Dear Fortran Reviewers,
I suppose that the slow summer season is upon us and that you are not
finding time for reviewing patches? Anyway, I have been quite busy and
have submitted a number of patches that have not been responded to
except for one remark on a typo and #6 being OK'd by Erik Edelmann.
I propose to commit the lot on Sunday, unless anybody has any
objections. Patches 2], 3], 4] and 5] are obvious or safe. I would
dearly like somebody to cast an eye over 1], even though I am confident
that it is an enormous improvement over the present handling of derived
type association.
1] reforms derived type association and puts it back in the frontend,
where it belongs; the logic is more or less correct, as it stands, but
it does have the two bugs and it just ain't pretty in trans-types.c! 6]
Fixes a problem with dependency chains in local variable declarations;
depending on what letter the symbol names start with, they can come out
in the wrong order, with wierd and wonderful consequences. Both of
these have been thoroughly tested; the gfortran testsuite, the
Polyhedron suite and tonto-2.3 all pass with flying colours.
1] Derived type association
PR28630 - ICE due to a module function returning a derived type
PR28601 - ICE on reexport of renamed type
http://gcc.gnu.org/ml/fortran/2006-08/msg00154.html
2] Derived type constructors
PR28425 - Derived type initialization via a derived type is broken
http://gcc.gnu.org/ml/fortran/2006-08/msg00127.html
3] Host association of the string length variable... or lack of
PR28600 - ICE on character pointer assignment
http://gcc.gnu.org/ml/fortran/2006-08/msg00155.html
4] Passing generic names
PR20886 - passing generic name as dummy argument
http://gcc.gnu.org/ml/fortran/2006-08/msg00166.html
5] A coding error by yours truly
PR28496 - Error during array initialization
http://gcc.gnu.org/ml/fortran/2006-08/msg00167.html
6] Local variable declarations happening in the wrong order
PR28660 - Spurious warning: 'ubound.6' is used uninitialized in this
function
http://gcc.gnu.org/ml/fortran/2006-08/msg00170.html Thanks for checking
this one
out, Erik!
Best regards
Paul