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]

[patch, libfortran] Fix PR 36773


Thomas Koenig wrote:
> OK for trunk?

OK with the following changes in the test case:

+  b = eoshift (a,1,boundary=c)
c  -> c(1,:)

+  b = eoshift (a,(/1/), boundary=c)

This line is completely wrong:
a) The rank of boundary is too large (scalar or rank 1)
b) The shape of shift and boundary must be the same
c) The shape of "a" (w/o dimension 1) must be the same as shift

(a) and (b) are detected by my updated patch to be submitted in a
moment.

How about:
  integer, dimension(0) :: j
  b = eoshift (a, j, boundary=c(1,:))

And also:
+   b = cshift (a,(/1/))

Here, one can also replace "[ 1 ]" by "j".

Tobias


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