35960 – run time abort with assignment of RESHAPEd zero sized array

Bug 35960 - run time abort with assignment of RESHAPEd zero sized array

Summary: run time abort with assignment of RESHAPEd zero sized array

Status:	RESOLVED FIXED

Alias:	None

Product:	gcc
Classification:	Unclassified
Component:	fortran (show other bugs)
Version:	4.4.0

Importance:	P3 normal
Target Milestone:	---
Assignee:	Thomas Koenig

URL:
Keywords:	wrong-code

Depends on:
Blocks:

Reported:	2008-04-16 21:18 UTC by Dick Hendrickson
Modified:	2008-04-20 20:07 UTC (History)
CC List:	2 users (show)

See Also:
Host:
Target:
Build:
Known to work:
Known to fail:
Last reconfirmed:	2008-04-17 21:58:22

Attachments
proposed patch (855 bytes, patch) 2008-04-19 14:11 UTC, Thomas Koenig	Details \| Diff
View All Add an attachment (proposed patch, testcase, etc.)

Note You need to log in before you can comment on or make changes to this bug.

Description Dick Hendrickson 2008-04-16 21:18:59 UTC

Each line of the following subroutine causes a run-time
abort. Except for the one labeled gf1069.  It works if
literal constants are used in place of the nf* variables.

Dick Hendrickson

      program try_gf1065


! fails on Windows XP
! gcc version 4.4.0 20080312 (experimental) [trunk revision 133139]


      call       gf1065(1,  2,  3,  4,  7,  8,  9)
      end

      SUBROUTINE GF1065(nf1,nf2,nf3,nf4,nf7,nf8,nf9)
!        also 1066 thru 1070, 1069 works

      REAL RDA(10,9)
      REAL RCA1(90)
      integer ila(2)
!gf1065
      RDA(NF9:NF8, NF7:NF3) = RESHAPE(RCA1,(/0,0/), (/1.0/),(/2,1/))
      print *, 'gf1065'

!gf1066
      rDA(NF9:NF8, NF7:NF3) = RESHAPE(rCA1,(/0,0/),ORDER=(/2,1/))
      print *, 'gf1066'

      ILA(1) = 5
      ILA(2) = 0
!gf1067
      rDA(NF4:NF8, NF7:NF3) = RESHAPE(rcA1,ILA)
      print *, 'gf1067'

!gf1068
      RdA(NF4:NF8, NF7:NF3) = RESHAPE(RcA1,ILA,PAD=(/-1.0/))
      print *, 'gf1068'

      ILA(1) = 0
      ILA(2) = 5
!gf1069    this one works
      RdA(NF9:NF8,NF4:NF8)=RESHAPE(RcA1,ILA,(/-1.0/),(/NF2,NF1/))
      print *, 'gf1069'

      ILA(1) = 5
      ILA(2) = 0
!gf1070
      RdA(NF4:NF8, NF7:NF3) = RESHAPE(RcA1,ILA,ORDER=(/NF1,NF2/))
      print *, 'gf1070'


      END SUBROUTINE


C:\gfortran:gfortran gf1065.f

C:\gfortran:a
Fortran runtime error: shape and target do not conform

Comment 1 Thomas Koenig 2008-04-17 21:58:22 UTC

Confirmed.

I'll take a look at this.

Comment 2 Thomas Koenig 2008-04-19 11:29:37 UTC

We don't check for empty arrays in reshape.

Should be fairly straightforward.

Comment 3 Thomas Koenig 2008-04-19 14:11:37 UTC

Created attachment 15496 [details]
proposed patch

Here's a patch.

I'll regression-test and commit tomorrow, probably.

Comment 4 Thomas Koenig 2008-04-20 19:56:53 UTC

Subject: Bug 35960

Author: tkoenig
Date: Sun Apr 20 19:56:07 2008
New Revision: 134490

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=134490
Log:
2008-04-20  Thomas Koenig  <tkoenig@gcc.gnu.org>

	PR libfortran/35960
	* intrinsics/reshape_generic.c (reshape_internal): If the size
	of the resized array is zero, as determined by the SHAPE
	argument, return early.
	* m4/reshape.m4:  Likewise.
	* generated/reshape_i4.c:  Regererated.
	* generated/reshape_i8.c:  Regenerated.
	* generated/reshape_i16.c:  Regenerated.
	* generated/reshape_r4.c:  Regenerated.
	* generated/reshape_r8.c:  Regenerated.
	* generated/reshape_r10.c:  Regenerated.
	* generated/reshape_r16.c:  Regenerated.
	* generated/reshape_c4.c:  Regenerated.
	* generated/reshape_c8.c:  Regenerated.
	* generated/reshape_c10.c:  Regenerated.
	* generated/reshape_c16.c:  Regenerated.

2008-04-20  Thomas Koenig  <tkoenig@gcc.gnu.org>

	PR libfortran/35960
	* gfortran.dg/reshape_zerosize_1.f90:  New file.


Added:
    trunk/gcc/testsuite/gfortran.dg/reshape_zerosize_1.f90
Modified:
    trunk/gcc/testsuite/ChangeLog
    trunk/libgfortran/ChangeLog
    trunk/libgfortran/generated/reshape_c10.c
    trunk/libgfortran/generated/reshape_c16.c
    trunk/libgfortran/generated/reshape_c4.c
    trunk/libgfortran/generated/reshape_c8.c
    trunk/libgfortran/generated/reshape_i16.c
    trunk/libgfortran/generated/reshape_i4.c
    trunk/libgfortran/generated/reshape_i8.c
    trunk/libgfortran/generated/reshape_r10.c
    trunk/libgfortran/generated/reshape_r16.c
    trunk/libgfortran/generated/reshape_r4.c
    trunk/libgfortran/generated/reshape_r8.c
    trunk/libgfortran/intrinsics/reshape_generic.c
    trunk/libgfortran/m4/reshape.m4

Comment 5 Thomas Koenig 2008-04-20 20:07:45 UTC

Fixed on trunk (no need to backport this to 4.3).

Closing.

Comment 6 Thomas Koenig 2008-04-25 20:12:09 UTC

Subject: Bug 35960

Author: tkoenig
Date: Fri Apr 25 20:11:19 2008
New Revision: 134677

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=134677
Log:
2008-04-25  Thomas Koenig  <tkoenig@gcc.gnu.org>

	PR libfortran/35960
	* m4/reshape.m4:  Fix typo in last commit.
	* generated/reshape_i4.c:  Regererated.
	* generated/reshape_i8.c:  Regenerated.
	* generated/reshape_i16.c:  Regenerated.
	* generated/reshape_r4.c:  Regenerated.
	* generated/reshape_r8.c:  Regenerated.
	* generated/reshape_r10.c:  Regenerated.
	* generated/reshape_r16.c:  Regenerated.
	* generated/reshape_c4.c:  Regenerated.
	* generated/reshape_c8.c:  Regenerated.
	* generated/reshape_c10.c:  Regenerated.
	* generated/reshape_c16.c:  Regenerated.


Modified:
    trunk/libgfortran/ChangeLog
    trunk/libgfortran/generated/reshape_c10.c
    trunk/libgfortran/generated/reshape_c16.c
    trunk/libgfortran/generated/reshape_c4.c
    trunk/libgfortran/generated/reshape_c8.c
    trunk/libgfortran/generated/reshape_i16.c
    trunk/libgfortran/generated/reshape_i4.c
    trunk/libgfortran/generated/reshape_i8.c
    trunk/libgfortran/generated/reshape_r10.c
    trunk/libgfortran/generated/reshape_r16.c
    trunk/libgfortran/generated/reshape_r4.c
    trunk/libgfortran/generated/reshape_r8.c
    trunk/libgfortran/m4/reshape.m4