Hi Thomas,
On 25.02.23 10:11, Thomas Schwinge wrote:
Do to the scan patterns need adjusting, or is something wrong?
The former. Regarding:
* gfortran.dg/goacc/finalize-1.f - for 'acc exit data':
!$ACC EXIT DATA FINALIZE DELETE (del_f_p(2:5))
Here, 'map\\(to:del_f_p [pointer set]' chang
Hi Tobias!
On 2023-02-23T17:42:08+0100, Tobias Burnus wrote:
> (Side note: this patch has been committed to OG12 as
> http://gcc.gnu.org/g:55a18d47442 )
I see og12 commit 55a18d4744258e3909568e425f9f473c49f9d13f
"Fortran/OpenMP: Fix mapping of array descriptors and deferred-length strings"
regr
Minor change to previous patch – it did not affect the mainline build
but it makes more sense this way and on OG12, i.e. with mapping
allocatable components (patch posted last year in Q1 during Stage 4),
it gave an ICE without this change as one '*' wasn't stripped.
--- b/gcc/fortran/trans-openmp
This patch moves some generic code for Fortran out of gimplify.cc
to trans-openmp.cc and fixes several issues related to mapping.
Tested with nvptx offloading.
OK for mainline?
Caveats:
Besides the issues shown in the comment-out code, there remains also an
issue with implicit mapping - at lea
