[Bug fortran/37319] New: [4.4 Regression] FAIL: gfortran.dg/function_kinds_5.f90

[Bug fortran/37319] New: [4.4 Regression] FAIL: gfortran.dg/function_kinds_5.f90

[dominiq at lps dot ens dot fr]

Between revisions 139588 (working) and 139622 (broken), the test
gfortran.dg/function_kinds_5.f90 started to fail: the expected error is no
longer emitted:

[ibook-dhum] f90/bug% gfc -c
/opt/gcc/_gcc_clean/gcc/testsuite/gfortran.dg/function_kinds_5.f90
[ibook-dhum] f90/bug% gfortran -c
/opt/gcc/_gcc_clean/gcc/testsuite/gfortran.dg/function_kinds_5.f90
/opt/gcc/_gcc_clean/gcc/testsuite/gfortran.dg/function_kinds_5.f90:8.6:

real (bad_kind(0d0)) function foo () ! { dg-error "must be an intrinsic or" }
     1
Error: Function 'bad_kind' in initialization expression at (1) must be an
intrinsic or a specification function

where gfortran is 4.3.2.

I have reported it and other regressions which appeared at this time in comment
#9 of pr37243.


-- 
           Summary: [4.4 Regression] FAIL: gfortran.dg/function_kinds_5.f90
           Product: gcc
           Version: 4.4.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: fortran
        AssignedTo: unassigned at gcc dot gnu dot org
        ReportedBy: dominiq at lps dot ens dot fr
 GCC build triplet: i686-apple-darwin9
  GCC host triplet: i686-apple-darwin9
GCC target triplet: i686-apple-darwin9


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37319

[Bug fortran/37319] [4.4 Regression] FAIL: gfortran.dg/function_kinds_5.f90

[jvdelisle at gcc dot gnu dot org]

------- Comment #1 from jvdelisle at gcc dot gnu dot org  2008-09-02 02:47 -------
Works for me here with latest trunk. x86-64-linux


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37319

[Bug fortran/37319] [4.4 Regression] FAIL: gfortran.dg/function_kinds_5.f90

[rguenth at gcc dot gnu dot org]

-- 

rguenth at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|---                         |4.4.0


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37319

[Bug fortran/37319] [4.4 Regression] FAIL: gfortran.dg/function_kinds_5.f90

[jakub at gcc dot gnu dot org]

-- 

jakub at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Priority|P3                          |P4


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37319

[Bug fortran/37319] [4.4 Regression] FAIL: gfortran.dg/function_kinds_5.f90

[dominiq at lps dot ens dot fr]

------- Comment #2 from dominiq at lps dot ens dot fr  2008-09-07 18:25 -------
Confirmed on http://gcc.gnu.org/ml/gcc-testresults/2008-09/msg00195.html and
http://gcc.gnu.org/ml/gcc-testresults/2008-09/msg00505.html.

This looks like a missing or wrong initialisation.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37319

[Bug fortran/37319] [4.4 Regression] FAIL: gfortran.dg/function_kinds_5.f90

[dominiq at lps dot ens dot fr]

------- Comment #3 from dominiq at lps dot ens dot fr  2008-09-07 18:35 -------
Note that in the window 139588 to 139622 there is no change in gfortran. The
most important change is the IRA merge at 139590. This may have caused a
miscompilation of some gfortran code (this happened with transfer) or only
exposed the problem on i686-apple-darwin9 (I don't see it on ppc).

Any idea on how to debug this pr?


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37319

[Bug fortran/37319] [4.4 Regression] FAIL: gfortran.dg/function_kinds_5.f90

[dominiq at lps dot ens dot fr]

------- Comment #4 from dominiq at lps dot ens dot fr  2008-09-11 16:33 -------
At r140286 with the patch in
http://gcc.gnu.org/ml/fortran/2008-09/msg00210.html, the failure is gone!-(


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37319

[Bug fortran/37319] FAIL the unexplained fix: gfortran.dg/function_kinds_5.f90

[pault at gcc dot gnu dot org]

------- Comment #5 from pault at gcc dot gnu dot org  2008-10-05 19:03 -------
(In reply to comment #4)
> At r140286 with the patch in
> http://gcc.gnu.org/ml/fortran/2008-09/msg00210.html, the failure is gone!-(
> 

I cannot explain it, other than the old version using something that was not
initialized, as you suggest.  If you do not mind, since it has cleared and the
testcase remains in place, I'll remove the regression label but will retain the
PR.

Cheers

Paul


-- 

pault at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
     Ever Confirmed|0                           |1
   Last reconfirmed|0000-00-00 00:00:00         |2008-10-05 19:03:35
               date|                            |
            Summary|[4.4 Regression] FAIL:      |FAIL the unexplained fix:
                   |gfortran.dg/function_kinds_5|gfortran.dg/function_kinds_5
                   |.f90                        |.f90


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37319

[Bug fortran/37319] [4.4 regression] gfortran.dg/function_kinds_5.f90 fails

[ebotcazou at gcc dot gnu dot org]

------- Comment #6 from ebotcazou at gcc dot gnu dot org  2008-11-24 19:32 -------
The problem still reproduces on the SPARC as of today.  When the compiler is
rebuilt at -O0, it goes away; when decl.c and parser.c are rebuilt at -O2, it
comes back.


-- 

ebotcazou at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
  GCC build triplet|i686-apple-darwin9          |
   GCC host triplet|i686-apple-darwin9          |
 GCC target triplet|i686-apple-darwin9          |
   Last reconfirmed|2008-10-05 19:03:35         |2008-11-24 19:32:05
               date|                            |
            Summary|FAIL the unexplained fix:   |[4.4 regression]
                   |gfortran.dg/function_kinds_5|gfortran.dg/function_kinds_5
                   |.f90                        |.f90 fails


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37319

[Bug fortran/37319] [4.4 regression] gfortran.dg/function_kinds_5.f90 fails

[ebotcazou at gcc dot gnu dot org]

------- Comment #7 from ebotcazou at gcc dot gnu dot org  2008-11-24 19:32 -------
Investigating.


-- 

ebotcazou at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
         AssignedTo|unassigned at gcc dot gnu   |ebotcazou at gcc dot gnu dot
                   |dot org                     |org
             Status|NEW                         |ASSIGNED
   Last reconfirmed|2008-11-24 19:32:05         |2008-11-24 19:32:44
               date|                            |


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37319

[Bug fortran/37319] [4.4 regression] gfortran.dg/function_kinds_5.f90 fails

[ebotcazou at gcc dot gnu dot org]

------- Comment #8 from ebotcazou at gcc dot gnu dot org  2008-11-24 19:59 -------
> This looks like a missing or wrong initialisation.

Confirmed, it's 'name' in match_deferred_characteristics:

  char name[GFC_MAX_SYMBOL_LEN + 1];
[...]
  /* Set the function locus correctly.  If we have not found the
     function name, there is an error.  */
  gfc_match ("function% %n", name);
  if (m == MATCH_YES && strcmp (name, gfc_current_block ()->name) == 0)
    {
      gfc_current_block ()->declared_at = gfc_current_locus;
      gfc_commit_symbols ();
    }
  else
    gfc_error_check ();

gfc_match ("function% %n", name) doesn't touch 'name' so the outcome of the
string comparison is random.  Would it make sense to test the return value
of gfc_match here?


-- 

ebotcazou at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |pault at gcc dot gnu dot org


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37319

[Bug fortran/37319] [4.4 regression] gfortran.dg/function_kinds_5.f90 fails

[kargl at gcc dot gnu dot org]

------- Comment #9 from kargl at gcc dot gnu dot org  2008-11-24 20:24 -------
(In reply to comment #8)
> > This looks like a missing or wrong initialisation.
> 
> Confirmed, it's 'name' in match_deferred_characteristics:
> 
>   char name[GFC_MAX_SYMBOL_LEN + 1];
> [...]
>   /* Set the function locus correctly.  If we have not found the
>      function name, there is an error.  */
>   gfc_match ("function% %n", name);

This should probably be

    m = gfc_match ("function% %n", name);

>   if (m == MATCH_YES && strcmp (name, gfc_current_block ()->name) == 0)

Otherwise, the 'm == MATCH_YES' is using an old value.

>     {
>       gfc_current_block ()->declared_at = gfc_current_locus;
>       gfc_commit_symbols ();
>     }


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37319

[Bug fortran/37319] [4.4 regression] gfortran.dg/function_kinds_5.f90 fails

[ebotcazou at gcc dot gnu dot org]

------- Comment #10 from ebotcazou at gcc dot gnu dot org  2008-11-24 20:33 -------
> This should probably be
> 
>     m = gfc_match ("function% %n", name);
> 
> >   if (m == MATCH_YES && strcmp (name, gfc_current_block ()->name) == 0)
> 
> Otherwise, the 'm == MATCH_YES' is using an old value.

But in this case the first assignment to 'm' would be dead.  What about

  if (m == MATCH_YES
      && gfc_match ("function% %n", name) == MATCH_YES
      && strcmp (name, gfc_current_block ()->name) == 0)
    {
      gfc_current_block ()->declared_at = gfc_current_locus;
      gfc_commit_symbols ();
    }
  else
    gfc_error_check ();

instead?


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37319

[Bug fortran/37319] [4.4 regression] gfortran.dg/function_kinds_5.f90 fails

[kargl at gcc dot gnu dot org]

------- Comment #11 from kargl at gcc dot gnu dot org  2008-11-24 20:46 -------
(In reply to comment #10)
> > This should probably be
> > 
> >     m = gfc_match ("function% %n", name);
> > 
> > >   if (m == MATCH_YES && strcmp (name, gfc_current_block ()->name) == 0)
> > 
> > Otherwise, the 'm == MATCH_YES' is using an old value.
> 
> But in this case the first assignment to 'm' would be dead.  What about
> 
>   if (m == MATCH_YES
>       && gfc_match ("function% %n", name) == MATCH_YES
>       && strcmp (name, gfc_current_block ()->name) == 0)
>     {
>       gfc_current_block ()->declared_at = gfc_current_locus;
>       gfc_commit_symbols ();
>     }
>   else
>     gfc_error_check ();
> 
> instead?

OK, I've found the code location in parse.c.  Yes, I think the
above is probably correct.  Note, I'm no longer a gfortran 
maintainer, so I can't approve the patch and I haven't tested
it.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37319

[Bug fortran/37319] [4.4 regression] gfortran.dg/function_kinds_5.f90 fails

[ebotcazou at gcc dot gnu dot org]

------- Comment #12 from ebotcazou at gcc dot gnu dot org  2008-11-24 20:55 -------
> OK, I've found the code location in parse.c.  Yes, I think the
> above is probably correct.  Note, I'm no longer a gfortran 
> maintainer, so I can't approve the patch and I haven't tested
> it.

Understood, I'm going to test and submit it.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37319

[Bug fortran/37319] [4.4 regression] gfortran.dg/function_kinds_5.f90 fails

[jvdelisle at gcc dot gnu dot org]

------- Comment #13 from jvdelisle at gcc dot gnu dot org  2008-11-25 02:35 -------
I will regression test on x86-64-gnu-linux and approve if it passes.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37319

[Bug fortran/37319] [4.4 regression] gfortran.dg/function_kinds_5.f90 fails

[ebotcazou at gcc dot gnu dot org]

------- Comment #14 from ebotcazou at gcc dot gnu dot org  2008-11-25 08:41 -------
Subject: Bug 37319

Author: ebotcazou
Date: Tue Nov 25 08:39:39 2008
New Revision: 142188

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=142188
Log:
        PR fortran/37319
        * parse.c (match_deferred_characteristics): Make sure 'name' is
        initialized before reading it.

Modified:
    trunk/gcc/fortran/ChangeLog
    trunk/gcc/fortran/parse.c


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37319

[Bug fortran/37319] [4.4 regression] gfortran.dg/function_kinds_5.f90 fails

[ebotcazou at gcc dot gnu dot org]

------- Comment #15 from ebotcazou at gcc dot gnu dot org  2008-11-25 08:44 -------
Patch installed.


-- 

ebotcazou at gcc dot gnu dot org changed:

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


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37319