This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [wwwdocs] IPA and LTO updates
* Andi Kleen wrote on Sat, Nov 20, 2010 at 02:41:50PM CET:
>
> > Will the final link also be driven by the makefile or are there parts
> > which are not? If the latter, then with --dry-run, you might not run
> > the parallel parts but do run the final serial parts; that either causes
> > a failure then (which is relatively harmless but a bit annoying), or
> > actually link the wrong (or insufficiently optimized) code, which can
> > be more annoying.
>
> I assume with --dry-run the gcc linker step is not invoked in the
> first place. So this situation should never happen.
That is not right. It's exactly the point I was trying to make in
<http://gcc.gnu.org/ml/gcc-patches/2010-10/msg00217.html> and referenced
therein:
If you add '+' to the recipe, then those commands are executed even in
dry-run mode. So the rule author now becomes responsible that this rule
obeys intended dry-run semantics.
Cheers,
Ralf