Re: [PATCH 3/9] directly recurse using c_type_print

[Tom Tromey <tromey at redhat dot com>]

>>>>> "Joel" == Joel Brobecker <brobecker@adacore.com> writes:

>> This is a minor fix to change the C type printer to directly recurse
>> rather than doing so via the generic type_print function.  This allows
>> us to pass the flags through more easily.

Joel> I've always wondered why we seem to be consistently recursing via
Joel> the generic entry point. The same happens with the expression evaluator,
Joel> for instance, at least for the Ada evaluator. I don't see a testcase
Joel> were recursing directly rather than via the generic interface would
Joel> be incorrect. Even in the case where the Ada evaluator calls the C
Joel> (aka "standard") evaluator, there is no reason that I can think of
Joel> that the evaluator would need to revert back to Ada for that subexp.

It might make sense in some situations where some other language type
printer defers a case to the C printer.  I'll look to see if this might
cause problems... to me this seems pretty fragile, though, and I'd
rather it be done some other way.

Tom