[Bug fortran/31591] New: internal compiler error on array initialization statement using 'ubound'

[Bug fortran/31591] New: internal compiler error on array initialization statement using 'ubound'

[Catherine dot M dot Moroney at jpl dot nasa dot gov]

See the bit of code below dealing with the initialization of array contents
with
the "ubound" results called on another array.  If the "module" statements plus
the write(*,*) statements are both commented out, then this bit of code
compiles.  However, if either (a) this code is made into a module, or
(b) it remains as a program but the 'write(*,*)' statement is put back in, then
the compilation fails with an "internal compiler error: in
gfc_conv_array_initializer, 
at fortran/trans-array.c:3460"

!MODULE TEST_CERES

PROGRAM Test_ceres

implicit none

real, dimension(10, 10,  10, 3, 4) ::  TCg_coef 
integer, parameter, dimension(5)  :: &
    TCg_bband_coef_ranks = ubound(TCg_coef)

!write(*,*) 'array_ranks = ',TCg_bband_coef_ranks

end program Test_ceres
!end module Test_ceres


-- 
           Summary: internal compiler error on array initialization
                    statement using 'ubound'
           Product: gcc
           Version: unknown
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: fortran
        AssignedTo: unassigned at gcc dot gnu dot org
        ReportedBy: Catherine dot M dot Moroney at jpl dot nasa dot gov


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=31591

[Bug fortran/31591] internal compiler error on array initialization statement using 'ubound'

[dfranke at gcc dot gnu dot org]

------- Comment #1 from dfranke at gcc dot gnu dot org  2007-04-16 20:05 -------
Tested gfortran v4.1.1, v4.2 and a recent svn checkout, none of them reported
an ICE. Which version do you use?


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=31591

[Bug fortran/31591] internal compiler error on array initialization statement using 'ubound'

[Catherine dot M dot Moroney at jpl dot nasa dot gov]

------- Comment #2 from Catherine dot M dot Moroney at jpl dot nasa dot gov  2007-04-16 20:07 -------
Subject: Re:  internal compiler error on array initialization statement using
'ubound'

cmm@sieglind.jpl.nasa.gov:/data/svn_workspace/cmm/sieglind/PGE/L2TC/ 
StandardProcessing/AZM/make  [69]>gfortran -v   Using built-in specs.
Target: x86_64-redhat-linux
Configured with: ../configure --prefix=/usr --mandir=/usr/share/man  
--infodir=/usr/share/info --enable-shared --enable-threads=posix  
--enable-checking=release --with-system-zlib --enable-__cxa_atexit  
--disable-libunwind-exceptions --enable-libgcj-multifile  
--enable-languages=c,c++,objc,obj-c++,java,fortran,ada  
--enable-java-awt=gtk --disable-dssi  
--with-java-home=/usr/lib/jvm/java-1.4.2-gcj-1.4.2.0/jre  
--with-cpu=generic --host=x86_64-redhat-linux
Thread model: posix
gcc version 4.1.1 20070105 (Red Hat 4.1.1-51)

Catherine

On Apr 16, 2007, at 12:05 PM, dfranke at gcc dot gnu dot org wrote:

>
>
> ------- Comment #1 from dfranke at gcc dot gnu dot org  2007-04-16  
> 20:05 -------
> Tested gfortran v4.1.1, v4.2 and a recent svn checkout, none of them  
> reported
> an ICE. Which version do you use?
>
>
> -- 
>
>
> http://gcc.gnu.org/bugzilla/show_bug.cgi?id=31591
>
> ------- You are receiving this mail because: -------
> You reported the bug, or are watching the reporter.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=31591

[Bug fortran/31591] ICE: on array initialization statement using 'ubound' (fortran/trans-array.c:3693)

[dfranke at gcc dot gnu dot org]

------- Comment #3 from dfranke at gcc dot gnu dot org  2007-04-16 20:42 -------
Ups, misread your report. After uncommenting the WRITE, all tested compilers
crash. 

$> gfortran -v
gcc version 4.3.0 20070331

Backtrace:
Breakpoint 1, fancy_abort (file=0x86e4fec
"../../../gcc/gcc/fortran/trans-array.c", line=3693, function=0x86e55a0
"gfc_conv_array_initializer")
    at ../../../gcc/gcc/diagnostic.c:656
656       internal_error ("in %s, at %s:%d", function, trim_filename (file),
line);
(gdb) bt
#0  fancy_abort (file=0x86e4fec "../../../gcc/gcc/fortran/trans-array.c",
line=3693, function=0x86e55a0 "gfc_conv_array_initializer")
    at ../../../gcc/gcc/diagnostic.c:656
#1  0x080b7183 in gfc_conv_array_initializer (type=0xb7ce4ec4, expr=0x8894d38)
at ../../../gcc/gcc/fortran/trans-array.c:3693
#2  0x080d093c in gfc_conv_initializer (expr=0x8894d38, ts=0x8894720,
type=0xb7ce4ec4, array=160 ' ', pointer=109 'm')
    at ../../../gcc/gcc/fortran/trans-expr.c:2740
#3  0x080c5f86 in gfc_get_symbol_decl (sym=0x8894710) at
../../../gcc/gcc/fortran/trans-decl.c:1018
#4  0x080ccb9d in gfc_conv_expr (se=0xbfea9304, expr=0x8894fd8) at
../../../gcc/gcc/fortran/trans-expr.c:391
#5  0x080cfa9e in gfc_conv_expr_lhs (se=0xbfea9304, expr=0x8894fd8) at
../../../gcc/gcc/fortran/trans-expr.c:3167
#6  0x080b77e6 in gfc_conv_ss_descriptor (block=0xbfea95e4, ss=0x8895220,
base=0) at ../../../gcc/gcc/fortran/trans-array.c:1838
#7  0x080ba62c in gfc_conv_expr_descriptor (se=0xbfea95e4, expr=0x8894fd8,
ss=0x8895220) at ../../../gcc/gcc/fortran/trans-array.c:4316
#8  0x080dac5e in gfc_trans_transfer (code=0x8895100) at
../../../gcc/gcc/fortran/trans-io.c:1838
#9  0x080b075d in gfc_trans_code (code=0x8895100) at
../../../gcc/gcc/fortran/trans.c:605
#10 0x080dd05d in build_dt (function=0xb7cdce0c, code=0x88951c0) at
../../../gcc/gcc/fortran/trans-io.c:1507
#11 0x080b073d in gfc_trans_code (code=0x88951c0) at
../../../gcc/gcc/fortran/trans.c:581
#12 0x080c7d06 in gfc_generate_function_code (ns=0x885a1e8) at
../../../gcc/gcc/fortran/trans-decl.c:3217
#13 0x0808c2b4 in gfc_parse_file () at ../../../gcc/gcc/fortran/parse.c:3261
#14 0x080ab660 in gfc_be_parse_file (set_yydebug=0) at
../../../gcc/gcc/fortran/f95-lang.c:305
#15 0x0830042d in toplev_main (argc=3, argv=0xbfea9954) at
../../../gcc/gcc/toplev.c:1050
#16 0x080ed162 in main (argc=1600349799, argv=0x766e6f63) at
../../../gcc/gcc/main.c:35


-- 

dfranke at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |dfranke at gcc dot gnu dot
                   |                            |org
             Status|UNCONFIRMED                 |NEW
     Ever Confirmed|0                           |1
   Last reconfirmed|0000-00-00 00:00:00         |2007-04-16 20:42:01
               date|                            |
            Summary|internal compiler error on  |ICE: on array initialization
                   |array initialization        |statement using 'ubound'
                   |statement using 'ubound'    |(fortran/trans-array.c:3693)
            Version|unknown                     |4.1.1


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=31591

[Bug fortran/31591] ICE: on array initialization statement using 'ubound' (fortran/trans-array.c:3693)

[burnus at gcc dot gnu dot org]

------- Comment #4 from burnus at gcc dot gnu dot org  2007-04-16 21:14 -------
The assert is trans-array.c's gfc_conv_array_initializer
  switch (expr->expr_type)
    {
    ...
    default:
      gcc_unreachable ();
Here, expr->expr_type is EXPR_FUNCTION.

By the way, the scalar version, i.e. removing "dimension" and using
"ubound(TCg_coef,1)", works.


-- 

burnus at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |burnus at gcc dot gnu dot
                   |                            |org
           Keywords|                            |ice-on-valid-code


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=31591

[Bug fortran/31591] UBOUND as initialization expression (lacking simplification)

[fxcoudert at gcc dot gnu dot org]

------- Comment #5 from fxcoudert at gcc dot gnu dot org  2007-04-17 22:48 -------
(In reply to comment #4)
> By the way, the scalar version, i.e. removing "dimension" and using
> "ubound(TCg_coef,1)", works.

Hum, I know that one (although I never thought it would trigger a bug when I
encountered it): the simplication function for {U,L}BOUND doesn't handle the
array case (ie when the second parameter is not present). It should be written,
using the other array simplication functions as guidance.

PS: other array intrinsics should be audited for the same problem


-- 

fxcoudert at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |fxcoudert at gcc dot gnu dot
                   |                            |org
      Known to fail|                            |4.1.3 4.2.0 4.3.0
   Last reconfirmed|2007-04-16 20:42:01         |2007-04-17 22:48:46
               date|                            |
            Summary|ICE: on array initialization|UBOUND as initialization
                   |statement using 'ubound'    |expression (lacking
                   |(fortran/trans-array.c:3693)|simplification)


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=31591

[Bug fortran/31591] UBOUND as initialization expression (lacking simplification)

[Catherine dot M dot Moroney at jpl dot nasa dot gov]

------- Comment #6 from Catherine dot M dot Moroney at jpl dot nasa dot gov  2007-04-17 23:09 -------
Subject: Re:  UBOUND as initialization expression (lacking simplification)

So, are you saying that LBOUND and UBOUND without the dimension
specifiers do not work at all?  Take the attached short test
program as an example.  If I comment out the assignment statement
from TCg_bband_coef_ranks to rank, then it compiles, but does it
actually work?  Is this program deceiving in that it compiles, but
underneath, is actually not working?  That's not good!

Catherine



On Apr 17, 2007, at 2:48 PM, fxcoudert at gcc dot gnu dot org wrote:

>
>
> ------- Comment #5 from fxcoudert at gcc dot gnu dot org  2007-04-17  
> 22:48 -------
> (In reply to comment #4)
>> By the way, the scalar version, i.e. removing "dimension" and using
>> "ubound(TCg_coef,1)", works.
>
> Hum, I know that one (although I never thought it would trigger a bug  
> when I
> encountered it): the simplication function for {U,L}BOUND doesn't  
> handle the
> array case (ie when the second parameter is not present). It should be  
> written,
> using the other array simplication functions as guidance.
>
> PS: other array intrinsics should be audited for the same problem
>
>
> --  
>
> fxcoudert at gcc dot gnu dot org changed:
>
>            What    |Removed                     |Added
> ----------------------------------------------------------------------- 
> -----
>                  CC|                            |fxcoudert at gcc dot  
> gnu dot
>                    |                            |org
>       Known to fail|                            |4.1.3 4.2.0 4.3.0
>    Last reconfirmed|2007-04-16 20:42:01         |2007-04-17 22:48:46
>                date|                            |
>             Summary|ICE: on array initialization|UBOUND as  
> initialization
>                    |statement using 'ubound'    |expression (lacking
>                    |(fortran/trans-array.c:3693)|simplification)
>
>
> http://gcc.gnu.org/bugzilla/show_bug.cgi?id=31591
>
> ------- You are receiving this mail because: -------
> You reported the bug, or are watching the reporter.


------- Comment #7 from Catherine dot M dot Moroney at jpl dot nasa dot gov  2007-04-17 23:09 -------
Created an attachment (id=13383)
 --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=13383&action=view)


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=31591

[Bug fortran/31591] UBOUND as initialization expression (lacking simplification)

[fxcoudert at gcc dot gnu dot org]

------- Comment #8 from fxcoudert at gcc dot gnu dot org  2007-04-17 23:22 -------
(In reply to comment #6)
> So, are you saying that LBOUND and UBOUND without the dimension
> specifiers do not work at all?

I'm saying that gfortran is missing the piece of code that allows is to use
UBOUND and LBOUND, without the DIM argument, as specification expression (ie as
value assigned to a PARAMETER). So:

integer, parameter, dimension(5)  :: &
    TCg_bband_coef_ranks = ubound(TCg_coef)

does not work, but

integer, parameter, dimension(5)  :: &
    TCg_bband_coef_ranks = (/ ubound(TCg_coef,1), ubound(TCg_coef,2), ... /)

works (if you fill the ...) and also the following works:

integer, dimension(5)  :: &
    TCg_bband_coef_ranks = ubound(TCg_coef)

> If I comment out the assignment statement
> from TCg_bband_coef_ranks to rank, then it compiles, but does it
> actually work?  Is this program deceiving in that it compiles, but
> underneath, is actually not working?  That's not good!

I *think* I see what you mean. When I say "it doesn't work", I mean "it doesn't
work if you try to use that PARAMETER later". When a PARAMETER is unused, the
missing bit of code inside gfortran is not triggered, and there's no failure to
compile. So it's not deveiving in the sense of "compiling but not working".


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=31591

[Bug fortran/31591] UBOUND as initialization expression (lacking simplification)

[fxcoudert at gcc dot gnu dot org]

------- Comment #9 from fxcoudert at gcc dot gnu dot org  2007-04-17 23:51 -------
Created an attachment (id=13384)
 --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=13384&action=view)
Patch to fix this PR

This patch should take care of the simplification. It hasn't been fully tested
yet, but it appears to work fine on your code and a few other cases I could
think of. It could take a week before I can formally submit it (thesis under
writing...)

Thanks for the bug report!


-- 

fxcoudert at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
         AssignedTo|unassigned at gcc dot gnu   |fxcoudert at gcc dot gnu dot
                   |dot org                     |org
             Status|NEW                         |ASSIGNED


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=31591

[Bug fortran/31591] UBOUND as initialization expression (lacking simplification)

[fxcoudert at gcc dot gnu dot org]

------- Comment #10 from fxcoudert at gcc dot gnu dot org  2007-04-18 00:08 -------
Created an attachment (id=13385)
 --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=13385&action=view)
Better patch


-- 

fxcoudert at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
  Attachment #13384|0                           |1
        is obsolete|                            |


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=31591

[Bug fortran/31591] UBOUND as initialization expression (lacking simplification)

[Catherine dot M dot Moroney at jpl dot nasa dot gov]

------- Comment #11 from Catherine dot M dot Moroney at jpl dot nasa dot gov  2007-04-18 00:12 -------
Subject: Re:  UBOUND as initialization expression (lacking simplification)

Well, thank-you for the quick turnaround on this patch!

Francois - good luck on your thesis!  I remember the struggle I had
getting mine finished.  Due on Monday, still running data Saturday
night ...  What's your thesis on?

Catherine

On Apr 17, 2007, at 3:51 PM, fxcoudert at gcc dot gnu dot org wrote:

>
>
> ------- Comment #9 from fxcoudert at gcc dot gnu dot org  2007-04-17  
> 23:51 -------
> Created an attachment (id=13384)
 --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=13384&action=view)
>  --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=13384&action=view)
> Patch to fix this PR
>
> This patch should take care of the simplification. It hasn't been  
> fully tested
> yet, but it appears to work fine on your code and a few other cases I  
> could
> think of. It could take a week before I can formally submit it (thesis  
> under
> writing...)
>
> Thanks for the bug report!
>
>
> --  
>
> fxcoudert at gcc dot gnu dot org changed:
>
>            What    |Removed                     |Added
> ----------------------------------------------------------------------- 
> -----
>          AssignedTo|unassigned at gcc dot gnu   |fxcoudert at gcc dot  
> gnu dot
>                    |dot org                     |org
>              Status|NEW                         |ASSIGNED
>
>
> http://gcc.gnu.org/bugzilla/show_bug.cgi?id=31591
>
> ------- You are receiving this mail because: -------
> You reported the bug, or are watching the reporter.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=31591

[Bug fortran/31591] UBOUND as initialization expression (lacking simplification)

[fxcoudert at gcc dot gnu dot org]

-- 

fxcoudert at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                URL|                            |http://gcc.gnu.org/ml/gcc-
                   |                            |patches/2007-
                   |                            |04/msg01990.html
           Keywords|                            |patch
   Target Milestone|---                         |4.3.0


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=31591

[Bug fortran/31591] UBOUND as initialization expression (lacking simplification)

[fxcoudert at gcc dot gnu dot org]

------- Comment #12 from fxcoudert at gcc dot gnu dot org  2007-04-29 17:04 -------
Subject: Bug 31591

Author: fxcoudert
Date: Sun Apr 29 17:03:58 2007
New Revision: 124281

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=124281
Log:
        PR fortran/31591

        * simplify.c (simplify_bound_dim): New function.
        (simplify_bound): Use the above. Perform simplification of LBOUND
        and UBOUND when DIM argument is not present.

        * gfortran.dg/bound_simplification_1.f90: New test.

Added:
    trunk/gcc/testsuite/gfortran.dg/bound_simplification_1.f90
Modified:
    trunk/gcc/fortran/ChangeLog
    trunk/gcc/fortran/simplify.c
    trunk/gcc/testsuite/ChangeLog


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=31591

[Bug fortran/31591] UBOUND as initialization expression (lacking simplification)

[fxcoudert at gcc dot gnu dot org]

------- Comment #13 from fxcoudert at gcc dot gnu dot org  2007-04-29 17:36 -------
Fixed on mainline.


-- 

fxcoudert at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|ASSIGNED                    |RESOLVED
         Resolution|                            |FIXED


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=31591