Le 24/10/2015 21:29, Dominique d'Humières a écrit :
At revision r229288 compiling the following test
[...]
while it compiles without error at r229261.
I believe the accesses to ref->u.ar should be guarded with ref->type ==
REF_ARRAY.
Steve, a patch doing that is preapproved.
Mikael
At revision r229288 compiling the following test
implicit none
type :: template_t
integer :: type
character(256) :: charset1, charset2
integer :: len1, len2
end type template_t
contains
subroutine match_quoted (tt, s, n, range)
type(template_t), intent(in) :: tt
c
Dear Steve,
This is OK to commit.
Thanks for the patch
Paul
On 23 October 2015 at 21:29, Steve Kargl
wrote:
> On Fri, Oct 23, 2015 at 12:28:14PM -0700, Steve Kargl wrote:
>> Built and regression tested on x86_64-*-freebsd.
>> OK to commit?
>>
>
> Now with the patch attached!
>
> --
> Steve
On Fri, Oct 23, 2015 at 12:28:14PM -0700, Steve Kargl wrote:
> Built and regression tested on x86_64-*-freebsd.
> OK to commit?
>
Now with the patch attached!
--
Steve
Index: gcc/fortran/array.c
===
--- gcc/fortran/array.c (revisio
All,
The attached patch fixes several ICEs caused by invalid
charlengths. The new testcase pr67805.f90 shows the
kinds of issues the patch will detect. An appropriate
error message is now issued.
Built and regression tested on x86_64-*-freebsd.
OK to commit?
2015-10-23 Steven G. Kargl
