[PATCH v2 1/7] DWARF: Don't add nameless modules to partial symbol table.

[PATCH v2 1/7] DWARF: Don't add nameless modules to partial symbol table.

[Pawel Wodkowski]

From: Bernhard Heckel <bernhard.heckel@intel.com>

A name for BLOCK DATA in Fortran is optional. If no
name has been assigned, GDB will crash during read-in of DWARF
when BLOCK DATA is represented via DW_TAG_module.
BLOCK DATA is used for one-time initialization of
non-pointer variables in named common blocks.

2016-06-15  Bernhard Heckel  <bernhard.heckel@intel.com>

gdb/Changelog:
	* dwarf2read.c (add_partial_symbol): Skip nameless modules.

gdb/Testsuite/Changelog:
	* gdb.fortran/block-data.f: New.
	* gdb.fortran/block-data.exp: New.
---
 gdb/dwarf2read.c                         | 13 ++++---
 gdb/testsuite/gdb.fortran/block-data.exp | 51 ++++++++++++++++++++++++++++
 gdb/testsuite/gdb.fortran/block-data.f   | 58 ++++++++++++++++++++++++++++++++
 3 files changed, 117 insertions(+), 5 deletions(-)
 create mode 100644 gdb/testsuite/gdb.fortran/block-data.exp
 create mode 100644 gdb/testsuite/gdb.fortran/block-data.f

diff --git a/gdb/dwarf2read.c b/gdb/dwarf2read.c
index d2a8cd44f9a5..89fd4ae15e80 100644
--- a/gdb/dwarf2read.c
+++ b/gdb/dwarf2read.c
@@ -8995,11 +8995,14 @@ add_partial_symbol (struct partial_die_info *pdi, struct dwarf2_cu *cu)
 			   0, cu->language, objfile);
       break;
     case DW_TAG_module:
-      add_psymbol_to_list (actual_name, strlen (actual_name),
-			   built_actual_name != NULL,
-			   MODULE_DOMAIN, LOC_TYPEDEF, -1,
-			   &objfile->global_psymbols,
-			   0, cu->language, objfile);
+      /* In Fortran 77 there might be a "BLOCK DATA" module available wihout
+         any name. If so, we skip the module as it doesn't bring any value */
+      if (actual_name != nullptr)
+	add_psymbol_to_list (actual_name, strlen (actual_name),
+			     built_actual_name != nullptr,
+			     MODULE_DOMAIN, LOC_TYPEDEF, -1,
+			     &objfile->global_psymbols,
+			     0, cu->language, objfile);
       break;
     case DW_TAG_class_type:
     case DW_TAG_interface_type:
diff --git a/gdb/testsuite/gdb.fortran/block-data.exp b/gdb/testsuite/gdb.fortran/block-data.exp
new file mode 100644
index 000000000000..2af250ad3886
--- /dev/null
+++ b/gdb/testsuite/gdb.fortran/block-data.exp
@@ -0,0 +1,51 @@
+# Copyright 2018 Free Software Foundation, Inc.
+#
+# Contributed by Intel Corp. <bernhard.heckel@intel.com>
+#
+# This program is free software; you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation; either version 3 of the License, or
+# (at your option) any later version.
+#
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# along with this program.  If not, see <http://www.gnu.org/licenses/>.
+
+# This test is supposed to test anonymous block-data statement.
+
+if { [skip_fortran_tests] } { return -1 }
+
+standard_testfile .f
+load_lib "fortran.exp"
+
+if {[prepare_for_testing $testfile.exp $testfile $srcfile {debug f90}]} {
+    return -1
+}
+
+if ![runto MAIN__] then {
+    untested "couldn't run to breakpoint MAIN__"
+    return -1
+}
+
+gdb_test "print doub1" "= 1.11\\d+" "print doub1, default values"
+gdb_test "print doub2" "= 2.22\\d+" "print doub2, default values"
+gdb_test "print char1" "= 'abcdef'" "print char1, default values"
+gdb_test "print char2" "= 'ghijkl'" "print char2, default values"
+
+gdb_breakpoint [gdb_get_line_number "! BP_BEFORE_SUB"]
+gdb_continue_to_breakpoint "! BP_BEFORE_SUB" ".*! BP_BEFORE_SUB.*"
+gdb_test "print doub1" "= 11.11\\d+" "print doub1, before sub"
+gdb_test "print doub2" "= 22.22\\d+" "print doub2, before sub"
+gdb_test "print char1" "= 'ABCDEF'" "print char1, before sub"
+gdb_test "print char2" "= 'GHIJKL'" "print char2, before sub"
+
+gdb_breakpoint [gdb_get_line_number "! BP_SUB"]
+gdb_continue_to_breakpoint "! BP_SUB" ".*! BP_SUB.*"
+gdb_test "print doub1" "= 11.11\\d+" "print char1, in sub"
+gdb_test "print doub2" "= 22.22\\d+" "print doub2, in sub"
+gdb_test "print char1" "= 'ABCDEF'" "print char1, in sub"
+gdb_test "print char2" "= 'GHIJKL'" "print char2, in sub"
diff --git a/gdb/testsuite/gdb.fortran/block-data.f b/gdb/testsuite/gdb.fortran/block-data.f
new file mode 100644
index 000000000000..a28e687ec885
--- /dev/null
+++ b/gdb/testsuite/gdb.fortran/block-data.f
@@ -0,0 +1,58 @@
+! Copyright 2018 Free Software Foundation, Inc.
+!
+! Contributed by Intel Corp. <bernhard.heckel@intel.com>
+!
+! This program is free software; you can redistribute it and/or modify
+! it under the terms of the GNU General Public License as published by
+! the Free Software Foundation; either version 3 of the License, or
+! (at your option) any later version.
+!
+! This program is distributed in the hope that it will be useful,
+! but WITHOUT ANY WARRANTY; without even the implied warranty of
+! MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+! GNU General Public License for more details.
+!
+! You should have received a copy of the GNU General Public License
+! along with this program.  If not, see <http://www.gnu.org/licenses/>.
+!
+! Test if GDB can handle block data without global name
+!
+! MAIN
+        PROGRAM bdata
+        DOUBLE PRECISION doub1, doub2
+        CHARACTER*6 char1, char2
+
+        COMMON /BLK1/ doub1, char1
+        COMMON /BLK2/ doub2, char2
+
+        doub1 = 11.111
+        doub2 = 22.222
+        char1 = 'ABCDEF'
+        char2 = 'GHIJKL'
+        CALL sub_block_data      ! BP_BEFORE_SUB
+        STOP
+        END
+
+! BLOCK DATA
+        BLOCK DATA
+
+        DOUBLE PRECISION doub1, doub2
+        CHARACTER*6 char1, char2
+
+        COMMON /BLK1/ doub1, char1
+        COMMON /BLK2/ doub2, char2
+        DATA doub1, doub2 /1.111, 2.222/
+        DATA char1, char2 /'abcdef', 'ghijkl'/
+        END
+
+! SUBROUTINE
+        SUBROUTINE sub_block_data
+
+        DOUBLE PRECISION doub1, doub2
+        CHARACTER*6 char1, char2
+
+        COMMON /BLK1/ doub1, char1
+        COMMON /BLK2/ doub2, char2
+
+        char1 = char2;    ! BP_SUB
+        END
-- 
2.7.4

[PATCH v2 6/7] Fortran: Nested functions, add scope parameter.

[Pawel Wodkowski]

From: Bernhard Heckel <bernhard.heckel@intel.com>

In order to avoid name clashing in GDB, we add a scope
to nested subroutines. Enveloping function gives the
scope.
---
 gdb/dwarf2read.c                           | 24 +++++++++++-
 gdb/testsuite/gdb.fortran/nested-funcs.exp | 20 ++++++++--
 gdb/testsuite/gdb.fortran/nested-funcs.f90 | 63 ++++++++++++++++++++++++++++--
 3 files changed, 98 insertions(+), 9 deletions(-)

diff --git a/gdb/dwarf2read.c b/gdb/dwarf2read.c
index dcf9da8372df..78f96ea0d157 100644
--- a/gdb/dwarf2read.c
+++ b/gdb/dwarf2read.c
@@ -8785,6 +8785,7 @@ partial_die_parent_scope (struct partial_die_info *pdi,
       return NULL;
     }
 
+  /* Internal (nested) subroutines in Fortran get a prefix.  */
   if (pdi->tag == DW_TAG_enumerator)
     /* Enumerators should not get the name of the enumeration as a prefix.  */
     parent->scope = grandparent_scope;
@@ -8794,7 +8795,10 @@ partial_die_parent_scope (struct partial_die_info *pdi,
       || parent->tag == DW_TAG_class_type
       || parent->tag == DW_TAG_interface_type
       || parent->tag == DW_TAG_union_type
-      || parent->tag == DW_TAG_enumeration_type)
+      || parent->tag == DW_TAG_enumeration_type
+      || (cu->language == language_fortran
+	  && parent->tag == DW_TAG_subprogram
+	  && pdi->tag == DW_TAG_subprogram))
     {
       if (grandparent_scope == NULL)
 	parent->scope = parent->name;
@@ -10651,8 +10655,14 @@ process_die (struct die_info *die, struct dwarf2_cu *cu)
     case DW_TAG_type_unit:
       read_type_unit_scope (die, cu);
       break;
-    case DW_TAG_entry_point:
     case DW_TAG_subprogram:
+      /* Internal subprograms in Fortran get a prefix.  */
+      if (cu->language == language_fortran
+	  && die->parent != NULL
+	  && die->parent->tag == DW_TAG_subprogram)
+	cu->processing_has_namespace_info = 1;
+      /* Fall through.  */
+    case DW_TAG_entry_point:
     case DW_TAG_inlined_subroutine:
       read_func_scope (die, cu);
       break;
@@ -22547,6 +22557,16 @@ determine_prefix (struct die_info *die, struct dwarf2_cu *cu)
 	      return name;
 	  }
 	return "";
+      case DW_TAG_subprogram:
+	/* Only internal subroutines in Fortran get a prefix with the name
+	   of the parent's subroutine.  */
+	if (cu->language == language_fortran)
+	  {
+	    if ((die->tag ==  DW_TAG_subprogram)
+		&& (dwarf2_name (parent, cu) != NULL))
+	      return dwarf2_name (parent, cu);
+	  }
+	return determine_prefix (parent, cu);
       case DW_TAG_enumeration_type:
 	parent_type = read_type_die (parent, cu);
 	if (TYPE_DECLARED_CLASS (parent_type))
diff --git a/gdb/testsuite/gdb.fortran/nested-funcs.exp b/gdb/testsuite/gdb.fortran/nested-funcs.exp
index 5e28b0e1a3e2..51c0053d575a 100755
--- a/gdb/testsuite/gdb.fortran/nested-funcs.exp
+++ b/gdb/testsuite/gdb.fortran/nested-funcs.exp
@@ -30,8 +30,8 @@ if ![runto MAIN__] then {
     continue
 }
 # Test if we can set a breakpoint in a nested function
-gdb_breakpoint "sub_nested_outer"
-gdb_continue_to_breakpoint "sub_nested_outer" ".*local_int = 19"
+gdb_breakpoint "testnestedfuncs::sub_nested_outer"
+gdb_continue_to_breakpoint "testnestedfuncs::sub_nested_outer" ".*local_int = 19"
 
 # Test if we can access local and
 # non-local variables defined one level up.
@@ -47,8 +47,8 @@ gdb_test "up"
 gdb_test "print index" "= 42" "print index at BP1, one frame up"
 
 # Test if we can set a breakpoint in a nested function
-gdb_breakpoint "sub_nested_inner"
-gdb_continue_to_breakpoint "sub_nested_inner" ".*local_int = 17"
+gdb_breakpoint "testnestedfuncs::sub_nested_inner"
+gdb_continue_to_breakpoint "testnestedfuncs::sub_nested_inner" ".*local_int = 17"
 
 # Test if we can access local and
 # non-local variables defined two level up.
@@ -64,6 +64,18 @@ gdb_continue_to_breakpoint "! BP_outer_2" ".*! BP_outer_2"
 gdb_test "print local_int" "= 19" \
   "print local_int in outer function, after sub_nested_inner"
 
+# Test if we can set a breakpoint in public routine with the same name as the internal
+gdb_breakpoint "sub_nested_outer"
+gdb_continue_to_breakpoint "sub_nested_outer" ".*name = 'sub_nested_outer external'"
+
+# Test if we can set a breakpoint in public routine with the same name as the internal
+gdb_breakpoint "sub_with_sub_nested_outer::sub_nested_outer"
+gdb_continue_to_breakpoint "sub_with_sub_nested_outer::sub_nested_outer" ".*local_int = 11"
+
+# Test if we can set a breakpoint in public routine with the same name as the internal
+gdb_breakpoint "mod1::sub_nested_outer"
+gdb_continue_to_breakpoint "mod1::sub_nested_outer" ".*name = 'sub_nested_outer_mod1'"
+
 # Sanity check in main.
 gdb_breakpoint [gdb_get_line_number "! BP_main"]
 gdb_continue_to_breakpoint "! BP_main" ".*! BP_main"
diff --git a/gdb/testsuite/gdb.fortran/nested-funcs.f90 b/gdb/testsuite/gdb.fortran/nested-funcs.f90
index 50a63b2c7c30..d4682b25465b 100755
--- a/gdb/testsuite/gdb.fortran/nested-funcs.f90
+++ b/gdb/testsuite/gdb.fortran/nested-funcs.f90
@@ -13,8 +13,62 @@
 ! You should have received a copy of the GNU General Public License
 ! along with this program.  If not, see <http://www.gnu.org/licenses/>.
 
-program TestNestedFuncs
+module mod1
+  integer :: var_i = 1
+  integer :: var_const
+  parameter (var_const = 20)
+
+CONTAINS
+
+  SUBROUTINE sub_nested_outer
+    integer :: local_int
+    character (len=20) :: name
+
+    name = 'sub_nested_outer_mod1'
+    local_int = 11
+
+  END SUBROUTINE sub_nested_outer
+end module mod1
+
+! Public sub_nested_outer
+SUBROUTINE sub_nested_outer
+  integer :: local_int
+  character (len=16) :: name
+
+  name = 'sub_nested_outer external'
+  local_int = 11
+END SUBROUTINE sub_nested_outer
+
+! Needed indirection to call public sub_nested_outer from main
+SUBROUTINE sub_nested_outer_ind
+  character (len=20) :: name
+
+  name = 'sub_nested_outer_ind'
+  CALL sub_nested_outer
+END SUBROUTINE sub_nested_outer_ind
+
+! public routine with internal subroutine
+SUBROUTINE sub_with_sub_nested_outer()
+  integer :: local_int
+  character (len=16) :: name
+
+  name = 'subroutine_with_int_sub'
+  local_int = 1
+
+  CALL sub_nested_outer  ! Should call the internal fct
+
+CONTAINS
 
+  SUBROUTINE sub_nested_outer
+    integer :: local_int
+    local_int = 11
+  END SUBROUTINE sub_nested_outer
+
+END SUBROUTINE sub_with_sub_nested_outer
+
+! Main
+program TestNestedFuncs
+  USE mod1, sub_nested_outer_use_mod1 => sub_nested_outer
   IMPLICIT NONE
 
   TYPE :: t_State
@@ -22,10 +76,13 @@ program TestNestedFuncs
   END TYPE t_State
 
   TYPE (t_State) :: v_state
-  integer index
+  integer index, local_int
 
   index = 13
-  CALL sub_nested_outer
+  CALL sub_nested_outer            ! Call internal sub_nested_outer
+  CALL sub_nested_outer_ind        ! Call external sub_nested_outer via sub_nested_outer_ind
+  CALL sub_with_sub_nested_outer   ! Call external routine with nested sub_nested_outer
+  CALL sub_nested_outer_use_mod1   ! Call sub_nested_outer imported via module
   index = 11              ! BP_main
   v_state%code = 27
 
-- 
2.7.4

[PATCH v2 3/7] Fortran: Accessing fields of inherited types via fully qualified name.

[Pawel Wodkowski]

From: Bernhard Heckel <bernhard.heckel@intel.com>

Fortran 2003 supports type extension. This patch allows access
to inherited members by using it's fully qualified name as
described in the fortran standard.

Before:
(gdb) print my_extended_obj%base_class_name%member_base
Syntax error near base_class_name%member_base

(gdb) print my_extended_obj%member_base
$1 = (10, 10, 10)

After:
(gdb) print my_extended_obj%base_clase_name%member_base
$1 = (10, 10, 10)

(gdb) print my_extended_obj%member_base
$1 = (10, 10, 10)

2016-04-22  Bernhard Heckel  <bernhard.heckel@intel.com>

gdb/Changelog:
	* f-exp.y (name): Allow TYPENAME.
	* valops.c (search_struct_method): Look also for baseclass.

gdb/Testsuite/Changelog:
	* gdb.fortran/oop-extend-type.f90: New.
	* gdb.fortran/oop-extend-type.exp: New.
---
 gdb/f-exp.y                                   |   7 +-
 gdb/testsuite/gdb.fortran/oop-extend-type.exp | 113 ++++++++++++++++++++++++++
 gdb/testsuite/gdb.fortran/oop-extend-type.f90 |  56 +++++++++++++
 gdb/valops.c                                  |   6 ++
 4 files changed, 180 insertions(+), 2 deletions(-)
 create mode 100644 gdb/testsuite/gdb.fortran/oop-extend-type.exp
 create mode 100644 gdb/testsuite/gdb.fortran/oop-extend-type.f90

diff --git a/gdb/f-exp.y b/gdb/f-exp.y
index 390bd45081b7..4c2e101699ac 100644
--- a/gdb/f-exp.y
+++ b/gdb/f-exp.y
@@ -604,8 +604,11 @@ nonempty_typelist
 		}
 	;
 
-name	:	NAME
-		{  $$ = $1.stoken; }
+name
+	:	NAME
+		{ $$ = $1.stoken; }
+	|	TYPENAME
+		{ $$ = $1.stoken; }
 	;
 
 name_not_typename :	NAME
diff --git a/gdb/testsuite/gdb.fortran/oop-extend-type.exp b/gdb/testsuite/gdb.fortran/oop-extend-type.exp
new file mode 100644
index 000000000000..8c3bb50a3ac6
--- /dev/null
+++ b/gdb/testsuite/gdb.fortran/oop-extend-type.exp
@@ -0,0 +1,113 @@
+# Copyright 2018 Free Software Foundation, Inc.
+
+# This program is free software; you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation; either version 3 of the License, or
+# (at your option) any later version.
+#
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# along with this program.  If not, see <http://www.gnu.org/licenses/>.
+
+standard_testfile ".f90"
+load_lib "fortran.exp"
+
+if { [prepare_for_testing ${testfile}.exp ${testfile} ${srcfile} \
+     {debug f90 quiet}] } {
+	 return -1
+}
+
+if ![runto_main] {
+    untested "could not run to main"
+    return -1
+}
+
+# Depending on the compiler being used, the type names can be printed differently.
+set real [fortran_real4]
+
+gdb_breakpoint [gdb_get_line_number "! Before vla allocation"]
+gdb_continue_to_breakpoint "! Before vla allocation" ".*! Before vla allocation"
+gdb_test "whatis wp_vla" "type = <not allocated>"
+
+gdb_breakpoint [gdb_get_line_number "! After value assignment"]
+gdb_continue_to_breakpoint "! After value assignment" ".*! After value assignment"
+set test "p wp%coo"
+gdb_test_multiple "$test" "$test" {
+    -re " = \\(1, 2, 1\\)\r\n$gdb_prompt $" {
+      pass "$test"
+    }
+    -re "There is no member named coo.\r\n$gdb_prompt $" {
+      kfail "gcc/49475" "$test"
+    }
+}
+gdb_test "p wp%point%coo" " = \\(1, 2, 1\\)"
+gdb_test "p wp%point" " = \\( coo = \\(1, 2, 1\\) \\)"
+gdb_test "p wp" " = \\( point = \\( coo = \\(1, 2, 1\\) \\), angle = 100 \\)"
+
+gdb_test "whatis wp" "type = Type waypoint"
+set output_pass [multi_line "type = Type, extends\\(point\\) :: waypoint" \
+              "    Type point :: point" \
+              "    $real :: angle" \
+              "End Type waypoint"]
+set output_kfail [multi_line "type = Type waypoint" \
+"    Type point :: point" \
+"    $real :: angle" \
+"End Type waypoint"]
+set test "ptype wp"
+gdb_test_multiple $test %test {
+    -re "$output_pass\r\n$gdb_prompt $" {
+      pass "$test"
+    }
+    -re "$output_kfail\r\n$gdb_prompt $" {
+      kfail "gcc/49475" "$test"
+    }
+}
+
+set test "ptype wp%coo"
+gdb_test_multiple "$test" "$test" {
+    -re "$real \\(3\\)\r\n$gdb_prompt $" {
+      pass "$test"
+    }
+    -re "There is no member named coo.\r\n$gdb_prompt $" {
+      kfail "gcc/49475" "$test"
+    }
+}
+gdb_test "ptype wp%point%coo" "$real \\(3\\)"
+
+set test "p wp_vla(1)%coo"
+gdb_test_multiple "$test" "$test" {
+    -re " = \\(10, 12, 10\\)\r\n$gdb_prompt $" {
+      pass "$test"
+    }
+    -re "There is no member named coo.\r\n$gdb_prompt $" {
+      kfail "gcc/49475" "$test"
+    }
+}
+gdb_test "p wp_vla(1)%point%coo" " = \\(10, 12, 10\\)"
+gdb_test "p wp_vla(1)%point" " = \\( coo = \\(10, 12, 10\\) \\)"
+gdb_test "p wp_vla(1)" " = \\( point = \\( coo = \\(10, 12, 10\\) \\), angle = 101 \\)"
+
+gdb_test "whatis wp_vla" "type = Type waypoint \\(3\\)"
+set test "ptype wp_vla"
+gdb_test_multiple $test %test {
+    -re "$output_pass \\(3\\)\r\n$gdb_prompt $" {
+      pass "$test"
+    }
+    -re "$output_kfail \\(3\\)\r\n$gdb_prompt $" {
+      kfail "gcc/49475" "$test"
+    }
+}
+set test "ptype wp_vla(1)%coo"
+gdb_test_multiple "$test" "$test" {
+    -re "$real \\(3\\)\r\n$gdb_prompt $" {
+      pass "$test"
+    }
+    -re "There is no member named coo.\r\n$gdb_prompt $" {
+      kfail "gcc/49475" "$test"
+    }
+}
+gdb_test "ptype wp_vla(1)%point%coo" "$real \\(3\\)"
diff --git a/gdb/testsuite/gdb.fortran/oop-extend-type.f90 b/gdb/testsuite/gdb.fortran/oop-extend-type.f90
new file mode 100644
index 000000000000..1fe8611f4632
--- /dev/null
+++ b/gdb/testsuite/gdb.fortran/oop-extend-type.f90
@@ -0,0 +1,56 @@
+! Copyright 2018 Free Software Foundation, Inc.
+! 
+! This program is free software; you can redistribute it and/or modify
+! it under the terms of the GNU General Public License as published by
+! the Free Software Foundation; either version 3 of the License, or
+! (at your option) any later version.
+! 
+! This program is distributed in the hope that it will be useful,
+! but WITHOUT ANY WARRANTY; without even the implied warranty of
+! MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+! GNU General Public License for more details.
+! 
+! You should have received a copy of the GNU General Public License
+! along with this program.  If not, see <http://www.gnu.org/licenses/>.
+
+module testmod
+    implicit none
+    type :: point
+        real :: coo(3)
+    end type
+
+    type, extends(point) :: waypoint
+        real :: angle
+    end type
+
+end module
+
+program testprog
+    use testmod
+    implicit none
+
+    logical l
+    type(waypoint) :: wp
+    type(waypoint), allocatable :: wp_vla(:)
+
+    l = allocated(wp_vla)
+    allocate(wp_vla(3))               ! Before vla allocation
+
+    l = allocated(wp_vla)             ! After vla allocation
+    wp%angle = 100.00
+    wp%point%coo(:) = 1.00
+    wp%point%coo(2) = 2.00
+
+    wp_vla(1)%angle = 101.00
+    wp_vla(1)%point%coo(:) = 10.00
+    wp_vla(1)%point%coo(2) = 12.00
+    wp_vla(2)%angle = 102.00
+    wp_vla(2)%point%coo(:) = 20.00
+    wp_vla(2)%point%coo(2) = 22.00
+    wp_vla(3)%angle = 103.00
+    wp_vla(3)%point%coo(:) = 30.00
+    wp_vla(3)%point%coo(2) = 32.00
+
+    print *, wp, wp_vla               ! After value assignment
+
+end program
diff --git a/gdb/valops.c b/gdb/valops.c
index c45caefbf1e0..a34e74b2bee9 100644
--- a/gdb/valops.c
+++ b/gdb/valops.c
@@ -2163,6 +2163,12 @@ value_struct_elt (struct value **argp, struct value **args,
       if (v)
 	return v;
 
+      /* fortran: If it is not a field it is the
+         type name of an inherited structure */
+      v = search_struct_field (name, *argp, t, 1);
+      if (v)
+	return v;
+
       /* C++: If it was not found as a data field, then try to
          return it as a pointer to a method.  */
       v = search_struct_method (name, argp, args, 0, 
-- 
2.7.4

[PATCH v2 5/7] Fortran: Enable setting breakpoint on nested functions.

[Pawel Wodkowski]

From: Bernhard Heckel <bernhard.heckel@intel.com>

Like in Ada, we want to be able to set a breakpoint on
nested functions, called "contained routines" in Fortran.

2016-06-30  Bernhard Heckel  <bernhard.heckel@intel.com>

gdb/Changelog:
	* dwarf2read.c (add_partial_symbol): store fortran subprograms
	in global scope.
	(add_partial_subprogram): Save fortran DW_TAG_subprogram.

gdb/Changelog:
	* gdb.fortran/nested-funcs.exp: Set breakpoint on contained routines.

Changes from V1 to V2:
	Fill commit message gaps.
---
 gdb/dwarf2read.c                           | 9 +++++++--
 gdb/testsuite/gdb.fortran/nested-funcs.exp | 7 +++++++
 2 files changed, 14 insertions(+), 2 deletions(-)

diff --git a/gdb/dwarf2read.c b/gdb/dwarf2read.c
index 88e57d7ab68e..dcf9da8372df 100644
--- a/gdb/dwarf2read.c
+++ b/gdb/dwarf2read.c
@@ -8910,7 +8910,9 @@ add_partial_symbol (struct partial_die_info *pdi, struct dwarf2_cu *cu)
     case DW_TAG_subprogram:
       addr = (gdbarch_adjust_dwarf2_addr (gdbarch, pdi->lowpc + baseaddr)
 	      - baseaddr);
-      if (pdi->is_external || cu->language == language_ada)
+      if (pdi->is_external
+	  || cu->language == language_ada
+	  || cu->language == language_fortran)
 	{
           /* brobecker/2007-12-26: Normally, only "external" DIEs are part
              of the global scope.  But in Ada, we want to be able to access
@@ -9206,6 +9208,8 @@ add_partial_subprogram (struct partial_die_info *pdi,
 	{
 	  if (pdi->tag == DW_TAG_entry_point)
 	    add_partial_entry_point (pdi, lowpc, highpc, set_addrmap, cu);
+	  else if (pdi->tag == DW_TAG_subprogram)
+	    add_partial_subprogram (pdi, lowpc, highpc, set_addrmap, cu);
 	  pdi = pdi->die_sibling;
 	}
     }
@@ -21543,7 +21547,8 @@ new_symbol (struct die_info *die, struct type *type, struct dwarf2_cu *cu,
 	  SYMBOL_ACLASS_INDEX (sym) = LOC_BLOCK;
 	  attr2 = dwarf2_attr (die, DW_AT_external, cu);
 	  if ((attr2 && (DW_UNSND (attr2) != 0))
-              || cu->language == language_ada)
+	      || cu->language == language_ada
+	      || cu->language == language_fortran)
 	    {
               /* Subprograms marked external are stored as a global symbol.
                  Ada subprograms, whether marked external or not, are always
diff --git a/gdb/testsuite/gdb.fortran/nested-funcs.exp b/gdb/testsuite/gdb.fortran/nested-funcs.exp
index 2481d256af69..5e28b0e1a3e2 100755
--- a/gdb/testsuite/gdb.fortran/nested-funcs.exp
+++ b/gdb/testsuite/gdb.fortran/nested-funcs.exp
@@ -29,6 +29,9 @@ if ![runto MAIN__] then {
     perror "couldn't run to breakpoint MAIN__"
     continue
 }
+# Test if we can set a breakpoint in a nested function
+gdb_breakpoint "sub_nested_outer"
+gdb_continue_to_breakpoint "sub_nested_outer" ".*local_int = 19"
 
 # Test if we can access local and
 # non-local variables defined one level up.
@@ -43,6 +46,10 @@ gdb_test "print local_int" "= 19" "print local_int in outer function"
 gdb_test "up"
 gdb_test "print index" "= 42" "print index at BP1, one frame up"
 
+# Test if we can set a breakpoint in a nested function
+gdb_breakpoint "sub_nested_inner"
+gdb_continue_to_breakpoint "sub_nested_inner" ".*local_int = 17"
+
 # Test if we can access local and
 # non-local variables defined two level up.
 gdb_breakpoint [gdb_get_line_number "! BP_inner"]
-- 
2.7.4

[PATCH v2 7/7] Fortran: Document scope operator.

[Pawel Wodkowski]

From: Bernhard Heckel <bernhard.heckel@intel.com>

Document scope operator in GDB manual.

2016-08-10  Bernhard Heckel  <bernhard.heckel@intel.com>

gdb/Changelog:
	* doc/gdb.texinfo (Fortran Operators): Describe scope operator.

Changes from V1 to V2:
	refactored entry in texinfo file, updated commit message.
---
 gdb/doc/gdb.texinfo | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/gdb/doc/gdb.texinfo b/gdb/doc/gdb.texinfo
index 944a2c438373..5eb3d6abfee5 100644
--- a/gdb/doc/gdb.texinfo
+++ b/gdb/doc/gdb.texinfo
@@ -15882,6 +15882,10 @@ The access component operator.  Normally used to access elements in derived
 types.  Also suitable for unions.  As unions aren't part of regular Fortran,
 this can only happen when accessing a register that uses a gdbarch-defined
 union type.
+@item ::
+The scope operator.  Normally used to access variables in modules or
+to set breakpoints on subroutines nested in modules or in other
+subroutines (internal subroutines).
 @end table
 
 @node Fortran Defaults
-- 
2.7.4

[PATCH v2 2/7] Dwarf: Fortran, support DW_TAG_entry_point.

[Pawel Wodkowski]

From: Bernhard Heckel <bernhard.heckel@intel.com>

Fortran provides additional entry-points to an subprogram.
Those entry-points may have only a subset of parameters
of the original subprogram as well.
Add support for parsing DW_TAG_entry_point's for Fortran.

2016-06-01  Bernhard Heckel  <bernhard.heckel@intel.com>

gdb/Changelog:
	* gdb/dwarf2read.c (add_partial_symbol): Handle DW_TAG_entry_point.
	(add_partial_entry_point): New.
	(add_partial_subprogram): Search for entry_points.
	(process_die): Handle DW_TAG_entry_point.
	(dwarf2_get_pc_bounds): Update low pc from DWARF.
	(load_partial_dies): Save DW_TAG_entry_point's.
	(load_partial_dies): Save DW_TAG_entry_point to hash table.
	(load_partial_dies): Look into child's of DW_TAG_sub_program
	for fortran.
	(new_symbol_full): Process DW_TAG_entry_point.
	(read_type_die_1): Handle DW_TAG_entry_point.

gdb/Testsuite/Changelog:
	* gdb.fortran/entry_point.f90: New.
	* gdb.fortran/entry_point.exp: New.
---
 gdb/dwarf2read.c                          | 98 ++++++++++++++++++++++++++++++-
 gdb/testsuite/gdb.fortran/entry-point.exp | 70 ++++++++++++++++++++++
 gdb/testsuite/gdb.fortran/entry-point.f90 | 48 +++++++++++++++
 3 files changed, 215 insertions(+), 1 deletion(-)
 create mode 100644 gdb/testsuite/gdb.fortran/entry-point.exp
 create mode 100644 gdb/testsuite/gdb.fortran/entry-point.f90

diff --git a/gdb/dwarf2read.c b/gdb/dwarf2read.c
index 89fd4ae15e80..88e57d7ab68e 100644
--- a/gdb/dwarf2read.c
+++ b/gdb/dwarf2read.c
@@ -1473,6 +1473,10 @@ static void add_partial_module (struct partial_die_info *pdi, CORE_ADDR *lowpc,
 static void add_partial_enumeration (struct partial_die_info *enum_pdi,
 				     struct dwarf2_cu *cu);
 
+static void add_partial_entry_point (struct partial_die_info *pdi,
+				     CORE_ADDR *lowpc, CORE_ADDR *highpc,
+				     int need_pc, struct dwarf2_cu *cu);
+
 static void add_partial_subprogram (struct partial_die_info *pdi,
 				    CORE_ADDR *lowpc, CORE_ADDR *highpc,
 				    int need_pc, struct dwarf2_cu *cu);
@@ -8876,6 +8880,32 @@ add_partial_symbol (struct partial_die_info *pdi, struct dwarf2_cu *cu)
 
   switch (pdi->tag)
     {
+    case DW_TAG_entry_point:
+      /* Don't know any other language than fortran which is
+	 using DW_TAG_entry_point.  */
+      if (cu->language == language_fortran)
+	{
+	  addr = gdbarch_adjust_dwarf2_addr (gdbarch, pdi->lowpc + baseaddr);
+	  /* DW_TAG_entry_point provides an additional entry_point to an
+	     existing sub_program. Therefore, we inherit the "external"
+	     attribute from the sub_program to which the entry_point
+	     belongs to.  */
+	  if (pdi->die_parent->is_external)
+	    add_psymbol_to_list (actual_name, strlen (actual_name),
+				 built_actual_name != nullptr,
+				 VAR_DOMAIN, LOC_BLOCK,
+				 SECT_OFF_TEXT (objfile),
+				 &objfile->global_psymbols,
+				 addr, cu->language, objfile);
+	  else
+	    add_psymbol_to_list (actual_name, strlen (actual_name),
+				 built_actual_name != nullptr,
+				 VAR_DOMAIN, LOC_BLOCK,
+				 SECT_OFF_TEXT (objfile),
+				 &objfile->static_psymbols,
+				 addr, cu->language, objfile);
+	}
+      break;
     case DW_TAG_inlined_subroutine:
     case DW_TAG_subprogram:
       addr = (gdbarch_adjust_dwarf2_addr (gdbarch, pdi->lowpc + baseaddr)
@@ -9082,6 +9112,17 @@ add_partial_module (struct partial_die_info *pdi, CORE_ADDR *lowpc,
     scan_partial_symbols (pdi->die_child, lowpc, highpc, set_addrmap, cu);
 }
 
+static void
+add_partial_entry_point (struct partial_die_info *pdi,
+			 CORE_ADDR *p_lowpc, CORE_ADDR *p_highpc,
+			 int set_addrmap, struct dwarf2_cu *cu)
+{
+  if (pdi->name == nullptr)
+    complaint (_("DW_TAG_entry_point must have a name"));
+  else
+    add_partial_symbol (pdi, cu);
+}
+
 /* Read a partial die corresponding to a subprogram or an inlined
    subprogram and create a partial symbol for that subprogram.
    When the CU language allows it, this routine also defines a partial
@@ -9158,6 +9199,16 @@ add_partial_subprogram (struct partial_die_info *pdi,
 	  pdi = pdi->die_sibling;
 	}
     }
+  else if (cu->language == language_fortran)
+    {
+      pdi = pdi->die_child;
+      while (pdi != nullptr)
+	{
+	  if (pdi->tag == DW_TAG_entry_point)
+	    add_partial_entry_point (pdi, lowpc, highpc, set_addrmap, cu);
+	  pdi = pdi->die_sibling;
+	}
+    }
 }
 
 /* Read a partial die corresponding to an enumeration type.  */
@@ -10596,6 +10647,7 @@ process_die (struct die_info *die, struct dwarf2_cu *cu)
     case DW_TAG_type_unit:
       read_type_unit_scope (die, cu);
       break;
+    case DW_TAG_entry_point:
     case DW_TAG_subprogram:
     case DW_TAG_inlined_subroutine:
       read_func_scope (die, cu);
@@ -14669,6 +14721,26 @@ dwarf2_get_pc_bounds (struct die_info *die, CORE_ADDR *lowpc,
   CORE_ADDR high = 0;
   enum pc_bounds_kind ret;
 
+  if (die->tag == DW_TAG_entry_point)
+    {
+      /* Entry_point is embedded in an subprogram. Therefore, we can use
+	 the highpc from it's enveloping subprogram and get the
+	 lowpc from DWARF.  */
+      if (0 == dwarf2_get_pc_bounds (die->parent, lowpc, highpc, cu, pst))
+	return PC_BOUNDS_NOT_PRESENT;
+
+      attr = dwarf2_attr (die, DW_AT_low_pc, cu);
+      if (!attr)
+	{
+	  complaint (_("DW_TAG_entry_point is missing DW_AT_low_pc"));
+	  return PC_BOUNDS_NOT_PRESENT;
+	}
+      low = attr_value_as_address (attr);
+      *lowpc = low;
+
+      return PC_BOUNDS_HIGH_LOW;
+    }
+
   attr_high = dwarf2_attr (die, DW_AT_high_pc, cu);
   if (attr_high)
     {
@@ -18399,6 +18471,7 @@ load_partial_dies (const struct die_reader_specs *reader,
 	  && !is_type_tag_for_partial (abbrev->tag)
 	  && abbrev->tag != DW_TAG_constant
 	  && abbrev->tag != DW_TAG_enumerator
+	  && abbrev->tag != DW_TAG_entry_point
 	  && abbrev->tag != DW_TAG_subprogram
 	  && abbrev->tag != DW_TAG_inlined_subroutine
 	  && abbrev->tag != DW_TAG_lexical_block
@@ -18529,6 +18602,7 @@ load_partial_dies (const struct die_reader_specs *reader,
 
       if (load_all
 	  || abbrev->tag == DW_TAG_constant
+	  || abbrev->tag == DW_TAG_entry_point
 	  || abbrev->tag == DW_TAG_subprogram
 	  || abbrev->tag == DW_TAG_variable
 	  || abbrev->tag == DW_TAG_namespace
@@ -18570,7 +18644,9 @@ load_partial_dies (const struct die_reader_specs *reader,
 		      || last_die->tag == DW_TAG_union_type))
 	      || (cu->language == language_ada
 		  && (last_die->tag == DW_TAG_subprogram
-		      || last_die->tag == DW_TAG_lexical_block))))
+		      || last_die->tag == DW_TAG_lexical_block))
+	      || (cu->language == language_fortran
+		  && last_die->tag == DW_TAG_subprogram)))
 	{
 	  nesting_level++;
 	  parent_die = last_die;
@@ -21442,6 +21518,25 @@ new_symbol (struct die_info *die, struct type *type, struct dwarf2_cu *cu,
 	  SYMBOL_ACLASS_INDEX (sym) = LOC_LABEL;
 	  dw2_add_symbol_to_list (sym, cu->list_in_scope);
 	  break;
+	case DW_TAG_entry_point:
+	  /* Don't know any other language than fortran which is
+	     using DW_TAG_entry_point.  */
+	  if (cu->language == language_fortran)
+	    {
+	      /* SYMBOL_BLOCK_VALUE (sym) will be filled in later by
+		 finish_block.  */
+	      SYMBOL_ACLASS_INDEX (sym) = LOC_BLOCK;
+	      /* DW_TAG_entry_point provides an additional entry_point to an
+		 existing sub_program. Therefore, we inherit the "external"
+		 attribute from the sub_program to which the entry_point
+		 belongs to.  */
+	      attr2 = dwarf2_attr (die->parent, DW_AT_external, cu);
+	      if (attr2 && (DW_UNSND (attr2) != 0))
+		list_to_add = cu->builder->get_global_symbols ();
+	      else
+		list_to_add = cu->list_in_scope;
+	    }
+	  break;
 	case DW_TAG_subprogram:
 	  /* SYMBOL_BLOCK_VALUE (sym) will be filled in later by
 	     finish_block.  */
@@ -22129,6 +22224,7 @@ read_type_die_1 (struct die_info *die, struct dwarf2_cu *cu)
     case DW_TAG_enumeration_type:
       this_type = read_enumeration_type (die, cu);
       break;
+    case DW_TAG_entry_point:
     case DW_TAG_subprogram:
     case DW_TAG_subroutine_type:
     case DW_TAG_inlined_subroutine:
diff --git a/gdb/testsuite/gdb.fortran/entry-point.exp b/gdb/testsuite/gdb.fortran/entry-point.exp
new file mode 100644
index 000000000000..a1f3f2bebdb9
--- /dev/null
+++ b/gdb/testsuite/gdb.fortran/entry-point.exp
@@ -0,0 +1,70 @@
+# Copyright 2018 Free Software Foundation, Inc.
+
+# This program is free software; you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation; either version 3 of the License, or
+# (at your option) any later version.
+#
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# along with this program.  If not, see <http://www.gnu.org/licenses/>.
+
+if { [skip_fortran_tests] } { return -1 }
+
+standard_testfile .f90
+load_lib "fortran.exp"
+
+if {[prepare_for_testing $testfile.exp $testfile $srcfile {debug f90}]} {
+    return -1
+}
+
+if ![runto MAIN__] then {
+    perror "couldn't run to breakpoint MAIN__"
+    continue
+}
+
+# Test if we can set a breakpoint via entry-point name
+set ept_name "foo"
+gdb_breakpoint $ept_name
+gdb_test "continue" \
+    [multi_line "Breakpoint $decimal, $ept_name \\(j=1, k=2, l=3, i1=4\\) at .*" \
+                ".*"] \
+    "continue to breakpoint: $ept_name"
+
+gdb_test "print j" "= 1" "print j, entered via $ept_name"
+gdb_test "print k" "= 2" "print k, entered via $ept_name"
+gdb_test "print l" "= 3" "print l, entered via $ept_name"
+gdb_test "print i1" "= 4" "print i1, entered via $ept_name"
+gdb_test "info args" \
+  [multi_line "j = 1" \
+              "k = 2" \
+              "l = 3" \
+              "i1 = 4"] \
+   "info args, entered via $ept_name"
+
+# Test if we can set a breakpoint via function name
+set ept_name "bar"
+gdb_breakpoint $ept_name
+gdb_test "continue" \
+    [multi_line "Breakpoint $decimal, $ept_name \\(i=4, j=5, k=6, i1=7\\) at .*" \
+                ".*"] \
+    "continue to breakpoint: $ept_name"
+
+gdb_test "print i" "= 4" "print i, entered via $ept_name"
+gdb_test "print j" "= 5" "print j, entered via $ept_name"
+gdb_test "print k" "= 6" "print k, entered via $ept_name"
+gdb_test "print i1" "= 7" "print i1, entered via $ept_name"
+
+set ept_name "tim"
+gdb_breakpoint $ept_name
+gdb_test "continue" \
+    [multi_line "Breakpoint $decimal, $ept_name \\(j=1\\) at .*" \
+                ".*"] \
+    "continue to breakpoint: $ept_name"
+
+gdb_test "print j" "= 1" "print j, entered via $ept_name"
+gdb_test "info args" "j = 1" "info args, entered via $ept_name"
diff --git a/gdb/testsuite/gdb.fortran/entry-point.f90 b/gdb/testsuite/gdb.fortran/entry-point.f90
new file mode 100644
index 000000000000..cb663b956982
--- /dev/null
+++ b/gdb/testsuite/gdb.fortran/entry-point.f90
@@ -0,0 +1,48 @@
+! Copyright 2018 Free Software Foundation, Inc.
+!
+! This program is free software; you can redistribute it and/or modify
+! it under the terms of the GNU General Public License as published by
+! the Free Software Foundation; either version 3 of the License, or
+! (at your option) any later version.
+!
+! This program is distributed in the hope that it will be useful,
+! but WITHOUT ANY WARRANTY; without even the implied warranty of
+! MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+! GNU General Public License for more details.
+!
+! You should have received a copy of the GNU General Public License
+! along with this program.  If not, see <http://www.gnu.org/licenses/>.
+
+program TestEntryPoint
+
+  call foo(1,2,3,4)
+  call bar(4,5,6,7)
+  call tim(1)
+
+end program TestEntryPoint
+
+  subroutine bar(I,J,K,I1)
+    INTEGER I,J,K,L,I1
+    INTEGER A
+    REAL    C
+
+    A = 0
+    C = 0.0
+
+    A = I + K + I1
+    goto 1000
+
+    entry foo(J,K,L,I1)
+    A = J + K + L + I1
+
+200 C = J
+    goto 1000
+
+    entry tim(J)
+    goto 200
+
+1000 A = C + 1
+     C = J * 1.5
+
+    return
+  end subroutine
-- 
2.7.4

[PATCH v2 4/7] Fortran: Ptype, print type extension.

[Pawel Wodkowski]

From: Bernhard Heckel <bernhard.heckel@intel.com>

Print base-class of an extended type when doing a ptype.

2016-05-24 Bernhard Heckel  <bernhard.heckel@intel.com>

gdb/Changelog:
	* gdb/f-typeprint.c (f_type_print_derivation_info): New.
	(f_type_print_base): Print baseclass info.

Changes from V1 to V2:
	clear commit message.
---
 gdb/f-typeprint.c | 30 +++++++++++++++++++++++++++---
 1 file changed, 27 insertions(+), 3 deletions(-)

diff --git a/gdb/f-typeprint.c b/gdb/f-typeprint.c
index 3f13c111ac56..133eaf9b9848 100644
--- a/gdb/f-typeprint.c
+++ b/gdb/f-typeprint.c
@@ -256,6 +256,25 @@ f_type_print_varspec_suffix (struct type *type, struct ui_file *stream,
     }
 }
 
+/* If TYPE is an extended type, then print out derivation information.
+
+   A typical output could look like this:
+   "Type, extends(point) :: waypoint"
+   "    Type point :: point"
+   "    real(kind=4) :: angle"
+   "End Type waypoint"
+ */
+
+static void
+f_type_print_derivation_info (struct type *type, struct ui_file *stream)
+{
+  /* Fortran doesn't support multiple inheritance. */
+  int i = 0;
+
+  if (TYPE_N_BASECLASSES (type) > 0)
+    fprintf_filtered (stream, ", extends(%s) ::", TYPE_NAME (TYPE_BASECLASS (type, i)));
+}
+
 /* Print the name of the type (or the ultimate pointer target,
    function value or array element), or the description of a
    structure or union.
@@ -372,10 +391,15 @@ f_type_print_base (struct type *type, struct ui_file *stream, int show,
     case TYPE_CODE_STRUCT:
     case TYPE_CODE_UNION:
       if (TYPE_CODE (type) == TYPE_CODE_UNION)
-	fprintfi_filtered (level, stream, "Type, C_Union :: ");
+	fprintfi_filtered (level, stream, "Type, C_Union ::");
       else
-	fprintfi_filtered (level, stream, "Type ");
-      fputs_filtered (TYPE_NAME (type), stream);
+	fprintfi_filtered (level, stream, "Type");
+
+      if (show > 0)
+	f_type_print_derivation_info (type, stream);
+
+      fprintf_filtered (stream, " %s", TYPE_NAME (type));
+
       /* According to the definition,
          we only print structure elements in case show > 0.  */
       if (show > 0)
-- 
2.7.4

Re: [PATCH v2 7/7] Fortran: Document scope operator.

[Eli Zaretskii]

> From: Pawel Wodkowski <pwodkowski@pl.sii.eu>
> Cc: tim.wiederhake@intel.com,	dragos.carciumaru@intel.com,	Bernhard Heckel <bernhard.heckel@intel.com>
> Date: Mon, 19 Nov 2018 22:38:50 +0100
> 
> From: Bernhard Heckel <bernhard.heckel@intel.com>
> 
> Document scope operator in GDB manual.
> 
> 2016-08-10  Bernhard Heckel  <bernhard.heckel@intel.com>
> 
> gdb/Changelog:
> 	* doc/gdb.texinfo (Fortran Operators): Describe scope operator.
> 
> Changes from V1 to V2:
> 	refactored entry in texinfo file, updated commit message.

OK for this part.

Thanks.

Re: [PATCH v2 7/7] Fortran: Document scope operator.

[Paweł Wódkowski]

On 20.11.2018 04:30, Eli Zaretskii wrote:
>> From: Pawel Wodkowski <pwodkowski@pl.sii.eu>
>> Cc: tim.wiederhake@intel.com,	dragos.carciumaru@intel.com,	Bernhard Heckel <bernhard.heckel@intel.com>
>> Date: Mon, 19 Nov 2018 22:38:50 +0100
>>
>> From: Bernhard Heckel <bernhard.heckel@intel.com>
>>
>> Document scope operator in GDB manual.
>>
>> 2016-08-10  Bernhard Heckel  <bernhard.heckel@intel.com>
>>
>> gdb/Changelog:
>> 	* doc/gdb.texinfo (Fortran Operators): Describe scope operator.
>>
>> Changes from V1 to V2:
>> 	refactored entry in texinfo file, updated commit message.
> 
> OK for this part.
> 
> Thanks.
> 

This is our first steps in GDB community so thanks for review.

Are we OK to push to all these changes or we need to wait for someone 
else to ack?

Pawel

Re: [PATCH v2 7/7] Fortran: Document scope operator.

[Eli Zaretskii]

> From: PaweÂł WĂłdkowski <pwodkowski@pl.sii.eu>
> CC: "gdb-patches@sourceware.org" <gdb-patches@sourceware.org>,
> 	MichaÂł UrbaĂąski <murbanski@pl.sii.eu>, "Sebastian
>  Basierski" <sbasierski@pl.sii.eu>, "tim.wiederhake@intel.com"
> 	<tim.wiederhake@intel.com>, "dragos.carciumaru@intel.com"
> 	<dragos.carciumaru@intel.com>, "bernhard.heckel@intel.com"
> 	<bernhard.heckel@intel.com>
> Date: Thu, 22 Nov 2018 16:13:19 +0000
> 
> >> From: Bernhard Heckel <bernhard.heckel@intel.com>
> >>
> >> Document scope operator in GDB manual.
> >>
> >> 2016-08-10  Bernhard Heckel  <bernhard.heckel@intel.com>
> >>
> >> gdb/Changelog:
> >> 	* doc/gdb.texinfo (Fortran Operators): Describe scope operator.
> >>
> >> Changes from V1 to V2:
> >> 	refactored entry in texinfo file, updated commit message.
> > 
> > OK for this part.
> > 
> > Thanks.
> > 
> 
> This is our first steps in GDB community so thanks for review.
> 
> Are we OK to push to all these changes or we need to wait for someone 
> else to ack?

I only approved part 7 of your patch series.  I cannot approve the
rest, so please wait for others to chime in.

Thanks.

PING Re: [PATCH v2 1/7] DWARF: Don't add nameless modules to partial symbol table.

[Pawel Wodkowski]

Hi,

Can kindly ask for review :)
Thanks

Pawel


On 19.11.2018 22:38, Pawel Wodkowski wrote:
> From: Bernhard Heckel <bernhard.heckel@intel.com>
>
> A name for BLOCK DATA in Fortran is optional. If no
> name has been assigned, GDB will crash during read-in of DWARF
> when BLOCK DATA is represented via DW_TAG_module.
> BLOCK DATA is used for one-time initialization of
> non-pointer variables in named common blocks.
>
> 2016-06-15  Bernhard Heckel  <bernhard.heckel@intel.com>
>
> gdb/Changelog:
> 	* dwarf2read.c (add_partial_symbol): Skip nameless modules.
>
> gdb/Testsuite/Changelog:
> 	* gdb.fortran/block-data.f: New.
> 	* gdb.fortran/block-data.exp: New.
> ---
>   gdb/dwarf2read.c                         | 13 ++++---
>   gdb/testsuite/gdb.fortran/block-data.exp | 51 ++++++++++++++++++++++++++++
>   gdb/testsuite/gdb.fortran/block-data.f   | 58 ++++++++++++++++++++++++++++++++
>   3 files changed, 117 insertions(+), 5 deletions(-)
>   create mode 100644 gdb/testsuite/gdb.fortran/block-data.exp
>   create mode 100644 gdb/testsuite/gdb.fortran/block-data.f
>
> diff --git a/gdb/dwarf2read.c b/gdb/dwarf2read.c
> index d2a8cd44f9a5..89fd4ae15e80 100644
> --- a/gdb/dwarf2read.c
> +++ b/gdb/dwarf2read.c
> @@ -8995,11 +8995,14 @@ add_partial_symbol (struct partial_die_info *pdi, struct dwarf2_cu *cu)
>   			   0, cu->language, objfile);
>         break;
>       case DW_TAG_module:
> -      add_psymbol_to_list (actual_name, strlen (actual_name),
> -			   built_actual_name != NULL,
> -			   MODULE_DOMAIN, LOC_TYPEDEF, -1,
> -			   &objfile->global_psymbols,
> -			   0, cu->language, objfile);
> +      /* In Fortran 77 there might be a "BLOCK DATA" module available wihout
> +         any name. If so, we skip the module as it doesn't bring any value */
> +      if (actual_name != nullptr)
> +	add_psymbol_to_list (actual_name, strlen (actual_name),
> +			     built_actual_name != nullptr,
> +			     MODULE_DOMAIN, LOC_TYPEDEF, -1,
> +			     &objfile->global_psymbols,
> +			     0, cu->language, objfile);
>         break;
>       case DW_TAG_class_type:
>       case DW_TAG_interface_type:
> diff --git a/gdb/testsuite/gdb.fortran/block-data.exp b/gdb/testsuite/gdb.fortran/block-data.exp
> new file mode 100644
> index 000000000000..2af250ad3886
> --- /dev/null
> +++ b/gdb/testsuite/gdb.fortran/block-data.exp
> @@ -0,0 +1,51 @@
> +# Copyright 2018 Free Software Foundation, Inc.
> +#
> +# Contributed by Intel Corp. <bernhard.heckel@intel.com>
> +#
> +# This program is free software; you can redistribute it and/or modify
> +# it under the terms of the GNU General Public License as published by
> +# the Free Software Foundation; either version 3 of the License, or
> +# (at your option) any later version.
> +#
> +# This program is distributed in the hope that it will be useful,
> +# but WITHOUT ANY WARRANTY; without even the implied warranty of
> +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
> +# GNU General Public License for more details.
> +#
> +# You should have received a copy of the GNU General Public License
> +# along with this program.  If not, see <http://www.gnu.org/licenses/>.
> +
> +# This test is supposed to test anonymous block-data statement.
> +
> +if { [skip_fortran_tests] } { return -1 }
> +
> +standard_testfile .f
> +load_lib "fortran.exp"
> +
> +if {[prepare_for_testing $testfile.exp $testfile $srcfile {debug f90}]} {
> +    return -1
> +}
> +
> +if ![runto MAIN__] then {
> +    untested "couldn't run to breakpoint MAIN__"
> +    return -1
> +}
> +
> +gdb_test "print doub1" "= 1.11\\d+" "print doub1, default values"
> +gdb_test "print doub2" "= 2.22\\d+" "print doub2, default values"
> +gdb_test "print char1" "= 'abcdef'" "print char1, default values"
> +gdb_test "print char2" "= 'ghijkl'" "print char2, default values"
> +
> +gdb_breakpoint [gdb_get_line_number "! BP_BEFORE_SUB"]
> +gdb_continue_to_breakpoint "! BP_BEFORE_SUB" ".*! BP_BEFORE_SUB.*"
> +gdb_test "print doub1" "= 11.11\\d+" "print doub1, before sub"
> +gdb_test "print doub2" "= 22.22\\d+" "print doub2, before sub"
> +gdb_test "print char1" "= 'ABCDEF'" "print char1, before sub"
> +gdb_test "print char2" "= 'GHIJKL'" "print char2, before sub"
> +
> +gdb_breakpoint [gdb_get_line_number "! BP_SUB"]
> +gdb_continue_to_breakpoint "! BP_SUB" ".*! BP_SUB.*"
> +gdb_test "print doub1" "= 11.11\\d+" "print char1, in sub"
> +gdb_test "print doub2" "= 22.22\\d+" "print doub2, in sub"
> +gdb_test "print char1" "= 'ABCDEF'" "print char1, in sub"
> +gdb_test "print char2" "= 'GHIJKL'" "print char2, in sub"
> diff --git a/gdb/testsuite/gdb.fortran/block-data.f b/gdb/testsuite/gdb.fortran/block-data.f
> new file mode 100644
> index 000000000000..a28e687ec885
> --- /dev/null
> +++ b/gdb/testsuite/gdb.fortran/block-data.f
> @@ -0,0 +1,58 @@
> +! Copyright 2018 Free Software Foundation, Inc.
> +!
> +! Contributed by Intel Corp. <bernhard.heckel@intel.com>
> +!
> +! This program is free software; you can redistribute it and/or modify
> +! it under the terms of the GNU General Public License as published by
> +! the Free Software Foundation; either version 3 of the License, or
> +! (at your option) any later version.
> +!
> +! This program is distributed in the hope that it will be useful,
> +! but WITHOUT ANY WARRANTY; without even the implied warranty of
> +! MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
> +! GNU General Public License for more details.
> +!
> +! You should have received a copy of the GNU General Public License
> +! along with this program.  If not, see <http://www.gnu.org/licenses/>.
> +!
> +! Test if GDB can handle block data without global name
> +!
> +! MAIN
> +        PROGRAM bdata
> +        DOUBLE PRECISION doub1, doub2
> +        CHARACTER*6 char1, char2
> +
> +        COMMON /BLK1/ doub1, char1
> +        COMMON /BLK2/ doub2, char2
> +
> +        doub1 = 11.111
> +        doub2 = 22.222
> +        char1 = 'ABCDEF'
> +        char2 = 'GHIJKL'
> +        CALL sub_block_data      ! BP_BEFORE_SUB
> +        STOP
> +        END
> +
> +! BLOCK DATA
> +        BLOCK DATA
> +
> +        DOUBLE PRECISION doub1, doub2
> +        CHARACTER*6 char1, char2
> +
> +        COMMON /BLK1/ doub1, char1
> +        COMMON /BLK2/ doub2, char2
> +        DATA doub1, doub2 /1.111, 2.222/
> +        DATA char1, char2 /'abcdef', 'ghijkl'/
> +        END
> +
> +! SUBROUTINE
> +        SUBROUTINE sub_block_data
> +
> +        DOUBLE PRECISION doub1, doub2
> +        CHARACTER*6 char1, char2
> +
> +        COMMON /BLK1/ doub1, char1
> +        COMMON /BLK2/ doub2, char2
> +
> +        char1 = char2;    ! BP_SUB
> +        END

Re: [PATCH v2 1/7] DWARF: Don't add nameless modules to partial symbol table.

[Simon Marchi]

Adding Richard to the CC list.  Since you are actually working with 
Fortran, you might have some comments about this series.

Simon

On 2018-11-19 16:38, Pawel Wodkowski wrote:
> From: Bernhard Heckel <bernhard.heckel@intel.com>
> 
> A name for BLOCK DATA in Fortran is optional. If no
> name has been assigned, GDB will crash during read-in of DWARF
> when BLOCK DATA is represented via DW_TAG_module.
> BLOCK DATA is used for one-time initialization of
> non-pointer variables in named common blocks.
> 
> 2016-06-15  Bernhard Heckel  <bernhard.heckel@intel.com>
> 
> gdb/Changelog:
> 	* dwarf2read.c (add_partial_symbol): Skip nameless modules.
> 
> gdb/Testsuite/Changelog:
> 	* gdb.fortran/block-data.f: New.
> 	* gdb.fortran/block-data.exp: New.
> ---
>  gdb/dwarf2read.c                         | 13 ++++---
>  gdb/testsuite/gdb.fortran/block-data.exp | 51 
> ++++++++++++++++++++++++++++
>  gdb/testsuite/gdb.fortran/block-data.f   | 58 
> ++++++++++++++++++++++++++++++++
>  3 files changed, 117 insertions(+), 5 deletions(-)
>  create mode 100644 gdb/testsuite/gdb.fortran/block-data.exp
>  create mode 100644 gdb/testsuite/gdb.fortran/block-data.f
> 
> diff --git a/gdb/dwarf2read.c b/gdb/dwarf2read.c
> index d2a8cd44f9a5..89fd4ae15e80 100644
> --- a/gdb/dwarf2read.c
> +++ b/gdb/dwarf2read.c
> @@ -8995,11 +8995,14 @@ add_partial_symbol (struct partial_die_info
> *pdi, struct dwarf2_cu *cu)
>  			   0, cu->language, objfile);
>        break;
>      case DW_TAG_module:
> -      add_psymbol_to_list (actual_name, strlen (actual_name),
> -			   built_actual_name != NULL,
> -			   MODULE_DOMAIN, LOC_TYPEDEF, -1,
> -			   &objfile->global_psymbols,
> -			   0, cu->language, objfile);
> +      /* In Fortran 77 there might be a "BLOCK DATA" module available 
> wihout
> +         any name. If so, we skip the module as it doesn't bring any 
> value */
> +      if (actual_name != nullptr)
> +	add_psymbol_to_list (actual_name, strlen (actual_name),
> +			     built_actual_name != nullptr,
> +			     MODULE_DOMAIN, LOC_TYPEDEF, -1,
> +			     &objfile->global_psymbols,
> +			     0, cu->language, objfile);
>        break;
>      case DW_TAG_class_type:
>      case DW_TAG_interface_type:
> diff --git a/gdb/testsuite/gdb.fortran/block-data.exp
> b/gdb/testsuite/gdb.fortran/block-data.exp
> new file mode 100644
> index 000000000000..2af250ad3886
> --- /dev/null
> +++ b/gdb/testsuite/gdb.fortran/block-data.exp
> @@ -0,0 +1,51 @@
> +# Copyright 2018 Free Software Foundation, Inc.
> +#
> +# Contributed by Intel Corp. <bernhard.heckel@intel.com>
> +#
> +# This program is free software; you can redistribute it and/or modify
> +# it under the terms of the GNU General Public License as published by
> +# the Free Software Foundation; either version 3 of the License, or
> +# (at your option) any later version.
> +#
> +# This program is distributed in the hope that it will be useful,
> +# but WITHOUT ANY WARRANTY; without even the implied warranty of
> +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
> +# GNU General Public License for more details.
> +#
> +# You should have received a copy of the GNU General Public License
> +# along with this program.  If not, see 
> <http://www.gnu.org/licenses/>.
> +
> +# This test is supposed to test anonymous block-data statement.
> +
> +if { [skip_fortran_tests] } { return -1 }
> +
> +standard_testfile .f
> +load_lib "fortran.exp"
> +
> +if {[prepare_for_testing $testfile.exp $testfile $srcfile {debug 
> f90}]} {
> +    return -1
> +}
> +
> +if ![runto MAIN__] then {
> +    untested "couldn't run to breakpoint MAIN__"
> +    return -1
> +}
> +
> +gdb_test "print doub1" "= 1.11\\d+" "print doub1, default values"
> +gdb_test "print doub2" "= 2.22\\d+" "print doub2, default values"
> +gdb_test "print char1" "= 'abcdef'" "print char1, default values"
> +gdb_test "print char2" "= 'ghijkl'" "print char2, default values"
> +
> +gdb_breakpoint [gdb_get_line_number "! BP_BEFORE_SUB"]
> +gdb_continue_to_breakpoint "! BP_BEFORE_SUB" ".*! BP_BEFORE_SUB.*"
> +gdb_test "print doub1" "= 11.11\\d+" "print doub1, before sub"
> +gdb_test "print doub2" "= 22.22\\d+" "print doub2, before sub"
> +gdb_test "print char1" "= 'ABCDEF'" "print char1, before sub"
> +gdb_test "print char2" "= 'GHIJKL'" "print char2, before sub"
> +
> +gdb_breakpoint [gdb_get_line_number "! BP_SUB"]
> +gdb_continue_to_breakpoint "! BP_SUB" ".*! BP_SUB.*"
> +gdb_test "print doub1" "= 11.11\\d+" "print char1, in sub"
> +gdb_test "print doub2" "= 22.22\\d+" "print doub2, in sub"
> +gdb_test "print char1" "= 'ABCDEF'" "print char1, in sub"
> +gdb_test "print char2" "= 'GHIJKL'" "print char2, in sub"
> diff --git a/gdb/testsuite/gdb.fortran/block-data.f
> b/gdb/testsuite/gdb.fortran/block-data.f
> new file mode 100644
> index 000000000000..a28e687ec885
> --- /dev/null
> +++ b/gdb/testsuite/gdb.fortran/block-data.f
> @@ -0,0 +1,58 @@
> +! Copyright 2018 Free Software Foundation, Inc.
> +!
> +! Contributed by Intel Corp. <bernhard.heckel@intel.com>
> +!
> +! This program is free software; you can redistribute it and/or modify
> +! it under the terms of the GNU General Public License as published by
> +! the Free Software Foundation; either version 3 of the License, or
> +! (at your option) any later version.
> +!
> +! This program is distributed in the hope that it will be useful,
> +! but WITHOUT ANY WARRANTY; without even the implied warranty of
> +! MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
> +! GNU General Public License for more details.
> +!
> +! You should have received a copy of the GNU General Public License
> +! along with this program.  If not, see 
> <http://www.gnu.org/licenses/>.
> +!
> +! Test if GDB can handle block data without global name
> +!
> +! MAIN
> +        PROGRAM bdata
> +        DOUBLE PRECISION doub1, doub2
> +        CHARACTER*6 char1, char2
> +
> +        COMMON /BLK1/ doub1, char1
> +        COMMON /BLK2/ doub2, char2
> +
> +        doub1 = 11.111
> +        doub2 = 22.222
> +        char1 = 'ABCDEF'
> +        char2 = 'GHIJKL'
> +        CALL sub_block_data      ! BP_BEFORE_SUB
> +        STOP
> +        END
> +
> +! BLOCK DATA
> +        BLOCK DATA
> +
> +        DOUBLE PRECISION doub1, doub2
> +        CHARACTER*6 char1, char2
> +
> +        COMMON /BLK1/ doub1, char1
> +        COMMON /BLK2/ doub2, char2
> +        DATA doub1, doub2 /1.111, 2.222/
> +        DATA char1, char2 /'abcdef', 'ghijkl'/
> +        END
> +
> +! SUBROUTINE
> +        SUBROUTINE sub_block_data
> +
> +        DOUBLE PRECISION doub1, doub2
> +        CHARACTER*6 char1, char2
> +
> +        COMMON /BLK1/ doub1, char1
> +        COMMON /BLK2/ doub2, char2
> +
> +        char1 = char2;    ! BP_SUB
> +        END

Re: [PATCH v2 1/7] DWARF: Don't add nameless modules to partial symbol table.

[Andrew Burgess]

* Pawel Wodkowski <pwodkowski@pl.sii.eu> [2018-11-19 22:38:44 +0100]:

> From: Bernhard Heckel <bernhard.heckel@intel.com>
> 
> A name for BLOCK DATA in Fortran is optional. If no
> name has been assigned, GDB will crash during read-in of DWARF
> when BLOCK DATA is represented via DW_TAG_module.
> BLOCK DATA is used for one-time initialization of
> non-pointer variables in named common blocks.

Thanks for this.

Locally I have:

  GNU Fortran (GCC) 7.3.1 20180712 (Red Hat 7.3.1-6)

For me at least this test passes with and without the change to GDB.

Can you confirm which Fortran compiler you see the issue with?  It
probably should be documented in the commit message and in comment
around the check in dwarf2read.c.  It could also be useful to include
examples of the problem DWARF in the commit message in cases like
this.

> 
> 2016-06-15  Bernhard Heckel  <bernhard.heckel@intel.com>
> 
> gdb/Changelog:
> 	* dwarf2read.c (add_partial_symbol): Skip nameless modules.
> 
> gdb/Testsuite/Changelog:

It's 'testsuite' all the patches in this series :)

> 	* gdb.fortran/block-data.f: New.
> 	* gdb.fortran/block-data.exp: New.
> ---
>  gdb/dwarf2read.c                         | 13 ++++---
>  gdb/testsuite/gdb.fortran/block-data.exp | 51 ++++++++++++++++++++++++++++
>  gdb/testsuite/gdb.fortran/block-data.f   | 58 ++++++++++++++++++++++++++++++++
>  3 files changed, 117 insertions(+), 5 deletions(-)
>  create mode 100644 gdb/testsuite/gdb.fortran/block-data.exp
>  create mode 100644 gdb/testsuite/gdb.fortran/block-data.f
> 
> diff --git a/gdb/dwarf2read.c b/gdb/dwarf2read.c
> index d2a8cd44f9a5..89fd4ae15e80 100644
> --- a/gdb/dwarf2read.c
> +++ b/gdb/dwarf2read.c
> @@ -8995,11 +8995,14 @@ add_partial_symbol (struct partial_die_info *pdi, struct dwarf2_cu *cu)
>  			   0, cu->language, objfile);
>        break;
>      case DW_TAG_module:
> -      add_psymbol_to_list (actual_name, strlen (actual_name),
> -			   built_actual_name != NULL,
> -			   MODULE_DOMAIN, LOC_TYPEDEF, -1,
> -			   &objfile->global_psymbols,
> -			   0, cu->language, objfile);
> +      /* In Fortran 77 there might be a "BLOCK DATA" module available wihout
> +         any name. If so, we skip the module as it doesn't bring any value */

You need a full-stop and two spaces at the end, like "...any value.  */".

> +      if (actual_name != nullptr)
> +	add_psymbol_to_list (actual_name, strlen (actual_name),
> +			     built_actual_name != nullptr,
> +			     MODULE_DOMAIN, LOC_TYPEDEF, -1,
> +			     &objfile->global_psymbols,
> +			     0, cu->language, objfile);
>        break;
>      case DW_TAG_class_type:
>      case DW_TAG_interface_type:
> diff --git a/gdb/testsuite/gdb.fortran/block-data.exp b/gdb/testsuite/gdb.fortran/block-data.exp
> new file mode 100644
> index 000000000000..2af250ad3886
> --- /dev/null
> +++ b/gdb/testsuite/gdb.fortran/block-data.exp
> @@ -0,0 +1,51 @@
> +# Copyright 2018 Free Software Foundation, Inc.
> +#
> +# Contributed by Intel Corp. <bernhard.heckel@intel.com>

Someone else might correct me, but I think "Contributed by..." lines
are not supposed to be added to GDB source..

> +#
> +# This program is free software; you can redistribute it and/or modify
> +# it under the terms of the GNU General Public License as published by
> +# the Free Software Foundation; either version 3 of the License, or
> +# (at your option) any later version.
> +#
> +# This program is distributed in the hope that it will be useful,
> +# but WITHOUT ANY WARRANTY; without even the implied warranty of
> +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
> +# GNU General Public License for more details.
> +#
> +# You should have received a copy of the GNU General Public License
> +# along with this program.  If not, see <http://www.gnu.org/licenses/>.
> +
> +# This test is supposed to test anonymous block-data statement.
> +
> +if { [skip_fortran_tests] } { return -1 }
> +
> +standard_testfile .f
> +load_lib "fortran.exp"
> +
> +if {[prepare_for_testing $testfile.exp $testfile $srcfile {debug f90}]} {
> +    return -1
> +}
> +
> +if ![runto MAIN__] then {
> +    untested "couldn't run to breakpoint MAIN__"
> +    return -1
> +}
> +
> +gdb_test "print doub1" "= 1.11\\d+" "print doub1, default values"
> +gdb_test "print doub2" "= 2.22\\d+" "print doub2, default values"
> +gdb_test "print char1" "= 'abcdef'" "print char1, default values"
> +gdb_test "print char2" "= 'ghijkl'" "print char2, default values"
> +
> +gdb_breakpoint [gdb_get_line_number "! BP_BEFORE_SUB"]
> +gdb_continue_to_breakpoint "! BP_BEFORE_SUB" ".*! BP_BEFORE_SUB.*"
> +gdb_test "print doub1" "= 11.11\\d+" "print doub1, before sub"
> +gdb_test "print doub2" "= 22.22\\d+" "print doub2, before sub"
> +gdb_test "print char1" "= 'ABCDEF'" "print char1, before sub"
> +gdb_test "print char2" "= 'GHIJKL'" "print char2, before sub"
> +
> +gdb_breakpoint [gdb_get_line_number "! BP_SUB"]
> +gdb_continue_to_breakpoint "! BP_SUB" ".*! BP_SUB.*"
> +gdb_test "print doub1" "= 11.11\\d+" "print char1, in sub"
> +gdb_test "print doub2" "= 22.22\\d+" "print doub2, in sub"
> +gdb_test "print char1" "= 'ABCDEF'" "print char1, in sub"
> +gdb_test "print char2" "= 'GHIJKL'" "print char2, in sub"
> diff --git a/gdb/testsuite/gdb.fortran/block-data.f b/gdb/testsuite/gdb.fortran/block-data.f
> new file mode 100644
> index 000000000000..a28e687ec885
> --- /dev/null
> +++ b/gdb/testsuite/gdb.fortran/block-data.f
> @@ -0,0 +1,58 @@
> +! Copyright 2018 Free Software Foundation, Inc.
> +!
> +! Contributed by Intel Corp. <bernhard.heckel@intel.com>
> +!
> +! This program is free software; you can redistribute it and/or modify
> +! it under the terms of the GNU General Public License as published by
> +! the Free Software Foundation; either version 3 of the License, or
> +! (at your option) any later version.
> +!
> +! This program is distributed in the hope that it will be useful,
> +! but WITHOUT ANY WARRANTY; without even the implied warranty of
> +! MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
> +! GNU General Public License for more details.
> +!
> +! You should have received a copy of the GNU General Public License
> +! along with this program.  If not, see <http://www.gnu.org/licenses/>.
> +!
> +! Test if GDB can handle block data without global name
> +!
> +! MAIN
> +        PROGRAM bdata
> +        DOUBLE PRECISION doub1, doub2
> +        CHARACTER*6 char1, char2
> +
> +        COMMON /BLK1/ doub1, char1
> +        COMMON /BLK2/ doub2, char2
> +
> +        doub1 = 11.111
> +        doub2 = 22.222
> +        char1 = 'ABCDEF'
> +        char2 = 'GHIJKL'
> +        CALL sub_block_data      ! BP_BEFORE_SUB
> +        STOP
> +        END
> +
> +! BLOCK DATA
> +        BLOCK DATA
> +
> +        DOUBLE PRECISION doub1, doub2
> +        CHARACTER*6 char1, char2
> +
> +        COMMON /BLK1/ doub1, char1
> +        COMMON /BLK2/ doub2, char2
> +        DATA doub1, doub2 /1.111, 2.222/
> +        DATA char1, char2 /'abcdef', 'ghijkl'/
> +        END
> +
> +! SUBROUTINE
> +        SUBROUTINE sub_block_data
> +
> +        DOUBLE PRECISION doub1, doub2
> +        CHARACTER*6 char1, char2
> +
> +        COMMON /BLK1/ doub1, char1
> +        COMMON /BLK2/ doub2, char2
> +
> +        char1 = char2;    ! BP_SUB
> +        END
> -- 
> 2.7.4
> 

Re: [PATCH v2 2/7] Dwarf: Fortran, support DW_TAG_entry_point.

[Andrew Burgess]

* Pawel Wodkowski <pwodkowski@pl.sii.eu> [2018-11-19 22:38:45 +0100]:

> From: Bernhard Heckel <bernhard.heckel@intel.com>
> 
> Fortran provides additional entry-points to an subprogram.
> Those entry-points may have only a subset of parameters
> of the original subprogram as well.
> Add support for parsing DW_TAG_entry_point's for Fortran.

As with patch #1 it's probably worth documenting which compilers you
see DW_TAG_entry_point in, along with an example of what the generated
DWARF looks like.

This means that in the future if/when others start to use this DWARF
feature we can figure out which other compilers we need to test.

> 
> 2016-06-01  Bernhard Heckel  <bernhard.heckel@intel.com>
> 
> gdb/Changelog:
> 	* gdb/dwarf2read.c (add_partial_symbol): Handle DW_TAG_entry_point.
> 	(add_partial_entry_point): New.
> 	(add_partial_subprogram): Search for entry_points.
> 	(process_die): Handle DW_TAG_entry_point.
> 	(dwarf2_get_pc_bounds): Update low pc from DWARF.
> 	(load_partial_dies): Save DW_TAG_entry_point's.
> 	(load_partial_dies): Save DW_TAG_entry_point to hash table.
> 	(load_partial_dies): Look into child's of DW_TAG_sub_program
> 	for fortran.
> 	(new_symbol_full): Process DW_TAG_entry_point.
> 	(read_type_die_1): Handle DW_TAG_entry_point.
> 
> gdb/Testsuite/Changelog:
> 	* gdb.fortran/entry_point.f90: New.
> 	* gdb.fortran/entry_point.exp: New.
> ---
>  gdb/dwarf2read.c                          | 98 ++++++++++++++++++++++++++++++-
>  gdb/testsuite/gdb.fortran/entry-point.exp | 70 ++++++++++++++++++++++
>  gdb/testsuite/gdb.fortran/entry-point.f90 | 48 +++++++++++++++
>  3 files changed, 215 insertions(+), 1 deletion(-)
>  create mode 100644 gdb/testsuite/gdb.fortran/entry-point.exp
>  create mode 100644 gdb/testsuite/gdb.fortran/entry-point.f90
> 
> diff --git a/gdb/dwarf2read.c b/gdb/dwarf2read.c
> index 89fd4ae15e80..88e57d7ab68e 100644
> --- a/gdb/dwarf2read.c
> +++ b/gdb/dwarf2read.c
> @@ -1473,6 +1473,10 @@ static void add_partial_module (struct partial_die_info *pdi, CORE_ADDR *lowpc,
>  static void add_partial_enumeration (struct partial_die_info *enum_pdi,
>  				     struct dwarf2_cu *cu);
>  
> +static void add_partial_entry_point (struct partial_die_info *pdi,
> +				     CORE_ADDR *lowpc, CORE_ADDR *highpc,
> +				     int need_pc, struct dwarf2_cu *cu);
> +
>  static void add_partial_subprogram (struct partial_die_info *pdi,
>  				    CORE_ADDR *lowpc, CORE_ADDR *highpc,
>  				    int need_pc, struct dwarf2_cu *cu);
> @@ -8876,6 +8880,32 @@ add_partial_symbol (struct partial_die_info *pdi, struct dwarf2_cu *cu)
>  
>    switch (pdi->tag)
>      {
> +    case DW_TAG_entry_point:
> +      /* Don't know any other language than fortran which is
> +	 using DW_TAG_entry_point.  */
> +      if (cu->language == language_fortran)
> +	{

I'm not sure the language check is needed here.  The description for
DW_TAG_entry_point in the DWARF spec seems pretty generic.  I don't
see how a different language would expect significantly different
results, and so, I would suggest we should just add this as general
purpose code.

> +	  addr = gdbarch_adjust_dwarf2_addr (gdbarch, pdi->lowpc + baseaddr);

When compared to DW_TAG_subprogram, this line doesn't include a '-
baseaddr'.  I think that deserves a comment explaining why.

> +	  /* DW_TAG_entry_point provides an additional entry_point to an
> +	     existing sub_program. Therefore, we inherit the "external"
> +	     attribute from the sub_program to which the entry_point
> +	     belongs to.  */
> +	  if (pdi->die_parent->is_external)
> +	    add_psymbol_to_list (actual_name, strlen (actual_name),
> +				 built_actual_name != nullptr,
> +				 VAR_DOMAIN, LOC_BLOCK,
> +				 SECT_OFF_TEXT (objfile),
> +				 &objfile->global_psymbols,
> +				 addr, cu->language, objfile);
> +	  else
> +	    add_psymbol_to_list (actual_name, strlen (actual_name),
> +				 built_actual_name != nullptr,
> +				 VAR_DOMAIN, LOC_BLOCK,
> +				 SECT_OFF_TEXT (objfile),
> +				 &objfile->static_psymbols,
> +				 addr, cu->language, objfile);

I think these two add_psymbol_to_list calls, and the two for
DW_TAG_subprogram should be factored out.  Just my opinion though...

> +	}
> +      break;
>      case DW_TAG_inlined_subroutine:
>      case DW_TAG_subprogram:
>        addr = (gdbarch_adjust_dwarf2_addr (gdbarch, pdi->lowpc + baseaddr)
> @@ -9082,6 +9112,17 @@ add_partial_module (struct partial_die_info *pdi, CORE_ADDR *lowpc,
>      scan_partial_symbols (pdi->die_child, lowpc, highpc, set_addrmap, cu);
>  }
>  
> +static void
> +add_partial_entry_point (struct partial_die_info *pdi,
> +			 CORE_ADDR *p_lowpc, CORE_ADDR *p_highpc,
> +			 int set_addrmap, struct dwarf2_cu *cu)
> +{
> +  if (pdi->name == nullptr)
> +    complaint (_("DW_TAG_entry_point must have a name"));
> +  else
> +    add_partial_symbol (pdi, cu);
> +}

This function should have a header comment, however, a quick look at
add_partial_subprogram seems to indicate that at this point GDB is not
complaining, but just ignoring weird looking DWARF.  For example this
code in add_partial_subprogram:

     /* Ignore subprogram DIEs that do not have a name, they are
        illegal.  Do not emit a complaint at this point, we will
        do so when we convert this psymtab into a symtab.  */
     if (pdi->name)
       add_partial_symbol (pdi, cu);

I think the same logic should apply for DW_TAG_entry_point maybe? So,
add_partial_entry_point is possibly redundant.

> +
>  /* Read a partial die corresponding to a subprogram or an inlined
>     subprogram and create a partial symbol for that subprogram.
>     When the CU language allows it, this routine also defines a partial
> @@ -9158,6 +9199,16 @@ add_partial_subprogram (struct partial_die_info *pdi,
>  	  pdi = pdi->die_sibling;
>  	}
>      }
> +  else if (cu->language == language_fortran)
> +    {

Like before, I'm not convinced we should tie this to Fortran...

> +      pdi = pdi->die_child;
> +      while (pdi != nullptr)
> +	{
> +	  if (pdi->tag == DW_TAG_entry_point)
> +	    add_partial_entry_point (pdi, lowpc, highpc, set_addrmap, cu);

You can probably just make the add_partial_symbol call directly here
(after a check that pdi->name is not null.

> +	  pdi = pdi->die_sibling;
> +	}
> +    }
>  }
>  
>  /* Read a partial die corresponding to an enumeration type.  */
> @@ -10596,6 +10647,7 @@ process_die (struct die_info *die, struct dwarf2_cu *cu)
>      case DW_TAG_type_unit:
>        read_type_unit_scope (die, cu);
>        break;
> +    case DW_TAG_entry_point:
>      case DW_TAG_subprogram:
>      case DW_TAG_inlined_subroutine:
>        read_func_scope (die, cu);
> @@ -14669,6 +14721,26 @@ dwarf2_get_pc_bounds (struct die_info *die, CORE_ADDR *lowpc,
>    CORE_ADDR high = 0;
>    enum pc_bounds_kind ret;
>  
> +  if (die->tag == DW_TAG_entry_point)
> +    {
> +      /* Entry_point is embedded in an subprogram. Therefore, we can use
> +	 the highpc from it's enveloping subprogram and get the
> +	 lowpc from DWARF.  */
> +      if (0 == dwarf2_get_pc_bounds (die->parent, lowpc, highpc, cu, pst))

Compare to 'PC_BOUNDS_NOT_PRESENT' not 0.

> +	return PC_BOUNDS_NOT_PRESENT;
> +
> +      attr = dwarf2_attr (die, DW_AT_low_pc, cu);
> +      if (!attr)
> +	{
> +	  complaint (_("DW_TAG_entry_point is missing DW_AT_low_pc"));
> +	  return PC_BOUNDS_NOT_PRESENT;
> +	}
> +      low = attr_value_as_address (attr);
> +      *lowpc = low;
> +
> +      return PC_BOUNDS_HIGH_LOW;
> +    }
> +
>    attr_high = dwarf2_attr (die, DW_AT_high_pc, cu);
>    if (attr_high)
>      {
> @@ -18399,6 +18471,7 @@ load_partial_dies (const struct die_reader_specs *reader,
>  	  && !is_type_tag_for_partial (abbrev->tag)
>  	  && abbrev->tag != DW_TAG_constant
>  	  && abbrev->tag != DW_TAG_enumerator
> +	  && abbrev->tag != DW_TAG_entry_point
>  	  && abbrev->tag != DW_TAG_subprogram
>  	  && abbrev->tag != DW_TAG_inlined_subroutine
>  	  && abbrev->tag != DW_TAG_lexical_block
> @@ -18529,6 +18602,7 @@ load_partial_dies (const struct die_reader_specs *reader,
>  
>        if (load_all
>  	  || abbrev->tag == DW_TAG_constant
> +	  || abbrev->tag == DW_TAG_entry_point
>  	  || abbrev->tag == DW_TAG_subprogram
>  	  || abbrev->tag == DW_TAG_variable
>  	  || abbrev->tag == DW_TAG_namespace
> @@ -18570,7 +18644,9 @@ load_partial_dies (const struct die_reader_specs *reader,
>  		      || last_die->tag == DW_TAG_union_type))
>  	      || (cu->language == language_ada
>  		  && (last_die->tag == DW_TAG_subprogram
> -		      || last_die->tag == DW_TAG_lexical_block))))
> +		      || last_die->tag == DW_TAG_lexical_block))
> +	      || (cu->language == language_fortran
> +		  && last_die->tag == DW_TAG_subprogram)))
>  	{
>  	  nesting_level++;
>  	  parent_die = last_die;
> @@ -21442,6 +21518,25 @@ new_symbol (struct die_info *die, struct type *type, struct dwarf2_cu *cu,
>  	  SYMBOL_ACLASS_INDEX (sym) = LOC_LABEL;
>  	  dw2_add_symbol_to_list (sym, cu->list_in_scope);
>  	  break;
> +	case DW_TAG_entry_point:
> +	  /* Don't know any other language than fortran which is
> +	     using DW_TAG_entry_point.  */
> +	  if (cu->language == language_fortran)
> +	    {
> +	      /* SYMBOL_BLOCK_VALUE (sym) will be filled in later by
> +		 finish_block.  */
> +	      SYMBOL_ACLASS_INDEX (sym) = LOC_BLOCK;
> +	      /* DW_TAG_entry_point provides an additional entry_point to an
> +		 existing sub_program. Therefore, we inherit the "external"
> +		 attribute from the sub_program to which the entry_point
> +		 belongs to.  */
> +	      attr2 = dwarf2_attr (die->parent, DW_AT_external, cu);
> +	      if (attr2 && (DW_UNSND (attr2) != 0))
> +		list_to_add = cu->builder->get_global_symbols ();
> +	      else
> +		list_to_add = cu->list_in_scope;
> +	    }
> +	  break;
>  	case DW_TAG_subprogram:
>  	  /* SYMBOL_BLOCK_VALUE (sym) will be filled in later by
>  	     finish_block.  */
> @@ -22129,6 +22224,7 @@ read_type_die_1 (struct die_info *die, struct dwarf2_cu *cu)
>      case DW_TAG_enumeration_type:
>        this_type = read_enumeration_type (die, cu);
>        break;
> +    case DW_TAG_entry_point:
>      case DW_TAG_subprogram:
>      case DW_TAG_subroutine_type:
>      case DW_TAG_inlined_subroutine:
> diff --git a/gdb/testsuite/gdb.fortran/entry-point.exp b/gdb/testsuite/gdb.fortran/entry-point.exp
> new file mode 100644
> index 000000000000..a1f3f2bebdb9
> --- /dev/null
> +++ b/gdb/testsuite/gdb.fortran/entry-point.exp
> @@ -0,0 +1,70 @@
> +# Copyright 2018 Free Software Foundation, Inc.
> +
> +# This program is free software; you can redistribute it and/or modify
> +# it under the terms of the GNU General Public License as published by
> +# the Free Software Foundation; either version 3 of the License, or
> +# (at your option) any later version.
> +#
> +# This program is distributed in the hope that it will be useful,
> +# but WITHOUT ANY WARRANTY; without even the implied warranty of
> +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
> +# GNU General Public License for more details.
> +#
> +# You should have received a copy of the GNU General Public License
> +# along with this program.  If not, see <http://www.gnu.org/licenses/>.
> +
> +if { [skip_fortran_tests] } { return -1 }
> +
> +standard_testfile .f90
> +load_lib "fortran.exp"
> +
> +if {[prepare_for_testing $testfile.exp $testfile $srcfile {debug f90}]} {
> +    return -1
> +}
> +
> +if ![runto MAIN__] then {
> +    perror "couldn't run to breakpoint MAIN__"
> +    continue
> +}
> +
> +# Test if we can set a breakpoint via entry-point name
> +set ept_name "foo"
> +gdb_breakpoint $ept_name
> +gdb_test "continue" \
> +    [multi_line "Breakpoint $decimal, $ept_name \\(j=1, k=2, l=3, i1=4\\) at .*" \
> +                ".*"] \
> +    "continue to breakpoint: $ept_name"
> +
> +gdb_test "print j" "= 1" "print j, entered via $ept_name"
> +gdb_test "print k" "= 2" "print k, entered via $ept_name"
> +gdb_test "print l" "= 3" "print l, entered via $ept_name"
> +gdb_test "print i1" "= 4" "print i1, entered via $ept_name"
> +gdb_test "info args" \
> +  [multi_line "j = 1" \
> +              "k = 2" \
> +              "l = 3" \
> +              "i1 = 4"] \
> +   "info args, entered via $ept_name"
> +
> +# Test if we can set a breakpoint via function name
> +set ept_name "bar"
> +gdb_breakpoint $ept_name
> +gdb_test "continue" \
> +    [multi_line "Breakpoint $decimal, $ept_name \\(i=4, j=5, k=6, i1=7\\) at .*" \
> +                ".*"] \
> +    "continue to breakpoint: $ept_name"
> +
> +gdb_test "print i" "= 4" "print i, entered via $ept_name"
> +gdb_test "print j" "= 5" "print j, entered via $ept_name"
> +gdb_test "print k" "= 6" "print k, entered via $ept_name"
> +gdb_test "print i1" "= 7" "print i1, entered via $ept_name"
> +
> +set ept_name "tim"
> +gdb_breakpoint $ept_name
> +gdb_test "continue" \
> +    [multi_line "Breakpoint $decimal, $ept_name \\(j=1\\) at .*" \
> +                ".*"] \
> +    "continue to breakpoint: $ept_name"
> +
> +gdb_test "print j" "= 1" "print j, entered via $ept_name"
> +gdb_test "info args" "j = 1" "info args, entered via $ept_name"
> diff --git a/gdb/testsuite/gdb.fortran/entry-point.f90 b/gdb/testsuite/gdb.fortran/entry-point.f90
> new file mode 100644
> index 000000000000..cb663b956982
> --- /dev/null
> +++ b/gdb/testsuite/gdb.fortran/entry-point.f90
> @@ -0,0 +1,48 @@
> +! Copyright 2018 Free Software Foundation, Inc.
> +!
> +! This program is free software; you can redistribute it and/or modify
> +! it under the terms of the GNU General Public License as published by
> +! the Free Software Foundation; either version 3 of the License, or
> +! (at your option) any later version.
> +!
> +! This program is distributed in the hope that it will be useful,
> +! but WITHOUT ANY WARRANTY; without even the implied warranty of
> +! MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
> +! GNU General Public License for more details.
> +!
> +! You should have received a copy of the GNU General Public License
> +! along with this program.  If not, see <http://www.gnu.org/licenses/>.
> +
> +program TestEntryPoint
> +
> +  call foo(1,2,3,4)
> +  call bar(4,5,6,7)
> +  call tim(1)
> +
> +end program TestEntryPoint
> +
> +  subroutine bar(I,J,K,I1)
> +    INTEGER I,J,K,L,I1
> +    INTEGER A
> +    REAL    C
> +
> +    A = 0
> +    C = 0.0
> +
> +    A = I + K + I1
> +    goto 1000
> +
> +    entry foo(J,K,L,I1)
> +    A = J + K + L + I1
> +
> +200 C = J
> +    goto 1000
> +
> +    entry tim(J)
> +    goto 200
> +
> +1000 A = C + 1
> +     C = J * 1.5
> +
> +    return
> +  end subroutine
> -- 
> 2.7.4
> 


Thanks,
Andrew

Re: [PATCH v2 5/7] Fortran: Enable setting breakpoint on nested functions.

[Richard Bunt]

Hi Pawel,

Thanks for this fix, I have a local patch for this so I'd be pleased to see an upstream variant.

Note: I cannot approve GDB patches but I thought I'd contribute a review since I was in this part
of the code recently.

On 11/19/18 9:38 PM, Pawel Wodkowski wrote:
> +# Test if we can set a breakpoint in a nested function
> +gdb_breakpoint "sub_nested_outer"
> +gdb_continue_to_breakpoint "sub_nested_outer" ".*local_int = 19"
>  
>  # Test if we can access local and
>  # non-local variables defined one level up.
> @@ -43,6 +46,10 @@ gdb_test "print local_int" "= 19" "print local_int in outer function"
>  gdb_test "up"
>  gdb_test "print index" "= 42" "print index at BP1, one frame up"
>  
> +# Test if we can set a breakpoint in a nested function
> +gdb_breakpoint "sub_nested_inner"
> +gdb_continue_to_breakpoint "sub_nested_inner" ".*local_int = 17"
> +
>  # Test if we can access local and
>  # non-local variables defined two level up.
>  gdb_breakpoint [gdb_get_line_number "! BP_inner"]
> 

This patch passed my local test case for this fix, so it looks good to me. The only test case that
I have locally that I think would be of value here, would be to test that breakpoints can be set on 
multiple functions in the same contains block (i.e. both sub_nested_outer and sub_nested_inner),
prior to program start. I found that such a test gives coverage to the changes in 
add_partial_subprogram as it tests the logic for subprograms which are linked as siblings.

What do you think?

>            /* brobecker/2007-12-26: Normally, only "external" DIEs are part
>               of the global scope.  But in Ada, we want to be able to access
> @@ -9206,6 +9208,8 @@ add_partial_subprogram (struct partial_die_info *pdi,
>  	{
>  	  if (pdi->tag == DW_TAG_entry_point)
>  	    add_partial_entry_point (pdi, lowpc, highpc, set_addrmap, cu);
> +	  else if (pdi->tag == DW_TAG_subprogram)
> +	    add_partial_subprogram (pdi, lowpc, highpc, set_addrmap, cu);
>  	  pdi = pdi->die_sibling;
>  	}
>      }


Many thanks,

Rich

Re: [PATCH v2 1/7] DWARF: Don't add nameless modules to partial symbol table.

[Paweł Wódkowski]

Thanks for review. See comments.

On 27.11.2018 22:41, Andrew Burgess wrote:
> * Pawel Wodkowski <pwodkowski@pl.sii.eu> [2018-11-19 22:38:44 +0100]:
> 
>> From: Bernhard Heckel <bernhard.heckel@intel.com>
>>
>> A name for BLOCK DATA in Fortran is optional. If no
>> name has been assigned, GDB will crash during read-in of DWARF
>> when BLOCK DATA is represented via DW_TAG_module.
>> BLOCK DATA is used for one-time initialization of
>> non-pointer variables in named common blocks.
> 
> Thanks for this.
> 
> Locally I have:
> 
>    GNU Fortran (GCC) 7.3.1 20180712 (Red Hat 7.3.1-6)
> 
> For me at least this test passes with and without the change to GDB.
> 
> Can you confirm which Fortran compiler you see the issue with?  It
> probably should be documented in the commit message and in comment
> around the check in dwarf2read.c.  It could also be useful to include
> examples of the problem DWARF in the commit message in cases like
> this.
> 

Indeed, I checked this for GNU fortran it there is no issue. But for 
Intel fortran compiler ifort I'm getting crash in add_partial_symbol 
function:

$ ifort block-data.f -g -o block-data
$ gdb --args ../../gdb ./block-data
<< ... >>
GNU gdb (GDB) 8.2.50.20181114-git
Copyright (C) 2018 Free Software Foundation, Inc.
License GPLv3+: GNU GPL version 3 or later 
<http://gnu.org/licenses/gpl.html>
This is free software: you are free to change and redistribute it.
<< ... >>
Program received signal SIGSEGV, Segmentation fault.
__strlen_sse2 () at ../sysdeps/x86_64/multiarch/../strlen.S:120
120     ../sysdeps/x86_64/multiarch/../strlen.S: No such file or directory.
(gdb) bt
#0  __strlen_sse2 () at ../sysdeps/x86_64/multiarch/../strlen.S:120
#1  0x000055555575dee9 in add_partial_symbol 
(pdi=pdi@entry=0x5555562ca490, cu=cu@entry=0x5555562c34b0) at 
dwarf2read.c:8998
#2  0x000055555575e59b in add_partial_namespace (cu=0x5555562c34b0, 
set_addrmap=0, highpc=0x7fffffffda90, lowpc=0x7fffffffda88, 
pdi=0x5555562ca490) at dwarf2read.c:9058
#3  scan_partial_symbols (first_die=<optimized out>, 
lowpc=lowpc@entry=0x7fffffffda88, highpc=highpc@entry=0x7fffffffda90, 
set_addrmap=0, cu=cu@entry=0x5555562c34b0) at dwarf2read.c:8682
#4  0x000055555575ee1e in process_psymtab_comp_unit_reader 
(reader=reader@entry=0x7fffffffdb50, info_ptr=0x55555617ddb3 "\002\025", 
<incomplete sequence \310>, comp_unit_die=0x5555562ca0f0,
     has_children=<optimized out>, data=data@entry=0x7fffffffdbf0) at 
dwarf2read.c:8033
<< ... >>

>>
>> 2016-06-15  Bernhard Heckel  <bernhard.heckel@intel.com>
>>
>> gdb/Changelog:
>> 	* dwarf2read.c (add_partial_symbol): Skip nameless modules.
>>
>> gdb/Testsuite/Changelog:
> 
> It's 'testsuite' all the patches in this series :)
> 
>> 	* gdb.fortran/block-data.f: New.
>> 	* gdb.fortran/block-data.exp: New.
>> ---
>>   gdb/dwarf2read.c                         | 13 ++++---
>>   gdb/testsuite/gdb.fortran/block-data.exp | 51 ++++++++++++++++++++++++++++
>>   gdb/testsuite/gdb.fortran/block-data.f   | 58 ++++++++++++++++++++++++++++++++
>>   3 files changed, 117 insertions(+), 5 deletions(-)
>>   create mode 100644 gdb/testsuite/gdb.fortran/block-data.exp
>>   create mode 100644 gdb/testsuite/gdb.fortran/block-data.f
>>
>> diff --git a/gdb/dwarf2read.c b/gdb/dwarf2read.c
>> index d2a8cd44f9a5..89fd4ae15e80 100644
>> --- a/gdb/dwarf2read.c
>> +++ b/gdb/dwarf2read.c
>> @@ -8995,11 +8995,14 @@ add_partial_symbol (struct partial_die_info *pdi, struct dwarf2_cu *cu)
>>   			   0, cu->language, objfile);
>>         break;
>>       case DW_TAG_module:
>> -      add_psymbol_to_list (actual_name, strlen (actual_name),
>> -			   built_actual_name != NULL,
>> -			   MODULE_DOMAIN, LOC_TYPEDEF, -1,
>> -			   &objfile->global_psymbols,
>> -			   0, cu->language, objfile);
>> +      /* In Fortran 77 there might be a "BLOCK DATA" module available wihout
>> +         any name. If so, we skip the module as it doesn't bring any value */
> 
> You need a full-stop and two spaces at the end, like "...any value.  */".
> 

Ok, will fix.

>> +      if (actual_name != nullptr)
>> +	add_psymbol_to_list (actual_name, strlen (actual_name),
>> +			     built_actual_name != nullptr,
>> +			     MODULE_DOMAIN, LOC_TYPEDEF, -1,
>> +			     &objfile->global_psymbols,
>> +			     0, cu->language, objfile);
>>         break;
>>       case DW_TAG_class_type:
>>       case DW_TAG_interface_type:
>> diff --git a/gdb/testsuite/gdb.fortran/block-data.exp b/gdb/testsuite/gdb.fortran/block-data.exp
>> new file mode 100644
>> index 000000000000..2af250ad3886
>> --- /dev/null
>> +++ b/gdb/testsuite/gdb.fortran/block-data.exp
>> @@ -0,0 +1,51 @@
>> +# Copyright 2018 Free Software Foundation, Inc.
>> +#
>> +# Contributed by Intel Corp. <bernhard.heckel@intel.com>
> 
> Someone else might correct me, but I think "Contributed by..." lines
> are not supposed to be added to GDB source..
> 

Will remove it.

>> +#
>> +# This program is free software; you can redistribute it and/or modify
>> +# it under the terms of the GNU General Public License as published by
>> +# the Free Software Foundation; either version 3 of the License, or
>> +# (at your option) any later version.
>> +#
>> +# This program is distributed in the hope that it will be useful,
>> +# but WITHOUT ANY WARRANTY; without even the implied warranty of
>> +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
>> +# GNU General Public License for more details.
>> +#
>> +# You should have received a copy of the GNU General Public License
>> +# along with this program.  If not, see <http://www.gnu.org/licenses/>.
>> +
>> +# This test is supposed to test anonymous block-data statement.
>> +
>> +if { [skip_fortran_tests] } { return -1 }
>> +
>> +standard_testfile .f
>> +load_lib "fortran.exp"
>> +
>> +if {[prepare_for_testing $testfile.exp $testfile $srcfile {debug f90}]} {
>> +    return -1
>> +}
>> +
>> +if ![runto MAIN__] then {
>> +    untested "couldn't run to breakpoint MAIN__"
>> +    return -1
>> +}
>> +
>> +gdb_test "print doub1" "= 1.11\\d+" "print doub1, default values"
>> +gdb_test "print doub2" "= 2.22\\d+" "print doub2, default values"
>> +gdb_test "print char1" "= 'abcdef'" "print char1, default values"
>> +gdb_test "print char2" "= 'ghijkl'" "print char2, default values"
>> +
>> +gdb_breakpoint [gdb_get_line_number "! BP_BEFORE_SUB"]
>> +gdb_continue_to_breakpoint "! BP_BEFORE_SUB" ".*! BP_BEFORE_SUB.*"
>> +gdb_test "print doub1" "= 11.11\\d+" "print doub1, before sub"
>> +gdb_test "print doub2" "= 22.22\\d+" "print doub2, before sub"
>> +gdb_test "print char1" "= 'ABCDEF'" "print char1, before sub"
>> +gdb_test "print char2" "= 'GHIJKL'" "print char2, before sub"
>> +
>> +gdb_breakpoint [gdb_get_line_number "! BP_SUB"]
>> +gdb_continue_to_breakpoint "! BP_SUB" ".*! BP_SUB.*"
>> +gdb_test "print doub1" "= 11.11\\d+" "print char1, in sub"
>> +gdb_test "print doub2" "= 22.22\\d+" "print doub2, in sub"
>> +gdb_test "print char1" "= 'ABCDEF'" "print char1, in sub"
>> +gdb_test "print char2" "= 'GHIJKL'" "print char2, in sub"
>> diff --git a/gdb/testsuite/gdb.fortran/block-data.f b/gdb/testsuite/gdb.fortran/block-data.f
>> new file mode 100644
>> index 000000000000..a28e687ec885
>> --- /dev/null
>> +++ b/gdb/testsuite/gdb.fortran/block-data.f
>> @@ -0,0 +1,58 @@
>> +! Copyright 2018 Free Software Foundation, Inc.
>> +!
>> +! Contributed by Intel Corp. <bernhard.heckel@intel.com>
>> +!
>> +! This program is free software; you can redistribute it and/or modify
>> +! it under the terms of the GNU General Public License as published by
>> +! the Free Software Foundation; either version 3 of the License, or
>> +! (at your option) any later version.
>> +!
>> +! This program is distributed in the hope that it will be useful,
>> +! but WITHOUT ANY WARRANTY; without even the implied warranty of
>> +! MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
>> +! GNU General Public License for more details.
>> +!
>> +! You should have received a copy of the GNU General Public License
>> +! along with this program.  If not, see <http://www.gnu.org/licenses/>.
>> +!
>> +! Test if GDB can handle block data without global name
>> +!
>> +! MAIN
>> +        PROGRAM bdata
>> +        DOUBLE PRECISION doub1, doub2
>> +        CHARACTER*6 char1, char2
>> +
>> +        COMMON /BLK1/ doub1, char1
>> +        COMMON /BLK2/ doub2, char2
>> +
>> +        doub1 = 11.111
>> +        doub2 = 22.222
>> +        char1 = 'ABCDEF'
>> +        char2 = 'GHIJKL'
>> +        CALL sub_block_data      ! BP_BEFORE_SUB
>> +        STOP
>> +        END
>> +
>> +! BLOCK DATA
>> +        BLOCK DATA
>> +
>> +        DOUBLE PRECISION doub1, doub2
>> +        CHARACTER*6 char1, char2
>> +
>> +        COMMON /BLK1/ doub1, char1
>> +        COMMON /BLK2/ doub2, char2
>> +        DATA doub1, doub2 /1.111, 2.222/
>> +        DATA char1, char2 /'abcdef', 'ghijkl'/
>> +        END
>> +
>> +! SUBROUTINE
>> +        SUBROUTINE sub_block_data
>> +
>> +        DOUBLE PRECISION doub1, doub2
>> +        CHARACTER*6 char1, char2
>> +
>> +        COMMON /BLK1/ doub1, char1
>> +        COMMON /BLK2/ doub2, char2
>> +
>> +        char1 = char2;    ! BP_SUB
>> +        END
>> -- 
>> 2.7.4
>>
> 

Re: [PATCH v2 2/7] Dwarf: Fortran, support DW_TAG_entry_point.

[Paweł Wódkowski]

On 27.11.2018 23:08, Andrew Burgess wrote:
> * Pawel Wodkowski <pwodkowski@pl.sii.eu> [2018-11-19 22:38:45 +0100]:
> 
>> From: Bernhard Heckel <bernhard.heckel@intel.com>
>>
>> Fortran provides additional entry-points to an subprogram.
>> Those entry-points may have only a subset of parameters
>> of the original subprogram as well.
>> Add support for parsing DW_TAG_entry_point's for Fortran.
> 
> As with patch #1 it's probably worth documenting which compilers you
> see DW_TAG_entry_point in, along with an example of what the generated
> DWARF looks like.
> 
> This means that in the future if/when others start to use this DWARF
> feature we can figure out which other compilers we need to test.
> 

As I checked gfortran will not generate entry point tags so for now it 
is only for Intel ifort compiler. The entry-point.f90 example compiled 
using ifort produce following entry point tag:

...
<2><9e>: Abbrev Number: 5 (DW_TAG_entry_point)
     <9f>   DW_AT_decl_line   : 35
     <a0>   DW_AT_decl_file   : 1
     <a1>   DW_AT_name        : foo
     <a5>   DW_AT_low_pc      : 0x402ba8
...
<3><c9>: Abbrev Number: 4 (DW_TAG_formal_parameter)
     <ca>   DW_AT_decl_line   : 35
     <cb>   DW_AT_decl_file   : 1
     <cc>   DW_AT_type        : <0x120>
     <d0>   DW_AT_name        : l
     <d2>   DW_AT_location    : 4 byte block: 76 f4 7e 6 
(DW_OP_breg6 (rbp): -140; DW_OP_deref)
...

I will add this to commit message.

Thanks
Pawel

>>
>> 2016-06-01  Bernhard Heckel  <bernhard.heckel@intel.com>
>>
>> gdb/Changelog:
>> 	* gdb/dwarf2read.c (add_partial_symbol): Handle DW_TAG_entry_point.
>> 	(add_partial_entry_point): New.
>> 	(add_partial_subprogram): Search for entry_points.
>> 	(process_die): Handle DW_TAG_entry_point.
>> 	(dwarf2_get_pc_bounds): Update low pc from DWARF.
>> 	(load_partial_dies): Save DW_TAG_entry_point's.
>> 	(load_partial_dies): Save DW_TAG_entry_point to hash table.
>> 	(load_partial_dies): Look into child's of DW_TAG_sub_program
>> 	for fortran.
>> 	(new_symbol_full): Process DW_TAG_entry_point.
>> 	(read_type_die_1): Handle DW_TAG_entry_point.
>>
>> gdb/Testsuite/Changelog:
>> 	* gdb.fortran/entry_point.f90: New.
>> 	* gdb.fortran/entry_point.exp: New.
>> ---
>>   gdb/dwarf2read.c                          | 98 ++++++++++++++++++++++++++++++-
>>   gdb/testsuite/gdb.fortran/entry-point.exp | 70 ++++++++++++++++++++++
>>   gdb/testsuite/gdb.fortran/entry-point.f90 | 48 +++++++++++++++
>>   3 files changed, 215 insertions(+), 1 deletion(-)
>>   create mode 100644 gdb/testsuite/gdb.fortran/entry-point.exp
>>   create mode 100644 gdb/testsuite/gdb.fortran/entry-point.f90
>>
>> diff --git a/gdb/dwarf2read.c b/gdb/dwarf2read.c
>> index 89fd4ae15e80..88e57d7ab68e 100644
>> --- a/gdb/dwarf2read.c
>> +++ b/gdb/dwarf2read.c
>> @@ -1473,6 +1473,10 @@ static void add_partial_module (struct partial_die_info *pdi, CORE_ADDR *lowpc,
>>   static void add_partial_enumeration (struct partial_die_info *enum_pdi,
>>   				     struct dwarf2_cu *cu);
>>   
>> +static void add_partial_entry_point (struct partial_die_info *pdi,
>> +				     CORE_ADDR *lowpc, CORE_ADDR *highpc,
>> +				     int need_pc, struct dwarf2_cu *cu);
>> +
>>   static void add_partial_subprogram (struct partial_die_info *pdi,
>>   				    CORE_ADDR *lowpc, CORE_ADDR *highpc,
>>   				    int need_pc, struct dwarf2_cu *cu);
>> @@ -8876,6 +8880,32 @@ add_partial_symbol (struct partial_die_info *pdi, struct dwarf2_cu *cu)
>>   
>>     switch (pdi->tag)
>>       {
>> +    case DW_TAG_entry_point:
>> +      /* Don't know any other language than fortran which is
>> +	 using DW_TAG_entry_point.  */
>> +      if (cu->language == language_fortran)
>> +	{
> 
> I'm not sure the language check is needed here.  The description for
> DW_TAG_entry_point in the DWARF spec seems pretty generic.  I don't
> see how a different language would expect significantly different
> results, and so, I would suggest we should just add this as general
> purpose code.
> 
>> +	  addr = gdbarch_adjust_dwarf2_addr (gdbarch, pdi->lowpc + baseaddr);
> 
> When compared to DW_TAG_subprogram, this line doesn't include a '-
> baseaddr'.  I think that deserves a comment explaining why.
> 
>> +	  /* DW_TAG_entry_point provides an additional entry_point to an
>> +	     existing sub_program. Therefore, we inherit the "external"
>> +	     attribute from the sub_program to which the entry_point
>> +	     belongs to.  */
>> +	  if (pdi->die_parent->is_external)
>> +	    add_psymbol_to_list (actual_name, strlen (actual_name),
>> +				 built_actual_name != nullptr,
>> +				 VAR_DOMAIN, LOC_BLOCK,
>> +				 SECT_OFF_TEXT (objfile),
>> +				 &objfile->global_psymbols,
>> +				 addr, cu->language, objfile);
>> +	  else
>> +	    add_psymbol_to_list (actual_name, strlen (actual_name),
>> +				 built_actual_name != nullptr,
>> +				 VAR_DOMAIN, LOC_BLOCK,
>> +				 SECT_OFF_TEXT (objfile),
>> +				 &objfile->static_psymbols,
>> +				 addr, cu->language, objfile);
> 
> I think these two add_psymbol_to_list calls, and the two for
> DW_TAG_subprogram should be factored out.  Just my opinion though...
> 
>> +	}
>> +      break;
>>       case DW_TAG_inlined_subroutine:
>>       case DW_TAG_subprogram:
>>         addr = (gdbarch_adjust_dwarf2_addr (gdbarch, pdi->lowpc + baseaddr)
>> @@ -9082,6 +9112,17 @@ add_partial_module (struct partial_die_info *pdi, CORE_ADDR *lowpc,
>>       scan_partial_symbols (pdi->die_child, lowpc, highpc, set_addrmap, cu);
>>   }
>>   
>> +static void
>> +add_partial_entry_point (struct partial_die_info *pdi,
>> +			 CORE_ADDR *p_lowpc, CORE_ADDR *p_highpc,
>> +			 int set_addrmap, struct dwarf2_cu *cu)
>> +{
>> +  if (pdi->name == nullptr)
>> +    complaint (_("DW_TAG_entry_point must have a name"));
>> +  else
>> +    add_partial_symbol (pdi, cu);
>> +}
> 
> This function should have a header comment, however, a quick look at
> add_partial_subprogram seems to indicate that at this point GDB is not
> complaining, but just ignoring weird looking DWARF.  For example this
> code in add_partial_subprogram:
> 
>       /* Ignore subprogram DIEs that do not have a name, they are
>          illegal.  Do not emit a complaint at this point, we will
>          do so when we convert this psymtab into a symtab.  */
>       if (pdi->name)
>         add_partial_symbol (pdi, cu);
> 
> I think the same logic should apply for DW_TAG_entry_point maybe? So,
> add_partial_entry_point is possibly redundant.
> 
>> +
>>   /* Read a partial die corresponding to a subprogram or an inlined
>>      subprogram and create a partial symbol for that subprogram.
>>      When the CU language allows it, this routine also defines a partial
>> @@ -9158,6 +9199,16 @@ add_partial_subprogram (struct partial_die_info *pdi,
>>   	  pdi = pdi->die_sibling;
>>   	}
>>       }
>> +  else if (cu->language == language_fortran)
>> +    {
> 
> Like before, I'm not convinced we should tie this to Fortran...
> 
>> +      pdi = pdi->die_child;
>> +      while (pdi != nullptr)
>> +	{
>> +	  if (pdi->tag == DW_TAG_entry_point)
>> +	    add_partial_entry_point (pdi, lowpc, highpc, set_addrmap, cu);
> 
> You can probably just make the add_partial_symbol call directly here
> (after a check that pdi->name is not null.
> 
>> +	  pdi = pdi->die_sibling;
>> +	}
>> +    }
>>   }
>>   
>>   /* Read a partial die corresponding to an enumeration type.  */
>> @@ -10596,6 +10647,7 @@ process_die (struct die_info *die, struct dwarf2_cu *cu)
>>       case DW_TAG_type_unit:
>>         read_type_unit_scope (die, cu);
>>         break;
>> +    case DW_TAG_entry_point:
>>       case DW_TAG_subprogram:
>>       case DW_TAG_inlined_subroutine:
>>         read_func_scope (die, cu);
>> @@ -14669,6 +14721,26 @@ dwarf2_get_pc_bounds (struct die_info *die, CORE_ADDR *lowpc,
>>     CORE_ADDR high = 0;
>>     enum pc_bounds_kind ret;
>>   
>> +  if (die->tag == DW_TAG_entry_point)
>> +    {
>> +      /* Entry_point is embedded in an subprogram. Therefore, we can use
>> +	 the highpc from it's enveloping subprogram and get the
>> +	 lowpc from DWARF.  */
>> +      if (0 == dwarf2_get_pc_bounds (die->parent, lowpc, highpc, cu, pst))
> 
> Compare to 'PC_BOUNDS_NOT_PRESENT' not 0.
> 
>> +	return PC_BOUNDS_NOT_PRESENT;
>> +
>> +      attr = dwarf2_attr (die, DW_AT_low_pc, cu);
>> +      if (!attr)
>> +	{
>> +	  complaint (_("DW_TAG_entry_point is missing DW_AT_low_pc"));
>> +	  return PC_BOUNDS_NOT_PRESENT;
>> +	}
>> +      low = attr_value_as_address (attr);
>> +      *lowpc = low;
>> +
>> +      return PC_BOUNDS_HIGH_LOW;
>> +    }
>> +
>>     attr_high = dwarf2_attr (die, DW_AT_high_pc, cu);
>>     if (attr_high)
>>       {
>> @@ -18399,6 +18471,7 @@ load_partial_dies (const struct die_reader_specs *reader,
>>   	  && !is_type_tag_for_partial (abbrev->tag)
>>   	  && abbrev->tag != DW_TAG_constant
>>   	  && abbrev->tag != DW_TAG_enumerator
>> +	  && abbrev->tag != DW_TAG_entry_point
>>   	  && abbrev->tag != DW_TAG_subprogram
>>   	  && abbrev->tag != DW_TAG_inlined_subroutine
>>   	  && abbrev->tag != DW_TAG_lexical_block
>> @@ -18529,6 +18602,7 @@ load_partial_dies (const struct die_reader_specs *reader,
>>   
>>         if (load_all
>>   	  || abbrev->tag == DW_TAG_constant
>> +	  || abbrev->tag == DW_TAG_entry_point
>>   	  || abbrev->tag == DW_TAG_subprogram
>>   	  || abbrev->tag == DW_TAG_variable
>>   	  || abbrev->tag == DW_TAG_namespace
>> @@ -18570,7 +18644,9 @@ load_partial_dies (const struct die_reader_specs *reader,
>>   		      || last_die->tag == DW_TAG_union_type))
>>   	      || (cu->language == language_ada
>>   		  && (last_die->tag == DW_TAG_subprogram
>> -		      || last_die->tag == DW_TAG_lexical_block))))
>> +		      || last_die->tag == DW_TAG_lexical_block))
>> +	      || (cu->language == language_fortran
>> +		  && last_die->tag == DW_TAG_subprogram)))
>>   	{
>>   	  nesting_level++;
>>   	  parent_die = last_die;
>> @@ -21442,6 +21518,25 @@ new_symbol (struct die_info *die, struct type *type, struct dwarf2_cu *cu,
>>   	  SYMBOL_ACLASS_INDEX (sym) = LOC_LABEL;
>>   	  dw2_add_symbol_to_list (sym, cu->list_in_scope);
>>   	  break;
>> +	case DW_TAG_entry_point:
>> +	  /* Don't know any other language than fortran which is
>> +	     using DW_TAG_entry_point.  */
>> +	  if (cu->language == language_fortran)
>> +	    {
>> +	      /* SYMBOL_BLOCK_VALUE (sym) will be filled in later by
>> +		 finish_block.  */
>> +	      SYMBOL_ACLASS_INDEX (sym) = LOC_BLOCK;
>> +	      /* DW_TAG_entry_point provides an additional entry_point to an
>> +		 existing sub_program. Therefore, we inherit the "external"
>> +		 attribute from the sub_program to which the entry_point
>> +		 belongs to.  */
>> +	      attr2 = dwarf2_attr (die->parent, DW_AT_external, cu);
>> +	      if (attr2 && (DW_UNSND (attr2) != 0))
>> +		list_to_add = cu->builder->get_global_symbols ();
>> +	      else
>> +		list_to_add = cu->list_in_scope;
>> +	    }
>> +	  break;
>>   	case DW_TAG_subprogram:
>>   	  /* SYMBOL_BLOCK_VALUE (sym) will be filled in later by
>>   	     finish_block.  */
>> @@ -22129,6 +22224,7 @@ read_type_die_1 (struct die_info *die, struct dwarf2_cu *cu)
>>       case DW_TAG_enumeration_type:
>>         this_type = read_enumeration_type (die, cu);
>>         break;
>> +    case DW_TAG_entry_point:
>>       case DW_TAG_subprogram:
>>       case DW_TAG_subroutine_type:
>>       case DW_TAG_inlined_subroutine:
>> diff --git a/gdb/testsuite/gdb.fortran/entry-point.exp b/gdb/testsuite/gdb.fortran/entry-point.exp
>> new file mode 100644
>> index 000000000000..a1f3f2bebdb9
>> --- /dev/null
>> +++ b/gdb/testsuite/gdb.fortran/entry-point.exp
>> @@ -0,0 +1,70 @@
>> +# Copyright 2018 Free Software Foundation, Inc.
>> +
>> +# This program is free software; you can redistribute it and/or modify
>> +# it under the terms of the GNU General Public License as published by
>> +# the Free Software Foundation; either version 3 of the License, or
>> +# (at your option) any later version.
>> +#
>> +# This program is distributed in the hope that it will be useful,
>> +# but WITHOUT ANY WARRANTY; without even the implied warranty of
>> +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
>> +# GNU General Public License for more details.
>> +#
>> +# You should have received a copy of the GNU General Public License
>> +# along with this program.  If not, see <http://www.gnu.org/licenses/>.
>> +
>> +if { [skip_fortran_tests] } { return -1 }
>> +
>> +standard_testfile .f90
>> +load_lib "fortran.exp"
>> +
>> +if {[prepare_for_testing $testfile.exp $testfile $srcfile {debug f90}]} {
>> +    return -1
>> +}
>> +
>> +if ![runto MAIN__] then {
>> +    perror "couldn't run to breakpoint MAIN__"
>> +    continue
>> +}
>> +
>> +# Test if we can set a breakpoint via entry-point name
>> +set ept_name "foo"
>> +gdb_breakpoint $ept_name
>> +gdb_test "continue" \
>> +    [multi_line "Breakpoint $decimal, $ept_name \\(j=1, k=2, l=3, i1=4\\) at .*" \
>> +                ".*"] \
>> +    "continue to breakpoint: $ept_name"
>> +
>> +gdb_test "print j" "= 1" "print j, entered via $ept_name"
>> +gdb_test "print k" "= 2" "print k, entered via $ept_name"
>> +gdb_test "print l" "= 3" "print l, entered via $ept_name"
>> +gdb_test "print i1" "= 4" "print i1, entered via $ept_name"
>> +gdb_test "info args" \
>> +  [multi_line "j = 1" \
>> +              "k = 2" \
>> +              "l = 3" \
>> +              "i1 = 4"] \
>> +   "info args, entered via $ept_name"
>> +
>> +# Test if we can set a breakpoint via function name
>> +set ept_name "bar"
>> +gdb_breakpoint $ept_name
>> +gdb_test "continue" \
>> +    [multi_line "Breakpoint $decimal, $ept_name \\(i=4, j=5, k=6, i1=7\\) at .*" \
>> +                ".*"] \
>> +    "continue to breakpoint: $ept_name"
>> +
>> +gdb_test "print i" "= 4" "print i, entered via $ept_name"
>> +gdb_test "print j" "= 5" "print j, entered via $ept_name"
>> +gdb_test "print k" "= 6" "print k, entered via $ept_name"
>> +gdb_test "print i1" "= 7" "print i1, entered via $ept_name"
>> +
>> +set ept_name "tim"
>> +gdb_breakpoint $ept_name
>> +gdb_test "continue" \
>> +    [multi_line "Breakpoint $decimal, $ept_name \\(j=1\\) at .*" \
>> +                ".*"] \
>> +    "continue to breakpoint: $ept_name"
>> +
>> +gdb_test "print j" "= 1" "print j, entered via $ept_name"
>> +gdb_test "info args" "j = 1" "info args, entered via $ept_name"
>> diff --git a/gdb/testsuite/gdb.fortran/entry-point.f90 b/gdb/testsuite/gdb.fortran/entry-point.f90
>> new file mode 100644
>> index 000000000000..cb663b956982
>> --- /dev/null
>> +++ b/gdb/testsuite/gdb.fortran/entry-point.f90
>> @@ -0,0 +1,48 @@
>> +! Copyright 2018 Free Software Foundation, Inc.
>> +!
>> +! This program is free software; you can redistribute it and/or modify
>> +! it under the terms of the GNU General Public License as published by
>> +! the Free Software Foundation; either version 3 of the License, or
>> +! (at your option) any later version.
>> +!
>> +! This program is distributed in the hope that it will be useful,
>> +! but WITHOUT ANY WARRANTY; without even the implied warranty of
>> +! MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
>> +! GNU General Public License for more details.
>> +!
>> +! You should have received a copy of the GNU General Public License
>> +! along with this program.  If not, see <http://www.gnu.org/licenses/>.
>> +
>> +program TestEntryPoint
>> +
>> +  call foo(1,2,3,4)
>> +  call bar(4,5,6,7)
>> +  call tim(1)
>> +
>> +end program TestEntryPoint
>> +
>> +  subroutine bar(I,J,K,I1)
>> +    INTEGER I,J,K,L,I1
>> +    INTEGER A
>> +    REAL    C
>> +
>> +    A = 0
>> +    C = 0.0
>> +
>> +    A = I + K + I1
>> +    goto 1000
>> +
>> +    entry foo(J,K,L,I1)
>> +    A = J + K + L + I1
>> +
>> +200 C = J
>> +    goto 1000
>> +
>> +    entry tim(J)
>> +    goto 200
>> +
>> +1000 A = C + 1
>> +     C = J * 1.5
>> +
>> +    return
>> +  end subroutine
>> -- 
>> 2.7.4
>>
> 
> 
> Thanks,
> Andrew
> 

Re: [PATCH v2 5/7] Fortran: Enable setting breakpoint on nested functions.

[Paweł Wódkowski]

Hi Richard,

Please see my comments.

On 28.11.2018 11:30, Richard Bunt wrote:
> Hi Pawel,
> 
> Thanks for this fix, I have a local patch for this so I'd be pleased to see an upstream variant.
> 
> Note: I cannot approve GDB patches but I thought I'd contribute a review since I was in this part
> of the code recently.
> 

Any review is great, especially from someone else developping for 
fortran these days ;)

> On 11/19/18 9:38 PM, Pawel Wodkowski wrote:
>> +# Test if we can set a breakpoint in a nested function
>> +gdb_breakpoint "sub_nested_outer"
>> +gdb_continue_to_breakpoint "sub_nested_outer" ".*local_int = 19"
>>   
>>   # Test if we can access local and
>>   # non-local variables defined one level up.
>> @@ -43,6 +46,10 @@ gdb_test "print local_int" "= 19" "print local_int in outer function"
>>   gdb_test "up"
>>   gdb_test "print index" "= 42" "print index at BP1, one frame up"
>>   
>> +# Test if we can set a breakpoint in a nested function
>> +gdb_breakpoint "sub_nested_inner"
>> +gdb_continue_to_breakpoint "sub_nested_inner" ".*local_int = 17"
>> +
>>   # Test if we can access local and
>>   # non-local variables defined two level up.
>>   gdb_breakpoint [gdb_get_line_number "! BP_inner"]
>>
> 
> This patch passed my local test case for this fix, so it looks good to me. The only test case that
> I have locally that I think would be of value here, would be to test that breakpoints can be set on
> multiple functions in the same contains block (i.e. both sub_nested_outer and sub_nested_inner),
> prior to program start. I found that such a test gives coverage to the changes in
> add_partial_subprogram as it tests the logic for subprograms which are linked as siblings.
> 
> What do you think?
> 

Yes, setting breakpoint before starting program is something that should 
work for sure so I checked it. I'm happy to say that it is working for 
both functions when set before MAIN__ :)
I will change the tests.


>>             /* brobecker/2007-12-26: Normally, only "external" DIEs are part
>>                of the global scope.  But in Ada, we want to be able to access
>> @@ -9206,6 +9208,8 @@ add_partial_subprogram (struct partial_die_info *pdi,
>>   	{
>>   	  if (pdi->tag == DW_TAG_entry_point)
>>   	    add_partial_entry_point (pdi, lowpc, highpc, set_addrmap, cu);
>> +	  else if (pdi->tag == DW_TAG_subprogram)
>> +	    add_partial_subprogram (pdi, lowpc, highpc, set_addrmap, cu);
>>   	  pdi = pdi->die_sibling;
>>   	}
>>       }
> 
> 
> Many thanks,
> 
> Rich
> 

Pawel

Re: [PATCH v2 3/7] Fortran: Accessing fields of inherited types via fully qualified name.

[Richard Bunt]

Hi Pawel,

Is there a compiler where all the marked kfails pass? I've given this patch a spin
with the Fortran compilers from various vendors, but there is a varying number of kfails.

I'd like to test everything if possible.

Many thanks,

Rich

On 11/19/18 9:38 PM, Pawel Wodkowski wrote:
> From: Bernhard Heckel <bernhard.heckel@intel.com>
> 
> Fortran 2003 supports type extension. This patch allows access
> to inherited members by using it's fully qualified name as
> described in the fortran standard.
> 
> Before:
> (gdb) print my_extended_obj%base_class_name%member_base
> Syntax error near base_class_name%member_base
> 
> (gdb) print my_extended_obj%member_base
> $1 = (10, 10, 10)
> 
> After:
> (gdb) print my_extended_obj%base_clase_name%member_base
> $1 = (10, 10, 10)
> 
> (gdb) print my_extended_obj%member_base
> $1 = (10, 10, 10)
> 
> 2016-04-22  Bernhard Heckel  <bernhard.heckel@intel.com>
> 
> gdb/Changelog:
> 	* f-exp.y (name): Allow TYPENAME.
> 	* valops.c (search_struct_method): Look also for baseclass.
> 
> gdb/Testsuite/Changelog:
> 	* gdb.fortran/oop-extend-type.f90: New.
> 	* gdb.fortran/oop-extend-type.exp: New.
> ---
>  gdb/f-exp.y                                   |   7 +-
>  gdb/testsuite/gdb.fortran/oop-extend-type.exp | 113 ++++++++++++++++++++++++++
>  gdb/testsuite/gdb.fortran/oop-extend-type.f90 |  56 +++++++++++++
>  gdb/valops.c                                  |   6 ++
>  4 files changed, 180 insertions(+), 2 deletions(-)
>  create mode 100644 gdb/testsuite/gdb.fortran/oop-extend-type.exp
>  create mode 100644 gdb/testsuite/gdb.fortran/oop-extend-type.f90
> 
> diff --git a/gdb/f-exp.y b/gdb/f-exp.y
> index 390bd45081b7..4c2e101699ac 100644
> --- a/gdb/f-exp.y
> +++ b/gdb/f-exp.y
> @@ -604,8 +604,11 @@ nonempty_typelist
>  		}
>  	;
>  
> -name	:	NAME
> -		{  $$ = $1.stoken; }
> +name
> +	:	NAME
> +		{ $$ = $1.stoken; }
> +	|	TYPENAME
> +		{ $$ = $1.stoken; }
>  	;
>  
>  name_not_typename :	NAME
> diff --git a/gdb/testsuite/gdb.fortran/oop-extend-type.exp b/gdb/testsuite/gdb.fortran/oop-extend-type.exp
> new file mode 100644
> index 000000000000..8c3bb50a3ac6
> --- /dev/null
> +++ b/gdb/testsuite/gdb.fortran/oop-extend-type.exp
> @@ -0,0 +1,113 @@
> +# Copyright 2018 Free Software Foundation, Inc.
> +
> +# This program is free software; you can redistribute it and/or modify
> +# it under the terms of the GNU General Public License as published by
> +# the Free Software Foundation; either version 3 of the License, or
> +# (at your option) any later version.
> +#
> +# This program is distributed in the hope that it will be useful,
> +# but WITHOUT ANY WARRANTY; without even the implied warranty of
> +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
> +# GNU General Public License for more details.
> +#
> +# You should have received a copy of the GNU General Public License
> +# along with this program.  If not, see <http://www.gnu.org/licenses/>.
> +
> +standard_testfile ".f90"
> +load_lib "fortran.exp"
> +
> +if { [prepare_for_testing ${testfile}.exp ${testfile} ${srcfile} \
> +     {debug f90 quiet}] } {
> +	 return -1
> +}
> +
> +if ![runto_main] {
> +    untested "could not run to main"
> +    return -1
> +}
> +
> +# Depending on the compiler being used, the type names can be printed differently.
> +set real [fortran_real4]
> +
> +gdb_breakpoint [gdb_get_line_number "! Before vla allocation"]
> +gdb_continue_to_breakpoint "! Before vla allocation" ".*! Before vla allocation"
> +gdb_test "whatis wp_vla" "type = <not allocated>"
> +
> +gdb_breakpoint [gdb_get_line_number "! After value assignment"]
> +gdb_continue_to_breakpoint "! After value assignment" ".*! After value assignment"
> +set test "p wp%coo"
> +gdb_test_multiple "$test" "$test" {
> +    -re " = \\(1, 2, 1\\)\r\n$gdb_prompt $" {
> +      pass "$test"
> +    }
> +    -re "There is no member named coo.\r\n$gdb_prompt $" {
> +      kfail "gcc/49475" "$test"
> +    }
> +}
> +gdb_test "p wp%point%coo" " = \\(1, 2, 1\\)"
> +gdb_test "p wp%point" " = \\( coo = \\(1, 2, 1\\) \\)"
> +gdb_test "p wp" " = \\( point = \\( coo = \\(1, 2, 1\\) \\), angle = 100 \\)"
> +
> +gdb_test "whatis wp" "type = Type waypoint"
> +set output_pass [multi_line "type = Type, extends\\(point\\) :: waypoint" \
> +              "    Type point :: point" \
> +              "    $real :: angle" \
> +              "End Type waypoint"]
> +set output_kfail [multi_line "type = Type waypoint" \
> +"    Type point :: point" \
> +"    $real :: angle" \
> +"End Type waypoint"]
> +set test "ptype wp"
> +gdb_test_multiple $test %test {
> +    -re "$output_pass\r\n$gdb_prompt $" {
> +      pass "$test"
> +    }
> +    -re "$output_kfail\r\n$gdb_prompt $" {
> +      kfail "gcc/49475" "$test"
> +    }
> +}
> +
> +set test "ptype wp%coo"
> +gdb_test_multiple "$test" "$test" {
> +    -re "$real \\(3\\)\r\n$gdb_prompt $" {
> +      pass "$test"
> +    }
> +    -re "There is no member named coo.\r\n$gdb_prompt $" {
> +      kfail "gcc/49475" "$test"
> +    }
> +}
> +gdb_test "ptype wp%point%coo" "$real \\(3\\)"
> +
> +set test "p wp_vla(1)%coo"
> +gdb_test_multiple "$test" "$test" {
> +    -re " = \\(10, 12, 10\\)\r\n$gdb_prompt $" {
> +      pass "$test"
> +    }
> +    -re "There is no member named coo.\r\n$gdb_prompt $" {
> +      kfail "gcc/49475" "$test"
> +    }
> +}
> +gdb_test "p wp_vla(1)%point%coo" " = \\(10, 12, 10\\)"
> +gdb_test "p wp_vla(1)%point" " = \\( coo = \\(10, 12, 10\\) \\)"
> +gdb_test "p wp_vla(1)" " = \\( point = \\( coo = \\(10, 12, 10\\) \\), angle = 101 \\)"
> +
> +gdb_test "whatis wp_vla" "type = Type waypoint \\(3\\)"
> +set test "ptype wp_vla"
> +gdb_test_multiple $test %test {
> +    -re "$output_pass \\(3\\)\r\n$gdb_prompt $" {
> +      pass "$test"
> +    }
> +    -re "$output_kfail \\(3\\)\r\n$gdb_prompt $" {
> +      kfail "gcc/49475" "$test"
> +    }
> +}
> +set test "ptype wp_vla(1)%coo"
> +gdb_test_multiple "$test" "$test" {
> +    -re "$real \\(3\\)\r\n$gdb_prompt $" {
> +      pass "$test"
> +    }
> +    -re "There is no member named coo.\r\n$gdb_prompt $" {
> +      kfail "gcc/49475" "$test"
> +    }
> +}
> +gdb_test "ptype wp_vla(1)%point%coo" "$real \\(3\\)"
> diff --git a/gdb/testsuite/gdb.fortran/oop-extend-type.f90 b/gdb/testsuite/gdb.fortran/oop-extend-type.f90
> new file mode 100644
> index 000000000000..1fe8611f4632
> --- /dev/null
> +++ b/gdb/testsuite/gdb.fortran/oop-extend-type.f90
> @@ -0,0 +1,56 @@
> +! Copyright 2018 Free Software Foundation, Inc.
> +! 
> +! This program is free software; you can redistribute it and/or modify
> +! it under the terms of the GNU General Public License as published by
> +! the Free Software Foundation; either version 3 of the License, or
> +! (at your option) any later version.
> +! 
> +! This program is distributed in the hope that it will be useful,
> +! but WITHOUT ANY WARRANTY; without even the implied warranty of
> +! MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
> +! GNU General Public License for more details.
> +! 
> +! You should have received a copy of the GNU General Public License
> +! along with this program.  If not, see <http://www.gnu.org/licenses/>.
> +
> +module testmod
> +    implicit none
> +    type :: point
> +        real :: coo(3)
> +    end type
> +
> +    type, extends(point) :: waypoint
> +        real :: angle
> +    end type
> +
> +end module
> +
> +program testprog
> +    use testmod
> +    implicit none
> +
> +    logical l
> +    type(waypoint) :: wp
> +    type(waypoint), allocatable :: wp_vla(:)
> +
> +    l = allocated(wp_vla)
> +    allocate(wp_vla(3))               ! Before vla allocation
> +
> +    l = allocated(wp_vla)             ! After vla allocation
> +    wp%angle = 100.00
> +    wp%point%coo(:) = 1.00
> +    wp%point%coo(2) = 2.00
> +
> +    wp_vla(1)%angle = 101.00
> +    wp_vla(1)%point%coo(:) = 10.00
> +    wp_vla(1)%point%coo(2) = 12.00
> +    wp_vla(2)%angle = 102.00
> +    wp_vla(2)%point%coo(:) = 20.00
> +    wp_vla(2)%point%coo(2) = 22.00
> +    wp_vla(3)%angle = 103.00
> +    wp_vla(3)%point%coo(:) = 30.00
> +    wp_vla(3)%point%coo(2) = 32.00
> +
> +    print *, wp, wp_vla               ! After value assignment
> +
> +end program
> diff --git a/gdb/valops.c b/gdb/valops.c
> index c45caefbf1e0..a34e74b2bee9 100644
> --- a/gdb/valops.c
> +++ b/gdb/valops.c
> @@ -2163,6 +2163,12 @@ value_struct_elt (struct value **argp, struct value **args,
>        if (v)
>  	return v;
>  
> +      /* fortran: If it is not a field it is the
> +         type name of an inherited structure */
> +      v = search_struct_field (name, *argp, t, 1);
> +      if (v)
> +	return v;
> +
>        /* C++: If it was not found as a data field, then try to
>           return it as a pointer to a method.  */
>        v = search_struct_method (name, argp, args, 0, 
> 

Re: [PATCH v2 3/7] Fortran: Accessing fields of inherited types via fully qualified name.

[Paweł Wódkowski]

On 03.12.2018 14:44, Richard Bunt wrote:
> Hi Pawel,
> 
> Is there a compiler where all the marked kfails pass? I've given this patch a spin
> with the Fortran compilers from various vendors, but there is a varying number of kfails.
> 

To be honest I'm testing gfortran and ifort and both are producing the 
same six kfails.

Which compilers are you testing against? I can add them to the CI if 
required.

Pawel

> I'd like to test everything if possible.
> 
> Many thanks,
> 
> Rich
> 
> On 11/19/18 9:38 PM, Pawel Wodkowski wrote:
>> From: Bernhard Heckel <bernhard.heckel@intel.com>
>>
>> Fortran 2003 supports type extension. This patch allows access
>> to inherited members by using it's fully qualified name as
>> described in the fortran standard.
>>
>> Before:
>> (gdb) print my_extended_obj%base_class_name%member_base
>> Syntax error near base_class_name%member_base
>>
>> (gdb) print my_extended_obj%member_base
>> $1 = (10, 10, 10)
>>
>> After:
>> (gdb) print my_extended_obj%base_clase_name%member_base
>> $1 = (10, 10, 10)
>>
>> (gdb) print my_extended_obj%member_base
>> $1 = (10, 10, 10)
>>
>> 2016-04-22  Bernhard Heckel  <bernhard.heckel@intel.com>
>>
>> gdb/Changelog:
>> 	* f-exp.y (name): Allow TYPENAME.
>> 	* valops.c (search_struct_method): Look also for baseclass.
>>
>> gdb/Testsuite/Changelog:
>> 	* gdb.fortran/oop-extend-type.f90: New.
>> 	* gdb.fortran/oop-extend-type.exp: New.
>> ---
>>   gdb/f-exp.y                                   |   7 +-
>>   gdb/testsuite/gdb.fortran/oop-extend-type.exp | 113 ++++++++++++++++++++++++++
>>   gdb/testsuite/gdb.fortran/oop-extend-type.f90 |  56 +++++++++++++
>>   gdb/valops.c                                  |   6 ++
>>   4 files changed, 180 insertions(+), 2 deletions(-)
>>   create mode 100644 gdb/testsuite/gdb.fortran/oop-extend-type.exp
>>   create mode 100644 gdb/testsuite/gdb.fortran/oop-extend-type.f90
>>
>> diff --git a/gdb/f-exp.y b/gdb/f-exp.y
>> index 390bd45081b7..4c2e101699ac 100644
>> --- a/gdb/f-exp.y
>> +++ b/gdb/f-exp.y
>> @@ -604,8 +604,11 @@ nonempty_typelist
>>   		}
>>   	;
>>   
>> -name	:	NAME
>> -		{  $$ = $1.stoken; }
>> +name
>> +	:	NAME
>> +		{ $$ = $1.stoken; }
>> +	|	TYPENAME
>> +		{ $$ = $1.stoken; }
>>   	;
>>   
>>   name_not_typename :	NAME
>> diff --git a/gdb/testsuite/gdb.fortran/oop-extend-type.exp b/gdb/testsuite/gdb.fortran/oop-extend-type.exp
>> new file mode 100644
>> index 000000000000..8c3bb50a3ac6
>> --- /dev/null
>> +++ b/gdb/testsuite/gdb.fortran/oop-extend-type.exp
>> @@ -0,0 +1,113 @@
>> +# Copyright 2018 Free Software Foundation, Inc.
>> +
>> +# This program is free software; you can redistribute it and/or modify
>> +# it under the terms of the GNU General Public License as published by
>> +# the Free Software Foundation; either version 3 of the License, or
>> +# (at your option) any later version.
>> +#
>> +# This program is distributed in the hope that it will be useful,
>> +# but WITHOUT ANY WARRANTY; without even the implied warranty of
>> +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
>> +# GNU General Public License for more details.
>> +#
>> +# You should have received a copy of the GNU General Public License
>> +# along with this program.  If not, see <http://www.gnu.org/licenses/>.
>> +
>> +standard_testfile ".f90"
>> +load_lib "fortran.exp"
>> +
>> +if { [prepare_for_testing ${testfile}.exp ${testfile} ${srcfile} \
>> +     {debug f90 quiet}] } {
>> +	 return -1
>> +}
>> +
>> +if ![runto_main] {
>> +    untested "could not run to main"
>> +    return -1
>> +}
>> +
>> +# Depending on the compiler being used, the type names can be printed differently.
>> +set real [fortran_real4]
>> +
>> +gdb_breakpoint [gdb_get_line_number "! Before vla allocation"]
>> +gdb_continue_to_breakpoint "! Before vla allocation" ".*! Before vla allocation"
>> +gdb_test "whatis wp_vla" "type = <not allocated>"
>> +
>> +gdb_breakpoint [gdb_get_line_number "! After value assignment"]
>> +gdb_continue_to_breakpoint "! After value assignment" ".*! After value assignment"
>> +set test "p wp%coo"
>> +gdb_test_multiple "$test" "$test" {
>> +    -re " = \\(1, 2, 1\\)\r\n$gdb_prompt $" {
>> +      pass "$test"
>> +    }
>> +    -re "There is no member named coo.\r\n$gdb_prompt $" {
>> +      kfail "gcc/49475" "$test"
>> +    }
>> +}
>> +gdb_test "p wp%point%coo" " = \\(1, 2, 1\\)"
>> +gdb_test "p wp%point" " = \\( coo = \\(1, 2, 1\\) \\)"
>> +gdb_test "p wp" " = \\( point = \\( coo = \\(1, 2, 1\\) \\), angle = 100 \\)"
>> +
>> +gdb_test "whatis wp" "type = Type waypoint"
>> +set output_pass [multi_line "type = Type, extends\\(point\\) :: waypoint" \
>> +              "    Type point :: point" \
>> +              "    $real :: angle" \
>> +              "End Type waypoint"]
>> +set output_kfail [multi_line "type = Type waypoint" \
>> +"    Type point :: point" \
>> +"    $real :: angle" \
>> +"End Type waypoint"]
>> +set test "ptype wp"
>> +gdb_test_multiple $test %test {
>> +    -re "$output_pass\r\n$gdb_prompt $" {
>> +      pass "$test"
>> +    }
>> +    -re "$output_kfail\r\n$gdb_prompt $" {
>> +      kfail "gcc/49475" "$test"
>> +    }
>> +}
>> +
>> +set test "ptype wp%coo"
>> +gdb_test_multiple "$test" "$test" {
>> +    -re "$real \\(3\\)\r\n$gdb_prompt $" {
>> +      pass "$test"
>> +    }
>> +    -re "There is no member named coo.\r\n$gdb_prompt $" {
>> +      kfail "gcc/49475" "$test"
>> +    }
>> +}
>> +gdb_test "ptype wp%point%coo" "$real \\(3\\)"
>> +
>> +set test "p wp_vla(1)%coo"
>> +gdb_test_multiple "$test" "$test" {
>> +    -re " = \\(10, 12, 10\\)\r\n$gdb_prompt $" {
>> +      pass "$test"
>> +    }
>> +    -re "There is no member named coo.\r\n$gdb_prompt $" {
>> +      kfail "gcc/49475" "$test"
>> +    }
>> +}
>> +gdb_test "p wp_vla(1)%point%coo" " = \\(10, 12, 10\\)"
>> +gdb_test "p wp_vla(1)%point" " = \\( coo = \\(10, 12, 10\\) \\)"
>> +gdb_test "p wp_vla(1)" " = \\( point = \\( coo = \\(10, 12, 10\\) \\), angle = 101 \\)"
>> +
>> +gdb_test "whatis wp_vla" "type = Type waypoint \\(3\\)"
>> +set test "ptype wp_vla"
>> +gdb_test_multiple $test %test {
>> +    -re "$output_pass \\(3\\)\r\n$gdb_prompt $" {
>> +      pass "$test"
>> +    }
>> +    -re "$output_kfail \\(3\\)\r\n$gdb_prompt $" {
>> +      kfail "gcc/49475" "$test"
>> +    }
>> +}
>> +set test "ptype wp_vla(1)%coo"
>> +gdb_test_multiple "$test" "$test" {
>> +    -re "$real \\(3\\)\r\n$gdb_prompt $" {
>> +      pass "$test"
>> +    }
>> +    -re "There is no member named coo.\r\n$gdb_prompt $" {
>> +      kfail "gcc/49475" "$test"
>> +    }
>> +}
>> +gdb_test "ptype wp_vla(1)%point%coo" "$real \\(3\\)"
>> diff --git a/gdb/testsuite/gdb.fortran/oop-extend-type.f90 b/gdb/testsuite/gdb.fortran/oop-extend-type.f90
>> new file mode 100644
>> index 000000000000..1fe8611f4632
>> --- /dev/null
>> +++ b/gdb/testsuite/gdb.fortran/oop-extend-type.f90
>> @@ -0,0 +1,56 @@
>> +! Copyright 2018 Free Software Foundation, Inc.
>> +!
>> +! This program is free software; you can redistribute it and/or modify
>> +! it under the terms of the GNU General Public License as published by
>> +! the Free Software Foundation; either version 3 of the License, or
>> +! (at your option) any later version.
>> +!
>> +! This program is distributed in the hope that it will be useful,
>> +! but WITHOUT ANY WARRANTY; without even the implied warranty of
>> +! MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
>> +! GNU General Public License for more details.
>> +!
>> +! You should have received a copy of the GNU General Public License
>> +! along with this program.  If not, see <http://www.gnu.org/licenses/>.
>> +
>> +module testmod
>> +    implicit none
>> +    type :: point
>> +        real :: coo(3)
>> +    end type
>> +
>> +    type, extends(point) :: waypoint
>> +        real :: angle
>> +    end type
>> +
>> +end module
>> +
>> +program testprog
>> +    use testmod
>> +    implicit none
>> +
>> +    logical l
>> +    type(waypoint) :: wp
>> +    type(waypoint), allocatable :: wp_vla(:)
>> +
>> +    l = allocated(wp_vla)
>> +    allocate(wp_vla(3))               ! Before vla allocation
>> +
>> +    l = allocated(wp_vla)             ! After vla allocation
>> +    wp%angle = 100.00
>> +    wp%point%coo(:) = 1.00
>> +    wp%point%coo(2) = 2.00
>> +
>> +    wp_vla(1)%angle = 101.00
>> +    wp_vla(1)%point%coo(:) = 10.00
>> +    wp_vla(1)%point%coo(2) = 12.00
>> +    wp_vla(2)%angle = 102.00
>> +    wp_vla(2)%point%coo(:) = 20.00
>> +    wp_vla(2)%point%coo(2) = 22.00
>> +    wp_vla(3)%angle = 103.00
>> +    wp_vla(3)%point%coo(:) = 30.00
>> +    wp_vla(3)%point%coo(2) = 32.00
>> +
>> +    print *, wp, wp_vla               ! After value assignment
>> +
>> +end program
>> diff --git a/gdb/valops.c b/gdb/valops.c
>> index c45caefbf1e0..a34e74b2bee9 100644
>> --- a/gdb/valops.c
>> +++ b/gdb/valops.c
>> @@ -2163,6 +2163,12 @@ value_struct_elt (struct value **argp, struct value **args,
>>         if (v)
>>   	return v;
>>   
>> +      /* fortran: If it is not a field it is the
>> +         type name of an inherited structure */
>> +      v = search_struct_field (name, *argp, t, 1);
>> +      if (v)
>> +	return v;
>> +
>>         /* C++: If it was not found as a data field, then try to
>>            return it as a pointer to a method.  */
>>         v = search_struct_method (name, argp, args, 0,
>>
> 

Re: [PATCH v2 3/7] Fortran: Accessing fields of inherited types via fully qualified name.

[Richard Bunt]

Hi Paweł,

I was testing with gfortran 7.3.0 and ifort 18.2. On the latter I observe
2 kfails relating to the missing extends in the ptype output. However, this
is because I do not have patch 4 applied. I will take a closer look at this
patch with ifort 18.2 since I can analyze the implementation in action.

On 12/4/18 9:05 PM, Paweł Wódkowski wrote:

>>> +gdb_test "whatis wp_vla" "type = <not allocated>"

A small nit while I'm here:
GDB tests needs to be uniquely identifiable, see https://sourceware.org/gdb/wiki/GDBTestcaseCookbook#Make_sure_test_messages_are_unique

      1 (gdb) PASS: gdb.fortran/oop-extend-type.exp: continue to breakpoint: ! After value assignment
      1 (gdb) PASS: gdb.fortran/oop-extend-type.exp: continue to breakpoint: ! Before vla allocation
      1 (gdb) PASS: gdb.fortran/oop-extend-type.exp: ptype wp%point%coo
      1 (gdb) PASS: gdb.fortran/oop-extend-type.exp: ptype wp_vla(1)%point%coo
      1 (gdb) PASS: gdb.fortran/oop-extend-type.exp: p wp
      1 (gdb) PASS: gdb.fortran/oop-extend-type.exp: p wp%point
      1 (gdb) PASS: gdb.fortran/oop-extend-type.exp: p wp%point%coo
      1 (gdb) PASS: gdb.fortran/oop-extend-type.exp: p wp_vla(1)
      1 (gdb) PASS: gdb.fortran/oop-extend-type.exp: p wp_vla(1)%point
      1 (gdb) PASS: gdb.fortran/oop-extend-type.exp: p wp_vla(1)%point%coo
      1 (gdb) PASS: gdb.fortran/oop-extend-type.exp: whatis wp
      2 (gdb) PASS: gdb.fortran/oop-extend-type.exp: whatis wp_vla


Many thanks,

Rich.

Re: [PATCH v2 3/7] Fortran: Accessing fields of inherited types via fully qualified name.

[Paweł Wódkowski]

On 06.12.2018 16:02, Richard Bunt wrote:
> Hi Paweł,
> 
> I was testing with gfortran 7.3.0 and ifort 18.2. On the latter I observe
> 2 kfails relating to the missing extends in the ptype output. However, this
> is because I do not have patch 4 applied. I will take a closer look at this
> patch with ifort 18.2 since I can analyze the implementation in action.
> 

Ok, will wait (perhaps Wednesday) with v3. Pls, let know if there is 
something more to change.

btw, I have ifort 19.0.1.144 20181018


> On 12/4/18 9:05 PM, Paweł Wódkowski wrote:
> 
>>>> +gdb_test "whatis wp_vla" "type = <not allocated>"
> 
> A small nit while I'm here:
> GDB tests needs to be uniquely identifiable, see https://sourceware.org/gdb/wiki/GDBTestcaseCookbook#Make_sure_test_messages_are_unique
> 
>        1 (gdb) PASS: gdb.fortran/oop-extend-type.exp: continue to breakpoint: ! After value assignment
>        1 (gdb) PASS: gdb.fortran/oop-extend-type.exp: continue to breakpoint: ! Before vla allocation
>        1 (gdb) PASS: gdb.fortran/oop-extend-type.exp: ptype wp%point%coo
>        1 (gdb) PASS: gdb.fortran/oop-extend-type.exp: ptype wp_vla(1)%point%coo
>        1 (gdb) PASS: gdb.fortran/oop-extend-type.exp: p wp
>        1 (gdb) PASS: gdb.fortran/oop-extend-type.exp: p wp%point
>        1 (gdb) PASS: gdb.fortran/oop-extend-type.exp: p wp%point%coo
>        1 (gdb) PASS: gdb.fortran/oop-extend-type.exp: p wp_vla(1)
>        1 (gdb) PASS: gdb.fortran/oop-extend-type.exp: p wp_vla(1)%point
>        1 (gdb) PASS: gdb.fortran/oop-extend-type.exp: p wp_vla(1)%point%coo
>        1 (gdb) PASS: gdb.fortran/oop-extend-type.exp: whatis wp
>        2 (gdb) PASS: gdb.fortran/oop-extend-type.exp: whatis wp_vla
> 
> 

Thx for the link. I will fix this.

Pawel

> Many thanks,
> 
> Rich.
> 

Re: [PATCH v2 3/7] Fortran: Accessing fields of inherited types via fully qualified name.

[Richard Bunt]

I have no further comments on v2 and look forward to v3.

Many thanks,

Rich.

On 12/9/18 8:40 PM, Paweł Wódkowski wrote:
> On 06.12.2018 16:02, Richard Bunt wrote:
>> Hi Paweł,
>>
>> I was testing with gfortran 7.3.0 and ifort 18.2. On the latter I observe
>> 2 kfails relating to the missing extends in the ptype output. However, this
>> is because I do not have patch 4 applied. I will take a closer look at this
>> patch with ifort 18.2 since I can analyze the implementation in action.
>>
> 
> Ok, will wait (perhaps Wednesday) with v3. Pls, let know if there is 
> something more to change.
> 
> btw, I have ifort 19.0.1.144 20181018
> 
> 
>> On 12/4/18 9:05 PM, Paweł Wódkowski wrote:
>>
>>>>> +gdb_test "whatis wp_vla" "type = <not allocated>"
>>
>> A small nit while I'm here:
>> GDB tests needs to be uniquely identifiable, see https://sourceware.org/gdb/wiki/GDBTestcaseCookbook#Make_sure_test_messages_are_unique
>>
>>        1 (gdb) PASS: gdb.fortran/oop-extend-type.exp: continue to breakpoint: ! After value assignment
>>        1 (gdb) PASS: gdb.fortran/oop-extend-type.exp: continue to breakpoint: ! Before vla allocation
>>        1 (gdb) PASS: gdb.fortran/oop-extend-type.exp: ptype wp%point%coo
>>        1 (gdb) PASS: gdb.fortran/oop-extend-type.exp: ptype wp_vla(1)%point%coo
>>        1 (gdb) PASS: gdb.fortran/oop-extend-type.exp: p wp
>>        1 (gdb) PASS: gdb.fortran/oop-extend-type.exp: p wp%point
>>        1 (gdb) PASS: gdb.fortran/oop-extend-type.exp: p wp%point%coo
>>        1 (gdb) PASS: gdb.fortran/oop-extend-type.exp: p wp_vla(1)
>>        1 (gdb) PASS: gdb.fortran/oop-extend-type.exp: p wp_vla(1)%point
>>        1 (gdb) PASS: gdb.fortran/oop-extend-type.exp: p wp_vla(1)%point%coo
>>        1 (gdb) PASS: gdb.fortran/oop-extend-type.exp: whatis wp
>>        2 (gdb) PASS: gdb.fortran/oop-extend-type.exp: whatis wp_vla
>>
>>
> 
> Thx for the link. I will fix this.
> 
> Pawel
> 
>> Many thanks,
>>
>> Rich.
>>
> 

Re: [PATCH v2 3/7] Fortran: Accessing fields of inherited types via fully qualified name.

[Paweł Wódkowski]

On 2019-01-16 12:11, Richard Bunt wrote:
> I have no further comments on v2 and look forward to v3.
> 
> Many thanks,

Great, I tough that this was forgotten ;)

If you don't mind I will  decouple this feature from unrelated patches 
in this series.

I will send v3 shortly.

Pawel

> 
> Rich.
> 
> On 12/9/18 8:40 PM, Paweł Wódkowski wrote:
>> On 06.12.2018 16:02, Richard Bunt wrote:
>>> Hi Paweł,
>>>
>>> I was testing with gfortran 7.3.0 and ifort 18.2. On the latter I observe
>>> 2 kfails relating to the missing extends in the ptype output. However, this
>>> is because I do not have patch 4 applied. I will take a closer look at this
>>> patch with ifort 18.2 since I can analyze the implementation in action.
>>>
>>
>> Ok, will wait (perhaps Wednesday) with v3. Pls, let know if there is
>> something more to change.
>>
>> btw, I have ifort 19.0.1.144 20181018
>>
>>
>>> On 12/4/18 9:05 PM, Paweł Wódkowski wrote:
>>>
>>>>>> +gdb_test "whatis wp_vla" "type = <not allocated>"
>>>
>>> A small nit while I'm here:
>>> GDB tests needs to be uniquely identifiable, see https://sourceware.org/gdb/wiki/GDBTestcaseCookbook#Make_sure_test_messages_are_unique
>>>
>>>         1 (gdb) PASS: gdb.fortran/oop-extend-type.exp: continue to breakpoint: ! After value assignment
>>>         1 (gdb) PASS: gdb.fortran/oop-extend-type.exp: continue to breakpoint: ! Before vla allocation
>>>         1 (gdb) PASS: gdb.fortran/oop-extend-type.exp: ptype wp%point%coo
>>>         1 (gdb) PASS: gdb.fortran/oop-extend-type.exp: ptype wp_vla(1)%point%coo
>>>         1 (gdb) PASS: gdb.fortran/oop-extend-type.exp: p wp
>>>         1 (gdb) PASS: gdb.fortran/oop-extend-type.exp: p wp%point
>>>         1 (gdb) PASS: gdb.fortran/oop-extend-type.exp: p wp%point%coo
>>>         1 (gdb) PASS: gdb.fortran/oop-extend-type.exp: p wp_vla(1)
>>>         1 (gdb) PASS: gdb.fortran/oop-extend-type.exp: p wp_vla(1)%point
>>>         1 (gdb) PASS: gdb.fortran/oop-extend-type.exp: p wp_vla(1)%point%coo
>>>         1 (gdb) PASS: gdb.fortran/oop-extend-type.exp: whatis wp
>>>         2 (gdb) PASS: gdb.fortran/oop-extend-type.exp: whatis wp_vla
>>>
>>>
>>
>> Thx for the link. I will fix this.
>>
>> Pawel
>>
>>> Many thanks,
>>>
>>> Rich.
>>>
>>
> 

Re: [PATCH v2 3/7] Fortran: Accessing fields of inherited types via fully qualified name.

[Richard Bunt]

Mail tends to drop off the top of the list pretty quickly, so please
keep pinging to help move the work along:

> The generally accepted ping frequency is 2 weeks after first submission, 
> and every week thereafter. 

https://sourceware.org/gdb/wiki/ContributionChecklist (13)

Many thanks,

Rich.

On 1/18/19 8:32 AM, Paweł Wódkowski wrote:
> On 2019-01-16 12:11, Richard Bunt wrote:
>> I have no further comments on v2 and look forward to v3.
>>
>> Many thanks,
> 
> Great, I tough that this was forgotten ;)
> 
> If you don't mind I will  decouple this feature from unrelated patches 
> in this series.
> 
> I will send v3 shortly.
> 
> Pawel
> 
>>
>> Rich.
>>
>> On 12/9/18 8:40 PM, Paweł Wódkowski wrote:
>>> On 06.12.2018 16:02, Richard Bunt wrote:
>>>> Hi Paweł,
>>>>
>>>> I was testing with gfortran 7.3.0 and ifort 18.2. On the latter I observe
>>>> 2 kfails relating to the missing extends in the ptype output. However, this
>>>> is because I do not have patch 4 applied. I will take a closer look at this
>>>> patch with ifort 18.2 since I can analyze the implementation in action.
>>>>
>>>
>>> Ok, will wait (perhaps Wednesday) with v3. Pls, let know if there is
>>> something more to change.
>>>
>>> btw, I have ifort 19.0.1.144 20181018
>>>
>>>
>>>> On 12/4/18 9:05 PM, Paweł Wódkowski wrote:
>>>>
>>>>>>> +gdb_test "whatis wp_vla" "type = <not allocated>"
>>>>
>>>> A small nit while I'm here:
>>>> GDB tests needs to be uniquely identifiable, see https://sourceware.org/gdb/wiki/GDBTestcaseCookbook#Make_sure_test_messages_are_unique
>>>>
>>>>         1 (gdb) PASS: gdb.fortran/oop-extend-type.exp: continue to breakpoint: ! After value assignment
>>>>         1 (gdb) PASS: gdb.fortran/oop-extend-type.exp: continue to breakpoint: ! Before vla allocation
>>>>         1 (gdb) PASS: gdb.fortran/oop-extend-type.exp: ptype wp%point%coo
>>>>         1 (gdb) PASS: gdb.fortran/oop-extend-type.exp: ptype wp_vla(1)%point%coo
>>>>         1 (gdb) PASS: gdb.fortran/oop-extend-type.exp: p wp
>>>>         1 (gdb) PASS: gdb.fortran/oop-extend-type.exp: p wp%point
>>>>         1 (gdb) PASS: gdb.fortran/oop-extend-type.exp: p wp%point%coo
>>>>         1 (gdb) PASS: gdb.fortran/oop-extend-type.exp: p wp_vla(1)
>>>>         1 (gdb) PASS: gdb.fortran/oop-extend-type.exp: p wp_vla(1)%point
>>>>         1 (gdb) PASS: gdb.fortran/oop-extend-type.exp: p wp_vla(1)%point%coo
>>>>         1 (gdb) PASS: gdb.fortran/oop-extend-type.exp: whatis wp
>>>>         2 (gdb) PASS: gdb.fortran/oop-extend-type.exp: whatis wp_vla
>>>>
>>>>
>>>
>>> Thx for the link. I will fix this.
>>>
>>> Pawel
>>>
>>>> Many thanks,
>>>>
>>>> Rich.
>>>>
>>>
>>
> 

Re: [PATCH v2 4/7] Fortran: Ptype, print type extension.

[Andrew Burgess]

* Pawel Wodkowski <pwodkowski@pl.sii.eu> [2018-11-19 22:38:47 +0100]:

> From: Bernhard Heckel <bernhard.heckel@intel.com>
> 
> Print base-class of an extended type when doing a ptype.
> 
> 2016-05-24 Bernhard Heckel  <bernhard.heckel@intel.com>
> 
> gdb/Changelog:
> 	* gdb/f-typeprint.c (f_type_print_derivation_info): New.
> 	(f_type_print_base): Print baseclass info.

Could this patch have a test case please.  Also there's a minor
whitespace issue I've pointed out below.

Thanks,
Andrew

> 
> Changes from V1 to V2:
> 	clear commit message.
> ---
>  gdb/f-typeprint.c | 30 +++++++++++++++++++++++++++---
>  1 file changed, 27 insertions(+), 3 deletions(-)
> 
> diff --git a/gdb/f-typeprint.c b/gdb/f-typeprint.c
> index 3f13c111ac56..133eaf9b9848 100644
> --- a/gdb/f-typeprint.c
> +++ b/gdb/f-typeprint.c
> @@ -256,6 +256,25 @@ f_type_print_varspec_suffix (struct type *type, struct ui_file *stream,
>      }
>  }
>  
> +/* If TYPE is an extended type, then print out derivation information.
> +
> +   A typical output could look like this:
> +   "Type, extends(point) :: waypoint"
> +   "    Type point :: point"
> +   "    real(kind=4) :: angle"
> +   "End Type waypoint"
> + */
> +
> +static void
> +f_type_print_derivation_info (struct type *type, struct ui_file *stream)
> +{
> +  /* Fortran doesn't support multiple inheritance. */

Extra whitespace after '.'.

> +  int i = 0;
> +
> +  if (TYPE_N_BASECLASSES (type) > 0)
> +    fprintf_filtered (stream, ", extends(%s) ::", TYPE_NAME (TYPE_BASECLASS (type, i)));
> +}
> +
>  /* Print the name of the type (or the ultimate pointer target,
>     function value or array element), or the description of a
>     structure or union.
> @@ -372,10 +391,15 @@ f_type_print_base (struct type *type, struct ui_file *stream, int show,
>      case TYPE_CODE_STRUCT:
>      case TYPE_CODE_UNION:
>        if (TYPE_CODE (type) == TYPE_CODE_UNION)
> -	fprintfi_filtered (level, stream, "Type, C_Union :: ");
> +	fprintfi_filtered (level, stream, "Type, C_Union ::");
>        else
> -	fprintfi_filtered (level, stream, "Type ");
> -      fputs_filtered (TYPE_NAME (type), stream);
> +	fprintfi_filtered (level, stream, "Type");
> +
> +      if (show > 0)
> +	f_type_print_derivation_info (type, stream);
> +
> +      fprintf_filtered (stream, " %s", TYPE_NAME (type));
> +
>        /* According to the definition,
>           we only print structure elements in case show > 0.  */
>        if (show > 0)
> -- 
> 2.7.4
> 

Re: [PATCH v2 3/7] Fortran: Accessing fields of inherited types via fully qualified name.

[Paweł Wódkowski]

Mail didn't dropped.

During rebase to latest master I spotted that last patch about scope
operator stopped working. I can't figure this out yet, but setting
breakpoint using scope isn't working any more despite that gdb see
proper scope when I dump the symbols.

Sorry but this will take some more time for V3 version as I want to be
1000% sure this is working.


On 30.01.2019 12:10, Richard Bunt wrote:
> Mail tends to drop off the top of the list pretty quickly, so please
> keep pinging to help move the work along:
>
>> The generally accepted ping frequency is 2 weeks after first submission, 
>> and every week thereafter. 
> https://sourceware.org/gdb/wiki/ContributionChecklist (13)
>
> Many thanks,
>
> Rich.
>
> On 1/18/19 8:32 AM, Paweł Wódkowski wrote:
>> On 2019-01-16 12:11, Richard Bunt wrote:
>>> I have no further comments on v2 and look forward to v3.
>>>
>>> Many thanks,
>> Great, I tough that this was forgotten ;)
>>
>> If you don't mind I will  decouple this feature from unrelated patches 
>> in this series.
>>
>> I will send v3 shortly.
>>
>> Pawel
>>
>>> Rich.
>>>
>>> On 12/9/18 8:40 PM, Paweł Wódkowski wrote:
>>>> On 06.12.2018 16:02, Richard Bunt wrote:
>>>>> Hi Paweł,
>>>>>
>>>>> I was testing with gfortran 7.3.0 and ifort 18.2. On the latter I observe
>>>>> 2 kfails relating to the missing extends in the ptype output. However, this
>>>>> is because I do not have patch 4 applied. I will take a closer look at this
>>>>> patch with ifort 18.2 since I can analyze the implementation in action.
>>>>>
>>>> Ok, will wait (perhaps Wednesday) with v3. Pls, let know if there is
>>>> something more to change.
>>>>
>>>> btw, I have ifort 19.0.1.144 20181018
>>>>
>>>>
>>>>> On 12/4/18 9:05 PM, Paweł Wódkowski wrote:
>>>>>
>>>>>>>> +gdb_test "whatis wp_vla" "type = <not allocated>"
>>>>> A small nit while I'm here:
>>>>> GDB tests needs to be uniquely identifiable, see https://sourceware.org/gdb/wiki/GDBTestcaseCookbook#Make_sure_test_messages_are_unique
>>>>>
>>>>>         1 (gdb) PASS: gdb.fortran/oop-extend-type.exp: continue to breakpoint: ! After value assignment
>>>>>         1 (gdb) PASS: gdb.fortran/oop-extend-type.exp: continue to breakpoint: ! Before vla allocation
>>>>>         1 (gdb) PASS: gdb.fortran/oop-extend-type.exp: ptype wp%point%coo
>>>>>         1 (gdb) PASS: gdb.fortran/oop-extend-type.exp: ptype wp_vla(1)%point%coo
>>>>>         1 (gdb) PASS: gdb.fortran/oop-extend-type.exp: p wp
>>>>>         1 (gdb) PASS: gdb.fortran/oop-extend-type.exp: p wp%point
>>>>>         1 (gdb) PASS: gdb.fortran/oop-extend-type.exp: p wp%point%coo
>>>>>         1 (gdb) PASS: gdb.fortran/oop-extend-type.exp: p wp_vla(1)
>>>>>         1 (gdb) PASS: gdb.fortran/oop-extend-type.exp: p wp_vla(1)%point
>>>>>         1 (gdb) PASS: gdb.fortran/oop-extend-type.exp: p wp_vla(1)%point%coo
>>>>>         1 (gdb) PASS: gdb.fortran/oop-extend-type.exp: whatis wp
>>>>>         2 (gdb) PASS: gdb.fortran/oop-extend-type.exp: whatis wp_vla
>>>>>
>>>>>
>>>> Thx for the link. I will fix this.
>>>>
>>>> Pawel
>>>>
>>>>> Many thanks,
>>>>>
>>>>> Rich.
>>>>>