public inbox for gcc-bugs@sourceware.org
help / color / mirror / Atom feed
From: "martin.diehl at kuleuven dot be" <gcc-bugzilla@gcc.gnu.org>
To: gcc-bugs@gcc.gnu.org
Subject: [Bug fortran/105687] New: non-contiguous data in assumed rank
Date: Sun, 22 May 2022 08:00:11 +0000	[thread overview]
Message-ID: <bug-105687-4@http.gcc.gnu.org/bugzilla/> (raw)

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105687

            Bug ID: 105687
           Summary: non-contiguous data in assumed rank
           Product: gcc
           Version: 12.1.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: fortran
          Assignee: unassigned at gcc dot gnu.org
          Reporter: martin.diehl at kuleuven dot be
  Target Milestone: ---

Created attachment 53014
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=53014&action=edit
small reproducer

When passing a strided array to a select-rank statement, the information that
that data is non-contiguous is somehow lost. At least when further passing the
data to an assumed-size function, the arriving data differs.

Note: This might be related to bug 105674. This report illustrates the actual
problem I've encountered (wrong data) while the other report shows intermediate
results when trying to write a minimal reproducer.


code
====
program test

  integer, dimension(2,2), target :: dim_2
  integer, dimension(:,:), pointer :: strided
  integer :: i

  dim_2 = reshape([(i,i=1,size(dim_2))],shape(dim_2))
  strided => dim_2(1:1,:)

  print*, 'is_contiguous(strided)', is_contiguous(strided)
  print*, 'is_contiguous((strided))', is_contiguous((strided))
  print*, ''

  print*, 'call show(strided)'
  call show(strided,size(strided))
  print*, 'call show((strided))'
  call show((strided),size((strided)))
  print*, 'call assumed_rank(strided)'
  call assumed_rank(strided)
  print*, 'call assumed_rank((strided))'
  call assumed_rank((strided))

  contains

  subroutine assumed_rank(dataset)

    integer, intent(in), dimension(..) :: dataset

    select rank(dataset)
      rank(2)
        print*, 'is contiguous in select', is_contiguous(dataset)
        print*, 'data in select', dataset
        call show(dataset,size(dataset))
    end select

    print*, ''

  end subroutine assumed_rank

  subroutine show(dataset,s)

    integer, intent(in) :: s
    integer, intent(in), dimension(*) :: dataset

    print*, 'is contiguous in show', is_contiguous(dataset)
    print*, 'data in show', dataset(1:s)
    print*, ''

  end subroutine show

end program test

output
======
 is_contiguous(strided) F
 is_contiguous((strided)) T

 call show(strided)
 is contiguous in show T
 data in show           1           3

 call show((strided))
 is contiguous in show T
 data in show           1           3

 call assumed_rank(strided)
 is contiguous in select T
 data in select           1           3
 is contiguous in show T
 data in show           1           2


 call assumed_rank((strided))
 is contiguous in select T
 data in select           1           3
 is contiguous in show T
 data in show           1           3

                 reply	other threads:[~2022-05-22  8:00 UTC|newest]

Thread overview: [no followups] expand[flat|nested]  mbox.gz  Atom feed

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=bug-105687-4@http.gcc.gnu.org/bugzilla/ \
    --to=gcc-bugzilla@gcc.gnu.org \
    --cc=gcc-bugs@gcc.gnu.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).