[Bug fortran/9395] can not acces allocatable array in fortran90

[Bug fortran/9395] can not acces allocatable array in fortran90

[joachim.protze at zih dot tu-dresden.de]

http://sourceware.org/bugzilla/show_bug.cgi?id=9395

--- Comment #5 from Joachim Protze <joachim.protze at zih dot tu-dresden.de> 2011-02-16 16:51:11 UTC ---
Created attachment 5246
  --> http://sourceware.org/bugzilla/attachment.cgi?id=5246
fortran program with 3d-array

I checked out the branch and found some strange behavior regarding subarrays
and at last gdb segfaulted:

(gdb) b 18
Breakpoint 1 at 0x400aae: file array-segf.f90, line 18.
(gdb) r
Starting program: array-segf 

Breakpoint 1, typesubarray () at array-segf.f90:18
18      buffer=100
(gdb) p buffer
$1 = (( ( 0, 1, 2, 3) ( 4, 5, 6, 7) ( 8, 9, 10, 11) ( 12, 13, 14, 15) ) ( ( 16,
17, 18, 19) ( 20, 21, 22, 23) ( 24, 25, 26, 27) ( 28, 29, 30, 31) ) ( ( 32, 33,
34, 35) ( 36, 37, 38, 39) ( 40, 41, 42, 43) ( 44, 45, 46, 47) ) ( ( 48, 49, 50,
51) ( 52, 53, 54, 55) ( 56, 57, 58, 59) ( 60, 61, 62, 63) ) )
(gdb) p buffer(:,:,0)
$2 = (( 0, 1, 2, 3) ( 4, 5, 6, 7) ( 8, 9, 10, 11) ( 12, 13, 14, 15) )
(gdb) p buffer(:,:,1)
$3 = (( 0, 1, 2, 3) ( 4, 5, 6, 7) ( 8, 9, 10, 11) ( 12, 13, 14, 15) )
(gdb) p buffer(:,:,4)
no such vector element
(gdb) p buffer(:,0,0)
$4 = (0, 1, 2, 3)
(gdb) p buffer(:,2,3)
$5 = (0, 1, 2, 3)
(gdb) p buffer(1,:,3)
$6 = (0, 4, 8, 12)
(gdb) p buffer(2:3,0,0)
$7 = (0, 1)
(gdb) p buffer(3,1,:)
segfault


The subarrays that are printed are choosen wrong. Size and dimension are
choosen right, but all subarrays are aligned to (0,0,0).
If the first dimension (last digit) is set to range, gdb always segfaults.

GNU Fortran (Ubuntu 4.4.3-4ubuntu5) 4.4.3
GNU gdb (GDB) 7.2.50.20110213-cvs

-- 
Configure bugmail: http://sourceware.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

[Bug fortran/9395] can not acces allocatable array in fortran90

[jan.kratochvil at redhat dot com]

http://sourceware.org/bugzilla/show_bug.cgi?id=9395

--- Comment #6 from Jan Kratochvil <jan.kratochvil at redhat dot com> 2011-02-18 00:01:51 UTC ---
Many thanks for the bugreport.
archer-jankratochvil-vla commit 678cbb7340d8edfa92ce6de931d8d4e9c3b86c06 should
have it fixed.

-- 
Configure bugmail: http://sourceware.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

[Bug fortran/9395] can not acces allocatable array in fortran90

[joachim.protze at zih dot tu-dresden.de]

http://sourceware.org/bugzilla/show_bug.cgi?id=9395

--- Comment #7 from Joachim Protze <joachim.protze at zih dot tu-dresden.de> 2011-02-25 13:59:14 UTC ---
Created attachment 5266
  --> http://sourceware.org/bugzilla/attachment.cgi?id=5266
fortran program with bigger 3d-array

It seems, that we now have infinite loops in some cases (for 300x300x300
array):

(gdb) p buffer(10,10,:10)
$1 = (3010, 93010, 183010, 273010, 363010, 453010, 543010, 633010, 723010,
813010, 903010)
(gdb) p buffer(10,9:10,10)
$2 = (902710, 903010)

but:
(gdb) p buffer(10,10,9:10)
-> hangs, should just display the last two values of $1

With 200x200x200 array the last one gives correct answer.

-- 
Configure bugmail: http://sourceware.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

[Bug fortran/9395] can not acces allocatable array in fortran90

[jan.kratochvil at redhat dot com]

http://sourceware.org/bugzilla/show_bug.cgi?id=9395

--- Comment #8 from Jan Kratochvil <jan.kratochvil at redhat dot com> 2011-02-25 18:13:48 UTC ---
(In reply to comment #7)
> It seems, that we now have infinite loops in some cases

Thanks, the code was reading some memory after the end of the array.

For small arrays it makes no difference but with such a large array GDB tried
to read unmapped memory pages which resulted in very ptrace reading and day it
would say:
Cannot access memory at address 0x7ffff0c75f58

Fixed now on archer-jankratochvil-vla by:
530ce8bb0aeeb513f566da7a7fae46d6414f8dc1

(gdb) p buffer(10,10,9:10)
$1 = (813010, 903010)

The VLA code needs a rewrite for merge, though.  I do not find viable to fix
the Comment 4 with the current codebase.

-- 
Configure bugmail: http://sourceware.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

[Bug fortran/9395] can not acces allocatable array in fortran90

[at_gdb at mathalacarte dot com]

http://sourceware.org/bugzilla/show_bug.cgi?id=9395

Fred Krogh <at_gdb at mathalacarte dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |at_gdb at mathalacarte dot
                   |                            |com

--- Comment #9 from Fred Krogh <at_gdb at mathalacarte dot com> 2011-08-06 19:18:00 UTC ---
It seems I'm having this problem when using GNU gdb (Gentoo 7.2 p1) 7.2, and
GNU Fortran (Gentoo 4.6.1 p1.0, pie-0.4.5) 4.6.1.  Here is the program I'm
running which includes comments on what gdb does.

  program test1
!   Gdb does not print entries in allocatable arrays
    integer, allocatable :: itest(:)
    allocate (itest(10))
    do i = 1, 10
      itest(i) = i
    end do
    print '("itest =", 10i3)', itest(1:10)
    print '("Stop gdb on this statement and look at itest")'
! p itest(1:2) says slice our of range
! p itest(1) gives 0
! p itest(2) gives -1   
    stop
  end program test1

-- 
Configure bugmail: http://sourceware.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

[Bug fortran/9395] can not acces allocatable array in fortran90

[at_gdb at mathalacarte dot com]

http://sourceware.org/bugzilla/show_bug.cgi?id=9395

--- Comment #10 from Fred Krogh <at_gdb at mathalacarte dot com> 2012-02-17 16:56:29 UTC ---
I just tried the program test1 that I posted in the above comment.  Got the
same (bad) result, but this time using GNU gdb (Gentoo 7.4 p1) 7.4, and GNU
Fortran (Gentoo 4.6.2 p1.1, pie-0.4.5) 4.6.2.  This bug really complicates the
debugging of Fortran programs using allocatable arrays.

-- 
Configure bugmail: http://sourceware.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

[Bug fortran/9395] can not acces allocatable array in fortran90

[at_gdb at mathalacarte dot com]

http://sourceware.org/bugzilla/show_bug.cgi?id=9395

Fred Krogh <at_gdb at mathalacarte dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
            Version|6.5                         |7.4

-- 
Configure bugmail: http://sourceware.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

[Bug fortran/9395] cannot access allocatable array in fortran90

[schneidermark at myfairpoint dot net]

http://sourceware.org/bugzilla/show_bug.cgi?id=9395

Mark Schneider <schneidermark at myfairpoint dot net> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Priority|P3                          |P1
                 CC|                            |schneidermark at
                   |                            |myfairpoint dot net
            Version|7.4                         |unknown
            Summary|can not acces allocatable   |cannot access allocatable
                   |array in fortran90          |array in fortran90
           Severity|normal                      |critical

--- Comment #11 from Mark Schneider <schneidermark at myfairpoint dot net> 2013-02-20 16:49:11 UTC ---
All IDE's based on gfortran/gdb have this problem, including Code:Blocks,
Simply Frotran, NetBeans and Eclipse. It makes it almost impossible to debug
newer code that depends on allocating arrays. And allocatable arrays in fortran
have been around since fortran 90.

-- 
Configure bugmail: http://sourceware.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

[Bug fortran/9395] cannot access allocatable array in fortran90

[at_gdb at mathalacarte dot com]

http://sourceware.org/bugzilla/show_bug.cgi?id=9395

--- Comment #12 from Fred Krogh <at_gdb at mathalacarte dot com> 2013-02-20 17:17:32 UTC ---
Just tried my little test1 program above with "GNU Fortran (Gentoo 4.7.2 p1.3,
pie-0.5.5) 4.7.2", and "GNU gdb (Gentoo 7.5.1 p2) 7.5.1". p itest(1:2) once
again gives slice out of range, but p itest(1) gives 6316032, and p itest(2)
gives 0.  Values output by the code are correct.  The optimistic view is that
any change from the previous result is some sort of progress.

-- 
Configure bugmail: http://sourceware.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

[Bug fortran/9395] cannot access allocatable array in fortran90

[schneidermark at myfairpoint dot net]

http://sourceware.org/bugzilla/show_bug.cgi?id=9395

--- Comment #13 from Mark Schneider <schneidermark at myfairpoint dot net> 2013-02-20 17:39:05 UTC ---
Created attachment 6883
  --> http://sourceware.org/bugzilla/attachment.cgi?id=6883
simple allocatable array

-- 
Configure bugmail: http://sourceware.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

[Bug fortran/9395] cannot access allocatable array in fortran90

[jan.kratochvil at redhat dot com]

http://sourceware.org/bugzilla/show_bug.cgi?id=9395

--- Comment #14 from Jan Kratochvil <jan.kratochvil at redhat dot com> 2013-02-20 17:42:13 UTC ---
plan: VLA (Variable Length Arrays) and Fortran dynamic types
http://sourceware.org/ml/gdb/2012-11/msg00094.html

-- 
Configure bugmail: http://sourceware.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

[Bug fortran/9395] cannot access allocatable array in fortran90

[schneidermark at myfairpoint dot net]

http://sourceware.org/bugzilla/show_bug.cgi?id=9395

--- Comment #15 from Mark Schneider <schneidermark at myfairpoint dot net> 2013-02-20 17:46:22 UTC ---
Using GDB 7.4, configured as "i686-pc-mingw32", Windows 7 Professional SP1
64bit, with gfortran 4.8 and Code:Blocks 12.11. When you try to look at an
array that is "allocatable", you are provided with nonsense, and you don't even
get the right amount of nonsense. However, if you change from allocatable to a
fixed size array everything displays correctly, including size (see attached
simple allocatable array example. As noted previously, every IDE I have tried
that uses the gnu tools has this problem.

-- 
Configure bugmail: http://sourceware.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

[Bug fortran/9395] cannot access allocatable array in fortran90

[marcialcz at gmail dot com]

http://sourceware.org/bugzilla/show_bug.cgi?id=9395

Marcial <marcialcz at gmail dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |marcialcz at gmail dot com

--- Comment #16 from Marcial <marcialcz at gmail dot com> ---
Does not work either on gdb7.4 as x86_64-apple-darwin10.8.0 but! 

Steven Vancoillie has found a way around this that may be useful. He explains
in his blog:    
http://numericalnoob.blogspot.mx/2012/08/fortran-allocatable-arrays-and-pointers.html 

he basically does this:

real, allocatable :: variableName(:)
allocate(variableName(4))
variableName = (/1,2,3,4/)
...
(gdb) p *((real*) variableName)@4
$1 = (1,2,3,4)

(gdb) p *((real*) variableName + 2)
$2 = 3

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug fortran/9395] cannot access allocatable array in fortran90

[sanimir.agovic at intel dot com]

https://sourceware.org/bugzilla/show_bug.cgi?id=9395

sagovic <sanimir.agovic at intel dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |sanimir.agovic at intel dot com

--- Comment #17 from sagovic <sanimir.agovic at intel dot com> ---
(In reply to Jan Kratochvil from comment #14)
> plan: VLA (Variable Length Arrays) and Fortran dynamic types
> http://sourceware.org/ml/gdb/2012-11/msg00094.html
Based on Jans plan we implemented variable length arrays for c99 and Fortran.
It is based on HEAD and tested with gcc/gfortran as well as some other C and
Fortran compilers.
Sources and documentation can be found here http://intel-gdb.github.io/

We are currently upreaming c99 vla support for gdb.
https://sourceware.org/ml/gdb-patches/2013-10/msg00632.html

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug fortran/9395] cannot access allocatable array in fortran90

[jackie.rosen at hushmail dot com]

https://sourceware.org/bugzilla/show_bug.cgi?id=9395

Jackie Rosen <jackie.rosen at hushmail dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |jackie.rosen at hushmail dot com

--- Comment #18 from Jackie Rosen <jackie.rosen at hushmail dot com> ---
*** Bug 260998 has been marked as a duplicate of this bug. ***
Seen from the domain http://volichat.com
Page where seen: http://volichat.com/adult-chat-rooms
Marked for reference. Resolved as fixed @bugzilla.

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug fortran/9395] cannot access allocatable array in fortran90

[dharenberg at ethz dot ch]

https://sourceware.org/bugzilla/show_bug.cgi?id=9395

dharenberg at ethz dot ch changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |dharenberg at ethz dot ch

--- Comment #19 from dharenberg at ethz dot ch ---
Workaround/ partial solution

I had the same problem with this version:
$ gdb -version
GNU gdb (Ubuntu/Linaro 7.4-2012.04-0ubuntu2.1) 7.4-2012.04

But not with the version that came with the intel compiler:
$ gdb-ia -version
No symbol table is loaded.  Use the "file" command.
GNU gdb (GDB) 7.5-1.0.160

However, in both versions, when the allocatable array is a component of a
derived type, then the problem essentially reappears. An easy workaround for
this case is to use the ASSOCIATE construct:
associate(my_array => this%allocatable_array)

Then I can see the contents of my_array, but not of this%allocatable_array.

I am a beginner, so I do all this in Eclipse, where it works by setting the
Debugger to gdb-ia in the launch configuration properties (given that you have
the Intel compiler installed). 

I guess this partially confirms comment 17 by sagovic.

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug fortran/9395] cannot access allocatable array in fortran90

[schwab at sourceware dot org]

https://sourceware.org/bugzilla/show_bug.cgi?id=9395

Andreas Schwab <schwab at sourceware dot org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|jackie.rosen at hushmail dot com   |

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug fortran/9395] cannot access allocatable array in fortran90

[tromey at sourceware dot org]

https://sourceware.org/bugzilla/show_bug.cgi?id=9395

Tom Tromey <tromey at sourceware dot org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |tromey at sourceware dot org
             Status|UNCONFIRMED                 |RESOLVED
         Resolution|---                         |FIXED

--- Comment #22 from Tom Tromey <tromey at sourceware dot org> ---
Reported as fixed, plus there's been work on allocatable
arrays.  If there's still a problem, though, feel free to
reopen w/ details or file a new bug.

-- 
You are receiving this mail because:
You are on the CC list for the bug.