[Bug fortran/45151] New: [4.6 regression] New Fortran failuires

[Bug fortran/45151] New: [4.6 regression] New Fortran failuires

[hjl dot tools at gmail dot com]

On Linux/ia32, revision 162776:

http://gcc.gnu.org/ml/gcc-cvs/2010-07/msg01130.html

caused:

FAIL: gfortran.dg/abstract_type_6.f03  -O  (internal compiler error)
FAIL: gfortran.dg/abstract_type_6.f03  -O  (test for excess errors)
FAIL: gfortran.dg/bind_c_dts_3.f03  -O  (internal compiler error)
FAIL: gfortran.dg/bind_c_dts_3.f03  -O  (test for excess errors)
FAIL: gfortran.dg/c_assoc_2.f03  -O  (internal compiler error)
FAIL: gfortran.dg/c_assoc_2.f03  -O  (test for excess errors)
FAIL: gfortran.dg/c_ptr_tests_5.f03  -O  (internal compiler error)
FAIL: gfortran.dg/c_ptr_tests_5.f03  -O  (test for excess errors)
FAIL: gfortran.dg/finalize_3.f03  -O  (internal compiler error)
FAIL: gfortran.dg/finalize_3.f03  -O  (test for excess errors)
FAIL: gfortran.dg/function_kinds_5.f90  -O  (internal compiler error)
FAIL: gfortran.dg/function_kinds_5.f90  -O  (test for excess errors)
FAIL: gfortran.dg/proc_decl_3.f90  -O  (internal compiler error)
FAIL: gfortran.dg/proc_decl_3.f90  -O  (test for excess errors)
FAIL: gfortran.dg/proc_ptr_comp_pass_4.f90  -O  (internal compiler error)
FAIL: gfortran.dg/proc_ptr_comp_pass_4.f90  -O  (test for excess errors)
FAIL: gfortran.dg/proc_ptr_result_2.f90  -O  (internal compiler error)
FAIL: gfortran.dg/proc_ptr_result_2.f90  -O  (test for excess errors)
FAIL: gfortran.dg/typebound_call_15.f03  -O  (internal compiler error)
FAIL: gfortran.dg/typebound_call_15.f03  -O  (test for excess errors)
FAIL: gfortran.dg/typebound_call_4.f03  -O  (internal compiler error)
FAIL: gfortran.dg/typebound_call_4.f03  -O  (test for excess errors)
FAIL: gfortran.dg/typebound_operator_2.f03  -O  (internal compiler error)
FAIL: gfortran.dg/typebound_operator_2.f03  -O  (test for excess errors)
FAIL: gfortran.dg/typebound_proc_13.f03  -O  (internal compiler error)
FAIL: gfortran.dg/typebound_proc_13.f03  -O  (test for excess errors)
FAIL: gfortran.dg/typebound_proc_15.f03  -O  (internal compiler error)
FAIL: gfortran.dg/typebound_proc_15.f03  -O  (test for excess errors)
FAIL: gfortran.dg/typebound_proc_5.f03  -O   (test for errors, line 109)
FAIL: gfortran.dg/typebound_proc_5.f03  -O  (internal compiler error)
FAIL: gfortran.dg/typebound_proc_5.f03  -O  (test for excess errors)
FAIL: gfortran.dg/typebound_proc_6.f03  -O  (internal compiler error)
FAIL: gfortran.dg/typebound_proc_6.f03  -O  (test for excess errors)


-- 
           Summary: [4.6 regression] New Fortran failuires
           Product: gcc
           Version: 4.6.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: fortran
        AssignedTo: unassigned at gcc dot gnu dot org
        ReportedBy: hjl dot tools at gmail dot com


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=45151

[Bug fortran/45151] [4.6 regression] New Fortran failuires

[dominiq at lps dot ens dot fr]

------- Comment #1 from dominiq at lps dot ens dot fr  2010-07-31 15:37 -------
Likely due to revision 162776:

[macbook] f90/bug% gfc -std=f2003
/opt/gcc/work/gcc/testsuite/gfortran.dg/typebound_proc_15.f03
/opt/gcc/work/gcc/testsuite/gfortran.dg/typebound_proc_15.f03:15.23:

  procedure :: bar, baz  ! { dg-error "PROCEDURE list" }
                       1
Error: Fortran 2008: PROCEDURE list at (1)
f951: internal compiler error: in next_statement, at fortran/parse.c:894


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=45151

[Bug fortran/45151] [4.6 regression] New Fortran failuires

[mikael at gcc dot gnu dot org]

------- Comment #2 from mikael at gcc dot gnu dot org  2010-08-01 12:23 -------
Confirmed, and mine I guess.
I don't know what I was testing when I said I was testing the patch for the
future revision 162776.
Actually, I was a bit surprised not to see any failure in the testsuite.

Anyway, the patch is OK, I think (It can be reverted temporarily if needed). 
I will look at the failures. 


-- 

mikael at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
         AssignedTo|unassigned at gcc dot gnu   |mikael at gcc dot gnu dot
                   |dot org                     |org
             Status|UNCONFIRMED                 |ASSIGNED
     Ever Confirmed|0                           |1
   Last reconfirmed|0000-00-00 00:00:00         |2010-08-01 12:23:04
               date|                            |


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=45151

[Bug fortran/45151] [4.6 regression] New Fortran failuires

[rguenth at gcc dot gnu dot org]

-- 

rguenth at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|---                         |4.6.0


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=45151

[Bug fortran/45151] [4.6 regression] New Fortran failuires

[janus at gcc dot gnu dot org]

------- Comment #3 from janus at gcc dot gnu dot org  2010-08-01 12:39 -------
Many of these failures are actually due to 'gfc_copy_formal_args' and can be
fixed with:


Index: gcc/fortran/symbol.c
===================================================================
--- gcc/fortran/symbol.c        (revision 162794)
+++ gcc/fortran/symbol.c        (working copy)
@@ -4030,6 +4030,9 @@ gfc_copy_formal_args (gfc_symbol *dest, gfc_symbol
   gfc_formal_arglist *formal_prev = NULL;
   /* Save current namespace so we can change it for formal args.  */
   gfc_namespace *parent_ns = gfc_current_ns;
+  
+  if (!src->formal)
+    return;

   /* Create a new namespace, which will be the formal ns (namespace
      of the formal args).  */
@@ -4070,6 +4073,8 @@ gfc_copy_formal_args (gfc_symbol *dest, gfc_symbol
     dest->formal_ns = gfc_current_ns;
   /* Restore the current namespace to what it was on entry.  */
   gfc_current_ns = parent_ns;
+
+  gfc_commit_symbols ();
 }


@@ -4181,6 +4186,8 @@ gfc_copy_formal_args_ppc (gfc_component *dest, gfc
     dest->formal_ns = gfc_current_ns;
   /* Restore the current namespace to what it was on entry.  */
   gfc_current_ns = parent_ns;
+
+  gfc_commit_symbols ();
 }



With this patch, the remaining failures are:

FAIL: gfortran.dg/bind_c_dts_3.f03  -O  (internal compiler error)
FAIL: gfortran.dg/c_assoc_2.f03  -O  (internal compiler error)
FAIL: gfortran.dg/c_ptr_tests_5.f03  -O  (internal compiler error)
FAIL: gfortran.dg/finalize_3.f03  -O  (internal compiler error)
FAIL: gfortran.dg/function_kinds_5.f90  -O  (internal compiler error)
FAIL: gfortran.dg/proc_ptr_result_2.f90  -O  (internal compiler error)


-- 

janus at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |janus at gcc dot gnu dot org
   Target Milestone|4.6.0                       |---


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=45151

[Bug fortran/45151] [4.6 regression] New Fortran failuires

[janus at gcc dot gnu dot org]

------- Comment #4 from janus at gcc dot gnu dot org  2010-08-01 12:50 -------
Btw, maybe 'changed_syms' should be made static again. Instead of the assert in
parse.c one can use 'gfc_symbol_state'.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=45151

[Bug fortran/45151] [4.6 regression] New Fortran failuires

[mikael at gcc dot gnu dot org]

------- Comment #5 from mikael at gcc dot gnu dot org  2010-08-01 12:52 -------
(In reply to comment #3)
> Many of these failures are actually due to 'gfc_copy_formal_args' and can be
> fixed with:
[...]
> 
> 
Agreed. The patch I was about to post had also this extra hunk :

Index: symbol.c
===================================================================
--- symbol.c    (revision 162798)
+++ symbol.c    (working copy)
@@ -4141,6 +4141,10 @@ gfc_copy_formal_args_ppc (gfc_component *dest, gfc
   /* Save current namespace so we can change it for formal args.  */
   gfc_namespace *parent_ns = gfc_current_ns;

+  /* Make sure gfc_commit_symbols at the end doesn't commit something
+     unrelated.  */
+  gcc_assert (changed_syms == NULL);
+
   /* Create a new namespace, which will be the formal ns (namespace
      of the formal args).  */
   gfc_current_ns = gfc_get_namespace (parent_ns, 0);


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=45151

[Bug fortran/45151] [4.6 regression] New Fortran failuires

[mikael at gcc dot gnu dot org]

------- Comment #6 from mikael at gcc dot gnu dot org  2010-08-01 12:54 -------
(In reply to comment #4)
> Btw, maybe 'changed_syms' should be made static again. Instead of the assert in
> parse.c one can use 'gfc_symbol_state'.
> 
Yes, makes sense.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=45151

[Bug fortran/45151] [4.6 regression] New Fortran failuires

[mikael at gcc dot gnu dot org]

------- Comment #7 from mikael at gcc dot gnu dot org  2010-08-01 13:12 -------
(In reply to comment #3)
> With this patch, the remaining failures are:
> 
> FAIL: gfortran.dg/bind_c_dts_3.f03  -O  (internal compiler error)
> FAIL: gfortran.dg/c_assoc_2.f03  -O  (internal compiler error)

fixed by this :

Index: intrinsic.c
===================================================================
--- intrinsic.c (revision 162798)
+++ intrinsic.c (working copy)
@@ -112,6 +112,8 @@ gfc_get_intrinsic_sub_symbol (const char *name)
   sym->attr.flavor = FL_PROCEDURE;
   sym->attr.proc = PROC_INTRINSIC;

+  gfc_commit_symbols ();
+
   return sym;
 }


Remaining :
> FAIL: gfortran.dg/c_ptr_tests_5.f03  -O  (internal compiler error)
> FAIL: gfortran.dg/finalize_3.f03  -O  (internal compiler error)
> FAIL: gfortran.dg/function_kinds_5.f90  -O  (internal compiler error)
> FAIL: gfortran.dg/proc_ptr_result_2.f90  -O  (internal compiler error)
> 


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=45151

[Bug fortran/45151] [4.6 regression] New Fortran failuires

[janus at gcc dot gnu dot org]

------- Comment #8 from janus at gcc dot gnu dot org  2010-08-01 21:30 -------
> > FAIL: gfortran.dg/proc_ptr_result_2.f90  -O  (internal compiler error)

This one is fixed by adding a 'gfc_commit_symbols' also in
'gfc_copy_formal_args_intr' (cf. comment #3).


Still Remaining :
> > FAIL: gfortran.dg/c_ptr_tests_5.f03  -O  (internal compiler error)
> > FAIL: gfortran.dg/finalize_3.f03  -O  (internal compiler error)
> > FAIL: gfortran.dg/function_kinds_5.f90  -O  (internal compiler error)


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=45151

[Bug fortran/45151] [4.6 regression] New Fortran failuires

[mikael at gcc dot gnu dot org]

------- Comment #9 from mikael at gcc dot gnu dot org  2010-08-01 22:29 -------
Created an attachment (id=21367)
 --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=21367&action=view)
Tentative patch

I will test & submit this tomorrow. 

I have decided to use gfc_commit_symbol (sym) instead of the full
gfc_commit_symbols to minimize interference with changed_sym in both parsing
mode and resolution mode (just in case someone wants to use the functions in
the other mode in the future). 

The remaining testcases were missing a reject_statement () or gfc_undo_symbol
thrown together with the error. The patch also corrects others cases in the
affected functions that may prove wrong in the future.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=45151

[Bug fortran/45151] [4.6 regression] New Fortran failuires

[dominiq at lps dot ens dot fr]

------- Comment #10 from dominiq at lps dot ens dot fr  2010-08-02 05:23 -------
The patch in comment #9 works as expected on the gfortran testsuite and on my
own tests. Note that while testing I have found unrelated pr45161. Thanks for
the patch.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=45151

[Bug fortran/45151] [4.6 regression] New Fortran failuires

[mikael at gcc dot gnu dot org]

------- Comment #11 from mikael at gcc dot gnu dot org  2010-08-02 15:31 -------
Subject: Bug 45151

Author: mikael
Date: Mon Aug  2 15:30:47 2010
New Revision: 162821

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=162821
Log:
2010-08-02  Mikael Morin  <mikael@gcc.gnu.org>
            Janus Weil  <janus@gcc.gnu.org>

        PR fortran/42051
        PR fortran/44064
        PR fortran/45151
        * intrinsic.c (gfc_get_intrinsic_sub_symbol): Commit changed symbol. 
        * symbol.c (gen_cptr_param, gen_fptr_param, gen_shape_param,
        gfc_copy_formal_args, gfc_copy_formal_args_intr,
        gfc_copy_formal_args_ppc, generate_isocbinding_symbol): Ditto.
        * parse.c (parse_derived_contains, parse_spec, parse_progunit): 
        Call reject_statement in case of error. 
        (match_deferred_characteritics): Call gfc_undo_symbols in case match
        fails.


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


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=45151

[Bug fortran/45151] [4.6 regression] New Fortran failuires

[burnus at gcc dot gnu dot org]

------- Comment #12 from burnus at gcc dot gnu dot org  2010-08-03 09:44 -------
This PR seems to be mostly fixed:

- x86_64-unknown-linux-gnu shows _no_ failures (-m32/-m64)
  http://gcc.gnu.org/ml/gcc-testresults/2010-08/msg00217.html

- ia64-unknown-linux-gnu shows _two_ failures:
  http://gcc.gnu.org/ml/gcc-testresults/2010-08/msg00213.html

FAIL: gfortran.dg/maxlocval_3.f90  -O3 -fomit-frame-pointer -funroll-loops 
(test for excess errors)
FAIL: gfortran.dg/maxlocval_3.f90  -O3 -fomit-frame-pointer -funroll-all-loops
-finline-functions  (test for excess errors)
FAIL: gfortran.dg/typebound_proc_15.f03  -O  (internal compiler error)
FAIL: gfortran.dg/typebound_proc_15.f03  -O  (test for excess errors)

The latter is PR 44584 for which a patch has just been posted.

HJ: Can you confirm that all failures are now gone (pending the PR 44584
committal), except for maxlocval_3.f90?

Can you post the excess error which is shown for maxlocval_3.f90?


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=45151

[Bug fortran/45151] [4.6 regression] New Fortran failuires

[hjl dot tools at gmail dot com]

------- Comment #13 from hjl dot tools at gmail dot com  2010-08-03 14:24 -------
gfortran.dg/maxlocval_3.f90 is due to assembler warning:

/tmp/cc9gn3uW.s:3475: Warning: Use of 'movl' may violate WAW dependency 'GR%, %
in 1 - 127' (impliedf), specific resource number is 24^M
/tmp/cc9gn3uW.s:3475: Warning: Only the first path encountering the conflict is
reported^M
/tmp/cc9gn3uW.s:3472: Warning: This is the location of the conflicting usage^M
output is:
/tmp/cc9gn3uW.s: Assembler messages:^M
/tmp/cc9gn3uW.s:3475: Warning: Use of 'movl' may violate WAW dependency 'GR%, %
in 1 - 127' (impliedf), specific resource number is 24^M
/tmp/cc9gn3uW.s:3475: Warning: Only the first path encountering the conflict is
reported^M 
/tmp/cc9gn3uW.s:3472: Warning: This is the location of the conflicting usage^M

FAIL: gfortran.dg/maxlocval_3.f90  -O3 -fomit-frame-pointer -funroll-loops 
(test for excess errors)


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=45151

[Bug fortran/45151] [4.6 regression] New Fortran failuires

[sje at cup dot hp dot com]

------- Comment #14 from sje at cup dot hp dot com  2010-08-03 15:42 -------
The assembler warning messages (shown in comment #13) that are causing the
failure of gfortran.dg/maxlocval_3.f90 are due to PR 15445 and is not a new
problem.


-- 

sje at cup dot hp dot com changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |sje at cup dot hp dot com


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=45151

[Bug fortran/45151] [4.6 regression] New Fortran failuires

[burnus at gcc dot gnu dot org]

------- Comment #15 from burnus at gcc dot gnu dot org  2010-08-03 16:13 -------
Based on the comments: Close as FIXED.

Thanks for the initial report HJ; thanks for the fix Mikael; and thanks for the
pointer to PR target/15445 Steve!


-- 

burnus at gcc dot gnu dot org changed:

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


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=45151

[Bug fortran/45151] [4.6 regression] New Fortran failuires

[mikael at gcc dot gnu dot org]

------- Comment #16 from mikael at gcc dot gnu dot org  2010-08-05 21:08 -------
Subject: Bug 45151

Author: mikael
Date: Thu Aug  5 21:08:36 2010
New Revision: 162921

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=162921
Log:
2010-08-05  Mikael Morin  <mikael@gcc.gnu.org>
            Janus Weil  <janus@gcc.gnu.org>

        PR fortran/42051
        PR fortran/44064
        PR fortran/45151
        * intrinsic.c (gfc_get_intrinsic_sub_symbol): Commit changed symbol. 
        * symbol.c (gen_cptr_param, gen_fptr_param, gen_shape_param,
        gfc_copy_formal_args, gfc_copy_formal_args_intr,
        gfc_copy_formal_args_ppc, generate_isocbinding_symbol): Ditto.
        (gfc_find_derived_vtab): Commit newly created symbols.
        * parse.c (parse_derived_contains, parse_spec, parse_progunit): 
        Call reject_statement in case of error. 
        (match_deferred_characteritics): Call gfc_undo_symbols in case match
        fails.


Modified:
    branches/gcc-4_5-branch/gcc/fortran/ChangeLog
    branches/gcc-4_5-branch/gcc/fortran/intrinsic.c
    branches/gcc-4_5-branch/gcc/fortran/parse.c
    branches/gcc-4_5-branch/gcc/fortran/symbol.c


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=45151