Re: [RFA] nuke CONST_PTR

[Michael Snyder <msnyder at redhat dot com>]

Andrew Cagney wrote:
> 
> > In gdb 4.18 and gdb 5.0, CONST_PTR was conditionally defined as nothing
> >> if MSC_VER was defined, and "const" otherwise.  In gdb 5.1 and gdb 5.1.1,
> >> CONST_PTR is always defined to "const".  So CONST_PTR has been
> >> unconditionally defined for only a few months.
> >>
> >> However, in gdb 4.18, gdb 5.0, gdb 5.1, and gdb 5.1, c-lang.c has also
> >> contained this line:
> >>
> >> struct type **const (cplus_builtin_types[]) =
> >>
> >> So this form has been in gdb source code for four releases already
> >> without drawing complaint.
> >>
> >> Testing: I built this on native i686-pc-linux-gnu and ran the test suite.
> >>
> >> Okay to apply?
> >
> >
> > Your reasoning seems good, but your testing doesn't.
> > Seems to me you need to test this when building with Microsoft C.
> >
> > OTOH, do we ever build with Microsoft C any more?
> > Is there any reason to support it?  Cygwin and Djgcc
> > are both self-hosting, aren't they?
> 
> Not since '98!  The line:
> 
>  >> struct type **const (cplus_builtin_types[]) =
> 
> was added in '98 so ever since then GDB hasn't been able to build with
> the MSC compiler CONST_PTR was working around.  As Michael pointed out,
> we've had 4 releases since then.

OK, then this change seems safe.  Maybe we can also
start yanking out other #if MSC code...