by hand, rather than from a
> makefile.  I suppose that it could still help when a makefile specifies an
> SDK that is not currently installed.  I'm just not convinced that's worth
> the effort of adding and maintaining a new warning, but I'm open to being
> convinced.
> >>
> >> My only concerns are false positives in fairly standard use cases. I
> think the warning could be useful when debugging problems and if you know
> you aren't in the use cases that might give a false positive you can turn
> it on by default (perhaps with a new umbrella?) but false positives are bad
> for everyone :)
> >
> > In what reality is "early in a system bootstrap" a fairly standard use
> case?
> >
> > Heh. That wasn't mine, but I was imagining a scenario more like "We have
> a sysroot command line option hard coded into our build system, but we
> don't always use it". Mostly lazy sort of thing, I don't care enough to
> hold up the patch, I was just asking if we wanted it on by default.
>
> Sure.  That doesn't seem like a strong argument against having this on by
> default either.
>

Fair enough. Was just curious if it'd been thought about. No objections
either.

-eric
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/cfe-commits/attachments/20121217/a964ae8c/attachment.html>


More information about the cfe-commits mailing list