Proposed gfortran development branch

[Thomas Koenig]

On Sat, 2009-03-21 at 10:08 +0100, Paul Richard Thomas wrote:
> Similarly, the array descriptor work should go to a
> development branch.... but should it be the same one?

Probably not; a separate branch would be a good idea.

I've been giving some thought on how to implement the array descriptor
changes in the library.  I'd like to

- create macros for accessing the ubound and extent of a descriptor

- implement the macros everywhere throughout the library, removing every
  instance of ubound, especially for stride calculation

- "poison" the old ubound by renaming it to u_bound in the header file

- fix compilation errors and fix any regressions :-)

- wait for whatever changes happen to the front end, then just change
  the ubound and extent calculation in once central place

This is a very big change in the library, text-wise.  I think it would
be good not to mix this with other changes.

Comments?

	Thomas