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

[Bernd Schmidt]

On 11/06/2013 12:31 PM, Richard Biener wrote:
> 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)).

What I seem to remember is being told that we'd use C++ mostly for new
code and not engage in a wholesale rewrite of the compiler.

Instead things happened exactly as predicted, we're getting immense
amounts of churn for little real gain, and in some cases the C++ changes
are not even improvements and look like really strange examples of C++
code. I'll offer as evidence the pass manager with its bizarre has_gate
and has_execute variables, member functions for gate and execute
wrapping non-member functions, and a separate pass_data outside of its
class (not even as a static member).

Now that the question of what would happen after a C++ conversion is no
longer theoretical but has been answered, can we maybe just stop and
think for a while whether this is really the way we should be going?


Bernd