[Bug fortran/35951] New: Unexpected erro Expected another dimension in array declaration

2008-04-15 Thread john dot young at jrc dot co dot jp 
The below code produces the error Expected another dimension in array
declaration at 1.  It only seems to occur when the passed argument that
contains the array dimension is a derived type and not an integer.  Also, it
only occurs when the derived-type declaration is below (not above) the
declaration of the array.  When the dimension is an integer and not a
derived-type component, either order compiles correctly.  I tested on g95 and
intel and obtained no error.  

My system is WindowsXP, cygwin, and gcc version 4.3.0 20080212 (experimental)
[trunk revision 132252]

John

!==
module test_module
implicit none

  type tt
integer :: n
  end type tt

contains
!===
!===
subroutine test_sub_1(z,t)
implicit none

real :: z(t%n)  !this order produces compiler error Expected another dimension
in array declaration at 1
type (tt) :: t

!type (tt) :: t   !this order compiles correctly
!real :: z(t%n) 


end subroutine test_sub_1
!===
!
subroutine test_sub_2(z,m)
implicit none

real:: z(m)   !this order compiles correctly
integer :: m

!integer,intent(in) :: m!this order compiles correctly too
!real, intent(out) :: z(m)

end subroutine test_sub_2
!===
!===
end module test_module


-- 
   Summary: Unexpected erro Expected another dimension in array
declaration
   Product: gcc
   Version: 4.3.0
Status: UNCONFIRMED
  Severity: normal
  Priority: P3
 Component: fortran
AssignedTo: unassigned at gcc dot gnu dot org
ReportedBy: john dot young at jrc dot co dot jp
GCC target triplet: i686-pc-cygwin


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=35951

[Bug fortran/35951] Unexpected erro Expected another dimension in array declaration

2008-04-15 Thread john dot young at jrc dot co dot jp 


--- Comment #1 from john dot young at jrc dot co dot jp  2008-04-16 01:51 
---
Created an attachment (id=15481)
 -- (http://gcc.gnu.org/bugzilla/attachment.cgi?id=15481action=view)
code to produce error described in bug report


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=35951

[Bug fortran/35951] Unexpected erro Expected another dimension in array declaration

2008-04-15 Thread john dot young at jrc dot co dot jp 


--- Comment #2 from john dot young at jrc dot co dot jp  2008-04-16 01:57 
---
I take back my comment that intel's compiler produces no error.  Intel 8.1
produces an error when the structure declaration is below the array
declaration.  I do not know the standard well enough to know whether this is a
bug or not.  G95 still compiles with no error.

John


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=35951

[Bug fortran/35952] New: Segmentation fault with character strings only when compiling with -funroll-loops and -O3

2008-04-15 Thread john dot young at jrc dot co dot jp 
I attached a short code that produces segmentation faults (or strange cygwin
errors).  The segmentation faults only occur when compiling with both
-funroll-loops and -O3.  If one or the other or neither are used, the program
runs fine.  The code has several examples.  I attached a stackdump for the case
of the three write statements.

System : WinXP, cygwin, gcc version 4.3.0 20080212 (experimental)

==
subroutine test_sub(label)

character(len=*),intent(in) :: label

character (len=(len(label)+20)) :: c1
character (len=(len(label)+20)) :: c2
character (len=(len(label)+20)) :: c3

! the below 3 lines cause a segmentation fault
!  print*, len(c1)
!  print*, len(c2)
!  print*, len(c3)

! the below 2 lines produce strange cygwin errors
!  a ### sig_send: wait for sig_complete event failed, signal -34, rc
-1, Win32 error 6
!  print*, len(c1)
!  print*, len(c2)

! the below 1 lines produces no problem
!  print*, len(c3)

! the below 3 lines cause a segmentation fault (I gave the stackdump for this
case)
  write(c1,'(A)') 'a'
  write(c2,'(A)') 'b'
  write(c3,'(A)') 'c'

! the below 2 lines produce strange cygwin erros
!  a ### sig_send: wait for sig_complete event failed, signal -34, rc
-1, Win32 error 6
!  write(c1,'(A)') 'a'
!  write(c2,'(A)') 'b'

! the below 1 lines produces no problem
!  write(c1,'(A)') 'a'

end subroutine test_sub
!==
program test

  call test_sub('a')
  print*, 'done'

end program test
!==


-- 
   Summary: Segmentation fault with character strings only when
compiling with -funroll-loops and -O3
   Product: gcc
   Version: 4.3.0
Status: UNCONFIRMED
  Severity: normal
  Priority: P3
 Component: fortran
AssignedTo: unassigned at gcc dot gnu dot org
ReportedBy: john dot young at jrc dot co dot jp
GCC target triplet: i686-pc-cywgin


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=35952

[Bug fortran/35952] Segmentation fault with character strings only when compiling with -funroll-loops and -O3

2008-04-15 Thread john dot young at jrc dot co dot jp 


--- Comment #1 from john dot young at jrc dot co dot jp  2008-04-16 05:34 
---
Created an attachment (id=15482)
 -- (http://gcc.gnu.org/bugzilla/attachment.cgi?id=15482action=view)
Code that exhibits segmentation fault.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=35952

[Bug fortran/35952] Segmentation fault with character strings only when compiling with -funroll-loops and -O3

2008-04-15 Thread john dot young at jrc dot co dot jp 


--- Comment #2 from john dot young at jrc dot co dot jp  2008-04-16 05:34 
---
Created an attachment (id=15483)
 -- (http://gcc.gnu.org/bugzilla/attachment.cgi?id=15483action=view)
stackdump for the case of three writes


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=35952

[Bug fortran/35865] New: Spurious warning for vector-valued functions passed as arguments

2008-04-07 Thread john dot young at jrc dot co dot jp 
In the code below, the following warning is produced

Waring: Actual argument contains too few elements for dummy argument 'func'
(1/3) at (1)

If the function result is changed to a scalar instead of a vector, no warning
is produced.  I do not know the standard well enough to know if this is a
problem, but other compilers (G95,Intel) do not produce a warning.  The
function arguments do match however.  I obtained this from GCC 4.3.0 20080212
(experimental) [trunk revision 132252] using cygwin under WinXP.


!===
module vec_func
implicit none

contains
!===
!===
!function vector_func(func,r1) result(res)
function vector_func(func) result(res)
implicit none

interface
  function func(r) result(res)
  implicit none
  real, intent(in) :: r
  real :: res(3)
  end function func
end interface

real :: res(3)

  res = func(0.)

end function vector_func
!==
function test_func(r) result(res)
implicit none

real,intent(in) :: r
real :: res(3)

  res = r

end function test_func
!=
end module vec_func
!=
program a
use vec_func
implicit none

real :: z(3)

  z = vector_func(test_func)
  print*, z

end program a
!=


-- 
   Summary: Spurious warning for vector-valued functions passed as
arguments
   Product: gcc
   Version: 4.3.0
Status: UNCONFIRMED
  Severity: minor
  Priority: P3
 Component: fortran
AssignedTo: unassigned at gcc dot gnu dot org
ReportedBy: john dot young at jrc dot co dot jp
GCC target triplet: i686-pc-cygwin


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=35865

[Bug fortran/35865] Spurious warning for vector-valued functions passed as arguments

2008-04-07 Thread john dot young at jrc dot co dot jp 


--- Comment #1 from john dot young at jrc dot co dot jp  2008-04-08 01:42 
---
Created an attachment (id=15444)
 -- (http://gcc.gnu.org/bugzilla/attachment.cgi?id=15444action=view)
code to produce spurious warning described in bug report


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=35865