[Bug fortran/26064] New: 4.1 regression introduced with patch for bug 18197

[Bug fortran/26064] New: 4.1 regression introduced with patch for bug 18197

[sven dot buijssen at math dot uni-dortmund dot de]

Erik Edelmann's patch for PR fortran/18197 introduced a regression. 

a) The following test case does not compile with gfortran >= Revision 108555

--- cut here ---
% cat > foo.h <<EOF
interface
  function foo() result (dresult)
    real :: dresult
  end function foo
end interface
EOF

% cat > regression.f90 <<EOF
module regression

contains

  subroutine helper(foo)
    include "foo.h"
  end subroutine helper

  subroutine master(foo)
    include "foo.h"
    call slave(foo, helper)
  end subroutine master

  subroutine slave(foo, subslave)
    include "foo.h"
    interface
      subroutine subslave(foo)
        include "foo.h"
      end subroutine subslave
    end interface
    call subslave(foo)
  end subroutine slave

end module regression
EOF
--- cut here ---

instead gfortran gives:

% gfortran -v -c regression.f90
Using built-in specs.
Target: i686-pc-linux-gnu
Configured with: ../../src/gcc/configure --enable-shared
--prefix=/home/[...]/20060201 --enable-languages=c,c++,fortran --enable-threads
--enable-__cxa_atexit
Thread model: posix
gcc version 4.2.0 20060201 (experimental)
 /home/[...]/20060201/bin/../libexec/gcc/i686-pc-linux-gnu/4.2.0/f951
regression.f90 -quiet -dumpbase regression.f90 -mtune=generic -auxbase
regression -version -o /tmp/ccNFDNgH.s
GNU F95 version 4.2.0 20060201 (experimental) (i686-pc-linux-gnu)
        compiled by GNU C version 4.2.0 20060201 (experimental).
GGC heuristics: --param ggc-min-expand=30 --param ggc-min-heapsize=4096

 In file regression.f90:11

    call slave(foo, helper)
                  1
Error: Type/rank mismatch in argument 'subslave' at (1)

b) Swapping the order of subroutines helper and master in above testcase
resolves the regression, gfortran compiles smoothly. (The complexity of the
code I actually tried to compile is higher and I did not find a suitable
permutation of subroutines yet.)

c) Compiling the current revision (110474) having reverted only Erik's patch:

--- cut here ---
% cat > undopatch18197 <<EOF
140a141,151
>           else
>             {
>                /* Set the type of the RESULT, then copy.  */
>               if (sym->result->ts.type == BT_UNKNOWN)
>                 gfc_set_default_type (sym->result, 1, sym->result->ns);
>
>               sym->ts = sym->result->ts;
>               if (sym->as == NULL)
>                 sym->as = gfc_copy_array_spec (sym->result->as);
>             }
>
EOF

% patch gcc/fortran/resolve.c < undopatch18197
% ./configure [....]; make bootstrap install
--- cut here ---

the above testcase compiles smoothly again - while reopening the error reported
in bug 18197 (and causing another error in my already mentioned code I could
not reduce to a small testcase so far.)


-- 
           Summary: 4.1 regression introduced with patch for bug 18197
           Product: gcc
           Version: 4.1.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: fortran
        AssignedTo: unassigned at gcc dot gnu dot org
        ReportedBy: sven dot buijssen at math dot uni-dortmund dot de
 GCC build triplet: i686-pc-linux-gnu
  GCC host triplet: i686-pc-linux-gnu
GCC target triplet: i686-pc-linux-gnu


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=26064

[Bug fortran/26064] 4.1 regression introduced with patch for bug 18197

[pinskia at gcc dot gnu dot org]

------- Comment #1 from pinskia at gcc dot gnu dot org  2006-02-01 14:45 -------
This looks related to PR 26041.
Can you try the patch at
http://gcc.gnu.org/ml/gcc-patches/2006-01/msg02261.html
?


-- 

pinskia at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
  BugsThisDependsOn|                            |26041


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=26064

[Bug fortran/26064] 4.1 regression introduced with patch for bug 18197

[sven dot buijssen at math dot uni-dortmund dot de]

------- Comment #2 from sven dot buijssen at math dot uni-dortmund dot de  2006-02-01 16:36 -------
The patch fixes the regression, but introduces a new internal compiler error
for the following testcase:

--- cut here ---
% cat > ice.f90 <<EOF
module ice
  implicit none
  contains

    subroutine foo()
    contains

      subroutine bar(baz)
        integer :: baz
        if (present(baz)) then
        endif
      end subroutine bar
    end subroutine foo
end module
EOF
--- cut here ---

% gfortran -c ice.f90
ice.f90: In function 'bar':
ice.f90:8: internal compiler error: in gfc_typenode_for_spec, at
fortran/trans-types.c:651
Please submit a full bug report,
with preprocessed source if appropriate.
See <URL:http://gcc.gnu.org/bugs.html> for instructions.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=26064

[Bug fortran/26064] 4.1 regression introduced with patch for bug 18197

[hjl at lucon dot org]

------- Comment #3 from hjl at lucon dot org  2006-02-01 16:49 -------
It is the same bug.

*** This bug has been marked as a duplicate of 26041 ***


-- 

hjl at lucon dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |RESOLVED
         Resolution|                            |DUPLICATE


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=26064

[Bug fortran/26064] 4.1 regression introduced with patch for bug 18197

[hjl at lucon dot org]

------- Comment #4 from hjl at lucon dot org  2006-02-01 17:00 -------
I didn't see it:

[hjl@gnu-16]$  cat xxx.f90
module ice
  implicit none
  contains

    subroutine foo()
    contains

      subroutine bar(baz)
        integer :: baz
        if (present(baz)) then
        endif
      end subroutine bar
    end subroutine foo
end module
[hjl@gnu-16]$ /export/build/gnu/gcc/build-x86_64-linux/gcc/gfortran
-B/export/build/gnu/gcc/build-x86_64-linux/gcc/  -S xxx.f90
 In file xxx.f90:10

        if (present(baz)) then
                   1
Error: 'a' argument of 'present' intrinsic at (1) must be of an OPTIONAL dummy
variable
[hjl@gnu-16]$


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=26064

[Bug fortran/26064] 4.1 regression introduced with patch for bug 18197

[sven dot buijssen at math dot uni-dortmund dot de]

------- Comment #5 from sven dot buijssen at math dot uni-dortmund dot de  2006-02-01 20:01 -------
I'm sorry. I had been experimenting which circumstances trigger the ice and
accidentally deleted the "optional" argument. To be valid the code needs to be
like

% cat > ice.f90 <<EOF
module ice
  implicit none
  contains

    subroutine foo()
    contains

      subroutine bar(baz)
        integer, optional :: baz
        if (present(baz)) then
        endif
      end subroutine bar
    end subroutine foo
end module
EOF

Just to be clear: gfortran revision 110475 does not complain about this
testcase. Only when I followed the advise of Andrew and integrated the patch
you suggested yesterday, H.J., then gfortran gives an ice.

% svn update -r HEAD
At revision 110475.
% svn diff
Index: gcc/fortran/resolve.c
===================================================================
--- gcc/fortran/resolve.c       (revision 110475)
+++ gcc/fortran/resolve.c       (working copy)
@@ -5771,21 +5771,14 @@
 }


-/* This function is called after a complete program unit has been compiled.
-   Its purpose is to examine all of the expressions associated with a program
-   unit, assign types to all intermediate expressions, make sure that all
-   assignments are to compatible types and figure out which names refer to
-   which functions or subroutines.  */
-
-void
-gfc_resolve (gfc_namespace * ns)
+static void
+gfc_resolve_type (gfc_namespace * ns)
 {
-  gfc_namespace *old_ns, *n;
+  gfc_namespace *n;
   gfc_charlen *cl;
   gfc_data *d;
   gfc_equiv *eq;

-  old_ns = gfc_current_ns;
   gfc_current_ns = ns;

   resolve_entries (ns);
@@ -5803,7 +5796,7 @@
                   "also be PURE", n->proc_name->name,
                   &n->proc_name->declared_at);

-      gfc_resolve (n);
+      gfc_resolve_type (n);
     }

   forall_flag = 0;
@@ -5827,12 +5820,37 @@
   for (eq = ns->equiv; eq; eq = eq->next)
     resolve_equivalence (eq);

-  cs_base = NULL;
-  resolve_code (ns->code, ns);
-
   /* Warn about unused labels.  */
   if (gfc_option.warn_unused_labels)
     warn_unused_label (ns->st_labels);
+}

+
+/* This function is called after a complete program unit has been compiled.
+   Its purpose is to examine all of the expressions associated with a program
+   unit, assign types to all intermediate expressions, make sure that all
+   assignments are to compatible types and figure out which names refer to
+   which functions or subroutines.  */
+
+void
+gfc_resolve (gfc_namespace * ns)
+{
+  gfc_namespace *old_ns, *n;
+
+  old_ns = gfc_current_ns;
+
+  gfc_resolve_type (ns);
+
+  for (n = ns->contained; n; n = n->sibling)
+    {
+      gfc_current_ns = n;
+      cs_base = NULL;
+      resolve_code (n->code, n);
+    }
+
+  gfc_current_ns = ns;
+  cs_base = NULL;
+  resolve_code (ns->code, ns);
+
   gfc_current_ns = old_ns;
 }


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=26064

[Bug fortran/26064] 4.1 regression introduced with patch for bug 18197

[hjl at lucon dot org]

------- Comment #6 from hjl at lucon dot org  2006-02-01 21:20 -------
I got

Breakpoint 3, gfc_sym_type (sym=0xb5b800)
    at /net/gnu-13/export/gnu/src/gcc-last/gcc/gcc/fortran/trans-types.c:1301
1301      type = gfc_typenode_for_spec (&sym->ts);
$8 = {name = 0xb51b46 "present", module = 0x0, declared_at = {
    nextc = 0xb53bbf "(baz)) then", lb = 0xb53b90}, ts = {type = BT_UNKNOWN,
    kind = 0, derived = 0x0, cl = 0x0}, attr = {allocatable = 0,
    dimension = 0, external = 0, intrinsic = 0, optional = 0, pointer = 0,
    save = 0, target = 0, dummy = 0, result = 0, assign = 0, data = 0,
    use_assoc = 0, in_namelist = 0, in_common = 0, in_equivalence = 0,
    function = 1, subroutine = 0, generic = 0, implicit_type = 0, untyped = 0,
    sequence = 0, elemental = 0, pure = 0, recursive = 0, unmaskable = 0,
    masked = 0, contained = 0, noreturn = 0, entry = 0, entry_master = 0,
    mixed_entry_master = 0, always_explicit = 0, referenced = 1,
    is_main_program = 0, access = ACCESS_UNKNOWN, intent = INTENT_UNKNOWN,
    flavor = FL_PROCEDURE, if_source = IFSRC_UNKNOWN, proc = PROC_UNKNOWN,
    cray_pointer = 0, cray_pointee = 0}, generic = 0x0,
  component_access = ACCESS_UNKNOWN, formal = 0x0, formal_ns = 0x0,
  value = 0x0, as = 0x0, result = 0xb5b800, components = 0x0,
  cp_pointer = 0x0, common_next = 0x0, common_head = 0x0, dummy_order = 0,
  namelist = 0x0, namelist_tail = 0x0, old_symbol = 0x0, tlink = 0x0,
  mark = 1, new = 0, equiv_built = 0, refs = 1, ns = 0xb5b170,
---Type <return> to continue, or q <return> to quit---
  backend_decl = 0x0}
(gdb)

Is "present" an intrinsic? Why isn't it marked so?


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=26064

[Bug fortran/26064] 4.1 regression introduced with patch for bug 18197

[hjl at lucon dot org]

------- Comment #7 from hjl at lucon dot org  2006-02-02 01:19 -------
Please try this one

http://gcc.gnu.org/ml/gcc-patches/2006-02/msg00102.html


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=26064

[Bug fortran/26064] 4.1 regression introduced with patch for bug 18197

[sven dot buijssen at math dot uni-dortmund dot de]

------- Comment #8 from sven dot buijssen at math dot uni-dortmund dot de  2006-02-02 11:30 -------
I tested the patch and encountered another internal compiler error. The
problem, however, is that this new error so far only occurs with a code
consisting of several thousand code lines and disappears as soon as I try to
remove some complexity. Nevertheless, I'm trying to deduce a reasonably sized
test case. I recon the error message I get does not help you to much in the
meantime:

internal compiler error: tree check: expected record_type or union_type or
qual_union_type, have omp_clause in gfc_get_derived_type, at fortran/trans-
types.c:1458


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=26064

[Bug fortran/26064] 4.1 regression introduced with patch for bug 18197

[hjl at lucon dot org]

------- Comment #9 from hjl at lucon dot org  2006-02-02 14:49 -------
If you back out the change for bug 18197, does several thousand lines of
code compile?


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=26064

[Bug fortran/26064] 4.1 regression introduced with patch for bug 18197

[sven dot buijssen at math dot uni-dortmund dot de]

------- Comment #10 from sven dot buijssen at math dot uni-dortmund dot de  2006-02-02 16:12 -------
They did with gfortran < Revision 108555. With current revision 110515 backing
out only the patch for 18197 (and not applying your patch) they do not.

The error message I get with revision 110515 backing out only the patch for
18197 is exactly the same as with revision 110515 integrating your patch. So,
it's probably a separate regression. I'll continue to distillate a test case
and file a separate bug report.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=26064

[Bug fortran/26064] 4.1 regression introduced with patch for bug 18197

[hjl at gcc dot gnu dot org]

------- Comment #11 from hjl at gcc dot gnu dot org  2006-02-05 19:52 -------
Subject: Bug 26064

Author: hjl
Date: Sun Feb  5 19:52:35 2006
New Revision: 110618

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=110618
Log:
2006-02-05  H.J. Lu  <hongjiu.lu@intel.com>

        PR fortran/26041
        PR fortran/26064
        * gfortran.dg/sibling_dummy_procedure_1.f90: New file.
        * gfortran.dg/sibling_dummy_procedure_2.f90: Likewise.
        * gfortran.dg/sibling_dummy_procedure_3.f90: Likewise.

Added:
    trunk/gcc/testsuite/gfortran.dg/sibling_dummy_procedure_1.f90
    trunk/gcc/testsuite/gfortran.dg/sibling_dummy_procedure_2.f90
    trunk/gcc/testsuite/gfortran.dg/sibling_dummy_procedure_3.f90
Modified:
    trunk/gcc/testsuite/ChangeLog


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=26064

[Bug fortran/26064] 4.1 regression introduced with patch for bug 18197

[hjl at gcc dot gnu dot org]

------- Comment #12 from hjl at gcc dot gnu dot org  2006-02-05 19:53 -------
Subject: Bug 26064

Author: hjl
Date: Sun Feb  5 19:53:00 2006
New Revision: 110619

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=110619
Log:
2006-02-05  H.J. Lu  <hongjiu.lu@intel.com>

        PR fortran/26041
        PR fortran/26064
        * resolve.c (resolve_types): New function.
        (resolve_codes): Likewise.
        (gfc_resolve): Use them.

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


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=26064

[Bug fortran/26064] 4.1 regression introduced with patch for bug 18197

[hjl at gcc dot gnu dot org]

------- Comment #13 from hjl at gcc dot gnu dot org  2006-03-07 22:29 -------
Subject: Bug 26064

Author: hjl
Date: Tue Mar  7 22:29:32 2006
New Revision: 111821

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=111821
Log:
2006-03-07  H.J. Lu  <hongjiu.lu@intel.com>

        PR fortran/26041
        PR fortran/26064
        Backport from mainline
        * resolve.c (resolve_types): New function.
        (resolve_codes): Likewise.
        (gfc_resolve): Use them.

Modified:
    branches/gcc-4_1-branch/gcc/fortran/ChangeLog
    branches/gcc-4_1-branch/gcc/fortran/resolve.c


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=26064

[Bug fortran/26064] 4.1 regression introduced with patch for bug 18197

[hjl at gcc dot gnu dot org]

------- Comment #14 from hjl at gcc dot gnu dot org  2006-03-07 22:31 -------
Subject: Bug 26064

Author: hjl
Date: Tue Mar  7 22:30:58 2006
New Revision: 111822

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=111822
Log:
2006-03-07  H.J. Lu  <hongjiu.lu@intel.com>

        PR fortran/26041
        PR fortran/26064
        Backport from mainline
        * gfortran.dg/sibling_dummy_procedure_1.f90: New file.
        * gfortran.dg/sibling_dummy_procedure_2.f90: Likewise.
        * gfortran.dg/sibling_dummy_procedure_3.f90: Likewise.

Added:
   
branches/gcc-4_1-branch/gcc/testsuite/gfortran.dg/sibling_dummy_procedure_1.f90
   
branches/gcc-4_1-branch/gcc/testsuite/gfortran.dg/sibling_dummy_procedure_2.f90
   
branches/gcc-4_1-branch/gcc/testsuite/gfortran.dg/sibling_dummy_procedure_3.f90
Modified:
    branches/gcc-4_1-branch/gcc/testsuite/ChangeLog


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=26064