Using C++ in gcc (was Re: [RFC] type safe trees)
Richard Kenner
kenner@vlsi1.ultra.nyu.edu
Fri Jun 25 18:48:00 GMT 2004
what would be the point of requiring a C++ compiler if we were not
then going to start using the language?
None.
I fail to see how anyone can interpret Paul's understanding of the
proposal as 'inventing a language and converting it to C', which is
what you seem to think he's saying, given that you've refuted his
comment as '... not inventing a language ...'.
I didn't think he was talking about that proposal.
I thought he was addressing the proposal to use a more complex version of the
GTY annotations to create C++-like model within GCC and convert it to C. I
interpreted "that languages" to mean the GTY annotations.
I thought his point was that it was better to use C++ than some invented
languages. I was agreeing with that, but pointing out that it was an "apples
to oranges" comparison because the invented languages would only be used in a
small isolated place while C++ would be used everywhere.
So far your only point has been that bootstrapping to a virgin system
would be harder. (a) it's hard full stop (b) the ada compiler is in
ada which appears to be a more complicated language than C++, and it
manages to do such bootstraps.
No, not at all! I would *never* consider doing the first bootstrap to a
virgin system to be for Ada. In my experience, even once the C bootstrap is
working solidly, it takes a considerable amount of time to get Ada
bootstrapped. But Ada and C++ are languages of roughly equivalent complexity
and so I wouldn't want to start with it either.
More information about the Gcc
mailing list