Re: [PATCH 0/6] Conversion of gimple types to C++ inheritance (v3)

[Jakub Jelinek <jakub@redhat.com>]

On Wed, Nov 06, 2013 at 12:31:00PM +0100, Richard Biener wrote:
> > Maybe we need to revisit it? As one of those who were not in favour of
> > the C++ move, can I ask you guys to step back for a moment and think
> > about - what do all of these changes buy us, exactly? Imagine the state
> > at the end, where everything is converted and supposedly the temporary
> > ugliness is gone, what have we gained over the code as it is now?
> 
> as_a gains us less runtime checking and more static type checking
> which is good.

But that really affects --enable-checking=yes builds (and only cases where
things aren't inlined).  If the price for that is uglier and less readable
code, then the price is just too high.

> > I still think all this effort is misdirected and distracts us from
> > making changes that improve gcc for its users.
> 
> That I agree to.  Instead of fixing the less than optimal separation / boundary
> between frontends and the middle-end, or fixing several other long-standing
> issues with GCC we spend a lot of time refactoring things to be C++.
> But that was kind of part of the decision (though I remember that we
> mainly wanted to convert containters and isolated stuff, not gimple
> or trees (I bet that'll be next)).
> 
> Of course I don't see contributors of "changes that improve gcc for its users"
> now wasting their time with converting code to C++.  That conversion
> may slow down those people, but only so much.  It'll get more interesting
> with branch maintainance ...

If the changes are done at least with some for users useful goal (like
perhaps making gcc usable as a library for JITting etc.), then I can see
the benefit, but as_s<> uglification doesn't seem to be beneficial to
users at all, and IMNSHO much better would be to instead
spend time on what is beneficial to users (restrict, vectorizer
cost model, vectorizer improvements, better diagnostics (say range
locations), ...).

	Jakub