Am 28.10.2013 23:26, schrieb Paul Richard Thomas:
This patch addresses issues arising from PR57893. It is entirely obvious.
Bootstraps and regtests on FC17/x86_64 - OK for trunk?
Thanks for the patch. It's OK except for the following change:
*** gcc/testsuite/gfortran.dg/unlimited_polymorphic_13.f90 (revision
204135)
--- gcc/testsuite/gfortran.dg/unlimited_polymorphic_13.f90 (working copy)
...
*** 36,43 ****
if (storage_size(o) /= sz) call abort()
type is (complex(c2))
if (storage_size(o) /= sz) call abort()
- end select
- select type (o)
type is (complex(c3))
if (storage_size(o) /= sz) call abort()
type is (complex(c4))
--- 42,47 ----
Please don't merge the two SELECT TYPE blocks. If a system has only two or
three complex kinds, that will lead to an error message like:
Error: CASE label at (1) overlaps with CASE label at (2)
Tobias
PS: The following patch is still pending review:
http://gcc.gnu.org/ml/gcc-patches/2013-10/msg01957.html
2013-10-29 Paul Thomas <pa...@gcc.gnu.org>
PR fortran 57893
* trans-types.c (gfc_typenode_for_spec): Add typenode for
BT_HOLLERITH. Note that the length is incorrect but unusable.
PR fortran 58858
* target-memory.c (gfc_element_size): Add element sizes for
BT_VOID and BT_ASSUMED, using gfc_typenode_for_spec.
2013-10-29 Paul Thomas <pa...@gcc.gnu.org>
PR fortran 57893
* gfortran.dg/unlimited_polymorphic_13.f90 : Use real variables
to determine sizes of real kinds.
PR fortran 58858
* gfortran.dg/unlimited_polymorphic_14.f90 : New test.
