This is the mail archive of the mailing list for the GCC project.

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

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

On 11/06/13 04:02, Bernd Schmidt wrote:
On 11/06/2013 10:31 AM, Richard Biener wrote:
We decided to move to C++.  As part of a later discussion we decided
to go with a single general dynamic-casting style, mimicing the "real"
C++ variant which is dynamic_cast < ... >.  Which resulted in

So yes, we've decided to go C++ so we have to live with certain
uglinesses of that decisions (and maybe over time those uglinesses
will fade away and we get used to it and like it).

Thus, there isn't another option besides using the is-a.h machinery
and enabling and using RTTI.  Sticking to C for gimple doesn't seem
to be consistent with the decision to move to C++.

Oh, I'm not saying I'm a big fan of as_a / is_a or C++ in general
as it plays out right now.  But well, we've had the discussion and
had a decision.

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?
I think Andrew outlined much of this at the Cauldron. I'd strongly recommend reviewing the talk/slides. My personal favorite is the type checking and everything that implies. For example, most/all of the tree checking crap just goes away as you'll almost always know what kind of tree you're working with. Checking becomes limited to locations where we we have to do conversions. It becomes a lot harder to mess something up.

I still think all this effort is misdirected and distracts us from
making changes that improve gcc for its users.
I disagree. This effort is meant to restructure GCC so that we can continue to improve it without going insane.

In many respects it's similar to the tree-ssa work we undertook 10+ years ago and to some degree it's cleaning up things we decided not to deal with at that time -- namely reusing trees in gimple/ssa.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]