[Bug fortran/67987] New: ICE on declaring and initializing character with negative len

[Bug fortran/67987] New: ICE on declaring and initializing character with negative len

[gerhard.steinmetz.fortran@t-online.de]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=67987

            Bug ID: 67987
           Summary: ICE on declaring and initializing character with
                    negative len
           Product: gcc
           Version: 5.2.1
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: fortran
          Assignee: unassigned at gcc dot gnu.org
          Reporter: gerhard.steinmetz.fortran@t-online.de
  Target Milestone: ---

Declaring and initializing a character with negative len (< -1)
requests memory of about 2^64 (= 18446744073709551616) :


$ cat z1.f90
program p
   character(-8) :: c = ' '
end


$ gfortran -g -O0 -Wall -fcheck=all -fno-frontend-optimize z1.f90
f951: out of memory allocating 18446744073709551588 bytes after a total of
434176 bytes

---

Warning without assignment :

$ cat z2.f90
program p
   character(-8) :: c
end

$ gfortran -g -O0 -Wall -fcheck=all -fno-frontend-optimize z2.f90
z2.f90:2:13:

    character(-8) :: c
             1
Warning: CHARACTER variable at (1) has negative length -8, the length has been
set to zero [-Wsurprising]

---

With len = -1, reaction depends on used compiler options :

$ cat z3.f90
program p
   character(-1) :: c = 'c'
end

$ gfortran -g -O0 -Wall -fcheck=all -fno-frontend-optimize z3.f90
gfortran: internal compiler error: Segmentation fault (program f951)

---

Same issue with "character parameter".

[Bug fortran/67987] ICE on declaring and initializing character with negative len

[kargl at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=67987

kargl at gcc dot gnu.org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
   Last reconfirmed|                            |2015-10-16
                 CC|                            |kargl at gcc dot gnu.org
           Assignee|unassigned at gcc dot gnu.org      |kargl at gcc dot gnu.org
     Ever confirmed|0                           |1
      Known to fail|                            |4.6.4, 4.7.4, 4.8.5, 4.9.4,
                   |                            |5.2.1, 6.0

--- Comment #1 from kargl at gcc dot gnu.org ---
I have a patch that I'm testing.  For the code in the
intial report I get

troutmask:sgk[238] gfc6 -c g6.f90
g6.f90:2:15:

    character(-8) :: c = ' '
               1
Error: LEN at (1) must be greater than or equal to 0

[Bug fortran/67987] ICE on declaring and initializing character with negative len

[kargl at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=67987

--- Comment #2 from kargl at gcc dot gnu.org ---
(In reply to kargl from comment #1)
> I have a patch that I'm testing.  For the code in the
> intial report I get
> 
> troutmask:sgk[238] gfc6 -c g6.f90
> g6.f90:2:15:
> 
>     character(-8) :: c = ' '
>                1
> Error: LEN at (1) must be greater than or equal to 0

Seems the patch causes a regression for char_length_2.f90.
This test came in with the fix for PR fortran/31250, where
negative length parameters are silently set to 0.  The 
fix for this regression is in resolve.c.  Unfortnately, 
this fix then exposes an issue with substring referecences.
For "string"(1:-1), this should result in a 0 length string,
but gfortran is setting the length to -1 in 
gfc_resolve_substring_charlen.

[Bug fortran/67987] ICE on declaring and initializing character with negative len

[sgk at troutmask dot apl.washington.edu]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=67987

--- Comment #3 from Steve Kargl <sgk at troutmask dot apl.washington.edu> ---
On Fri, Oct 16, 2015 at 06:03:06PM +0000, kargl at gcc dot gnu.org wrote:
> https://gcc.gnu.org/bugzilla/show_bug.cgi?id=67987
> 
> --- Comment #2 from kargl at gcc dot gnu.org ---
> (In reply to kargl from comment #1)
> > I have a patch that I'm testing.  For the code in the
> > intial report I get
> > 
> > troutmask:sgk[238] gfc6 -c g6.f90
> > g6.f90:2:15:
> > 
> >     character(-8) :: c = ' '
> >                1
> > Error: LEN at (1) must be greater than or equal to 0
> 
> Seems the patch causes a regression for char_length_2.f90.
> This test came in with the fix for PR fortran/31250, where
> negative length parameters are silently set to 0.  The 
> fix for this regression is in resolve.c.  Unfortnately, 
> this fix then exposes an issue with substring referecences.
> For "string"(1:-1), this should result in a 0 length string,
> but gfortran is setting the length to -1 in 
> gfc_resolve_substring_charlen.
> 

Go the substring issued ironed out.

[Bug fortran/67987] ICE on declaring and initializing character with negative len

[kargl at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=67987

--- Comment #4 from kargl at gcc dot gnu.org ---
*** Bug 67988 has been marked as a duplicate of this bug. ***

[Bug fortran/67987] ICE on declaring and initializing character with negative len

[kargl at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=67987

--- Comment #5 from kargl at gcc dot gnu.org ---
Author: kargl
Date: Sat Oct 17 16:50:47 2015
New Revision: 228933

URL: https://gcc.gnu.org/viewcvs?rev=228933&root=gcc&view=rev
Log:
2015-10-17  Steven G. Kargl  <kargl@gcc.gnu.org>

        PR fortran/67987
        * decl.c (char_len_param_value): Unwrap unlong line.  If LEN < 0,
        force it to zero per the Fortran 90, 95, 2003, and 2008 Standards. 
        * resolve.c (gfc_resolve_substring_charlen): Unwrap unlong line.
        If 'start' is larger than 'end', length of substring is negative,
        so explicitly set it to zero.
        (resolve_charlen): Remove -Wsurprising warning.  Update comment to
        reflect that the text is from the F2008 standard.

2015-10-17  Steven G. Kargl  <kargl@gcc.gnu.org>

        PR fortran/67987
        * gfortran.df/pr67987.f90: New test.
        * gfortran.dg/char_length_2.f90: Update testcase.

Added:
    trunk/gcc/testsuite/gfortran.dg/pr67987.f90
Modified:
    trunk/gcc/fortran/ChangeLog
    trunk/gcc/fortran/decl.c
    trunk/gcc/fortran/resolve.c
    trunk/gcc/testsuite/ChangeLog
    trunk/gcc/testsuite/gfortran.dg/char_length_2.f90

[Bug fortran/67987] ICE on declaring and initializing character with negative len

[kargl at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=67987

kargl at gcc dot gnu.org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|---                         |FIXED
   Target Milestone|---                         |5.3

--- Comment #7 from kargl at gcc dot gnu.org ---
Fixed on trunk and 5-branch.  Thanks for the bug report.