public inbox for gcc-bugs@sourceware.org
help / color / mirror / Atom feed
* [Bug libfortran/63689] New: -fcheck=array-temps should warn for temporaries during intrinsic assignment
@ 2014-10-30 17:45 bardeau at iram dot fr
  2014-12-07 14:34 ` [Bug fortran/63689] " dominiq at lps dot ens.fr
  0 siblings, 1 reply; 2+ messages in thread
From: bardeau at iram dot fr @ 2014-10-30 17:45 UTC (permalink / raw)
  To: gcc-bugs

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63689

            Bug ID: 63689
           Summary: -fcheck=array-temps should warn for temporaries during
                    intrinsic assignment
           Product: gcc
           Version: unknown
            Status: UNCONFIRMED
          Severity: enhancement
          Priority: P3
         Component: libfortran
          Assignee: unassigned at gcc dot gnu.org
          Reporter: bardeau at iram dot fr

Hi!

please consider this code:

program test
  real(kind=4), pointer :: a(:),b(:)
  allocate(a(2),b(2))
  b = a
end program test

In the assignment, because the LHS and RHS are pointers, they may alias
each-other, so the compiler generates a temporary/intermediate array for
evaluating the RHS. Fine.

This use-case is well detected by -Warray-temporaries:

$ gfortran test.f90 -Warray-temporaries -o test
test.f90:4.6:

  b = a
      1
Warning: Creating array temporary at (1)

On the other hand, -fcheck=array-temps gives no message at run-time:
$ gfortran test.f90 -fcheck=array-temps -o test && ./test
$

I know that the documentation for -fcheck=array-temps clearly states that it
warns for temporary arrays passed to a procedure. However, it would be useful
to extend the warnings to temporaries during intrinsic assignment because the
option name is opened to all temporaries, and because -Warray-temporaries warns
in both cases.

I am currently tracking temporaries, but I prefer runtime checks (temporaries
*are* actually created) instead of compilation checks (temporaries *may* be
created, too many false positives in my program).

I was not sure which version (all opened?) and which component could be
affected, I let you correct this. Thanks!


^ permalink raw reply	[flat|nested] 2+ messages in thread

* [Bug fortran/63689] -fcheck=array-temps should warn for temporaries during intrinsic assignment
  2014-10-30 17:45 [Bug libfortran/63689] New: -fcheck=array-temps should warn for temporaries during intrinsic assignment bardeau at iram dot fr
@ 2014-12-07 14:34 ` dominiq at lps dot ens.fr
  0 siblings, 0 replies; 2+ messages in thread
From: dominiq at lps dot ens.fr @ 2014-12-07 14:34 UTC (permalink / raw)
  To: gcc-bugs

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63689

Dominique d'Humieres <dominiq at lps dot ens.fr> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
   Last reconfirmed|                            |2014-12-07
          Component|libfortran                  |fortran
     Ever confirmed|0                           |1

--- Comment #1 from Dominique d'Humieres <dominiq at lps dot ens.fr> ---
> I am currently tracking temporaries, but I prefer runtime checks
> (temporaries *are* actually created) instead of compilation checks
> (temporaries *may* be created, too many false positives in my program).

See pr48655 for an instance of false positive.

> I was not sure which version (all opened?) and which component could be
> affected, I let you correct this.

Confirmed from 4.8 up to trunk (5.0). Since the checking code is emitted by the
front-end, I erasing the component to 'fortran'.


^ permalink raw reply	[flat|nested] 2+ messages in thread

end of thread, other threads:[~2014-12-07 14:34 UTC | newest]

Thread overview: 2+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2014-10-30 17:45 [Bug libfortran/63689] New: -fcheck=array-temps should warn for temporaries during intrinsic assignment bardeau at iram dot fr
2014-12-07 14:34 ` [Bug fortran/63689] " dominiq at lps dot ens.fr

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).