PR 23551: why should we coalesce inlined variables?

[Richard Guenther]

On 5/10/07, Daniel Jacobowitz <drow@false.org> wrote:
> On Thu, May 10, 2007 at 11:30:21AM +0200, Richard Guenther wrote:
> > I would expect this to cause a noticable memory usage increase on testcases
> > like tramp3d.  Can you check that, too?  (c.f.
> > www.suse.de/~rguenther/tramp3d/tramp3d-v4.cpp.gz)
> > So I guess while it may look reasonable for C, it isn't really for C++.
>
> But doesn't Alex's point entirely apply there too?

For debugging?  Sure.  But we don't want to have debug data explode, neither
compile-time memory usage or compile-time.  This is why I asked him to check
the effect of his patch on tramp3d.  Iff we actually end up removing
the variables
anyway (not because of coalescing but because we should still copyprop over
those variables and remove the dead copies) the effect shouldn't be
that big, nor
the gain.

Richard.