This is the mail archive of the fortran@gcc.gnu.org mailing list for the GNU Fortran 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 RFA: -Wc++-compat warning about ++/-- with enum value


NightStrike wrote:

> Not only is this an accurate statement, but you also have to remember
> that the learning curve of a project this large written in OO fashion
> is far less.  If a project follows simple OO techniques, it's much
> easier for a newcomer to pick an isolated piece of the puzzle and
> learn that part, instead of drowning in the giant sea of source code
> for the bigger picture.

  Well, even more so: GCC is in fact a very largely OO-based design already.
It's just that we've done the whole thing manually, rather than leaving the
grunt work to the compiler, where it belongs.

  Personally I'm looking forward to "class tree", and not having to worry any
more about trying to use the wrong polymorphic accessor, that will reduce our
exposure to a whole category of bugs, and I'd guess that we'll be able to do
away with large chunks of --enable-checking, replacing runtime overheads with
static compile-time correctness guarantees.  I don't suppose we'd need to use
anything more tricky than simple classes with virtual functions and maybe the
odd template to get 90% of the benefits that C++ can bring us.

    cheers,
      DaveK


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