Remove DW_STRING and DW_STRING_IS_CANONICAL

This removes DW_STRING and DW_STRING_IS_CANONICAL, replacing them with
accessor methods on struct attribute.  The new code ensures that a
string value will only ever be used when the form allows it.

gdb/ChangeLog
2020-09-29  Tom Tromey  <tom@tromey.com>

	* dwarf2/read.c (read_cutu_die_from_dwo)
	(read_attribute_reprocess, read_attribute_value, read_attribute)
	(dwarf2_const_value_attr, dwarf2_name, dump_die_shallow)
	(dwarf2_fetch_constant_bytes): Update.
	* dwarf2/attribute.h (struct attribute) <form_is_string>: Declare.
	<set_string_noncanonical, set_string_canonical>: New methods.
	<string_is_canonical>: Update comment.
	<canonical_string_p>: Add assert.
	(DW_STRING, DW_STRING_IS_CANONICAL): Remove.
	* dwarf2/attribute.c (attribute::form_is_string): New method.
	(attribute::string): Use it.

3 files changed, 67 insertions, 49 deletions

diff --git a/gdb/dwarf2/attribute.c b/gdb/dwarf2/attribute.c
index af24d66..3fe1f08 100644
--- a/gdb/dwarf2/attribute.c
+++ b/gdb/dwarf2/attribute.c
@@ -61,18 +61,26 @@ attribute::as_address () const
 
 /* See attribute.h.  */
 
+bool
+attribute::form_is_string () const
+{
+  return (form == DW_FORM_strp || form == DW_FORM_line_strp
+	  || form == DW_FORM_string
+	  || form == DW_FORM_strx
+	  || form == DW_FORM_strx1
+	  || form == DW_FORM_strx2
+	  || form == DW_FORM_strx3
+	  || form == DW_FORM_strx4
+	  || form == DW_FORM_GNU_str_index
+	  || form == DW_FORM_GNU_strp_alt);
+}
+
+/* See attribute.h.  */
+
 const char *
 attribute::as_string () const
 {
-  if (form == DW_FORM_strp || form == DW_FORM_line_strp
-      || form == DW_FORM_string
-      || form == DW_FORM_strx
-      || form == DW_FORM_strx1
-      || form == DW_FORM_strx2
-      || form == DW_FORM_strx3
-      || form == DW_FORM_strx4
-      || form == DW_FORM_GNU_str_index
-      || form == DW_FORM_GNU_strp_alt)
+  if (form_is_string ())
     return u.str;
   return nullptr;
 }
diff --git a/gdb/dwarf2/attribute.h b/gdb/dwarf2/attribute.h
index a2c41f2..a9b77c1 100644
--- a/gdb/dwarf2/attribute.h
+++ b/gdb/dwarf2/attribute.h
@@ -98,6 +98,9 @@ struct attribute
 
   bool form_is_block () const;
 
+  /* Check if the attribute's form is a string form.  */
+  bool form_is_string () const;
+
   /* Return DIE offset of this attribute.  Return 0 with complaint if
      the attribute is not of the required kind.  */
 
@@ -122,16 +125,34 @@ struct attribute
      flag indicates whether the value has been canonicalized.  */
   bool canonical_string_p () const
   {
+    gdb_assert (form_is_string ());
     return string_is_canonical;
   }
 
+  /* Initialize this attribute to hold a non-canonical string
+     value.  */
+  void set_string_noncanonical (const char *str)
+  {
+    gdb_assert (form_is_string ());
+    u.str = str;
+    string_is_canonical = 0;
+  }
+
+  /* Set the canonical string value for this attribute.  */
+  void set_string_canonical (const char *str)
+  {
+    gdb_assert (form_is_string ());
+    u.str = str;
+    string_is_canonical = 1;
+  }
+
 
   ENUM_BITFIELD(dwarf_attribute) name : 16;
   ENUM_BITFIELD(dwarf_form) form : 15;
 
-  /* Has DW_STRING already been updated by dwarf2_canonicalize_name?  This
-     field should be in u.str (existing only for DW_STRING) but it is kept
-     here for better struct attribute alignment.  */
+  /* Has u.str already been updated by dwarf2_canonicalize_name?  This
+     field should be in u.str but it is kept here for better struct
+     attribute alignment.  */
   unsigned int string_is_canonical : 1;
 
   union
@@ -154,8 +175,6 @@ private:
 
 /* Get at parts of an attribute structure.  */
 
-#define DW_STRING(attr)    ((attr)->u.str)
-#define DW_STRING_IS_CANONICAL(attr) ((attr)->string_is_canonical)
 #define DW_UNSND(attr)     ((attr)->u.unsnd)
 #define DW_BLOCK(attr)     ((attr)->u.blk)
 #define DW_SND(attr)       ((attr)->u.snd)
diff --git a/gdb/dwarf2/read.c b/gdb/dwarf2/read.c
index f090cb3..fa78be1 100644
--- a/gdb/dwarf2/read.c
+++ b/gdb/dwarf2/read.c
@@ -6872,8 +6872,7 @@ read_cutu_die_from_dwo (dwarf2_cu *cu,
       comp_dir = XOBNEW (&cu->comp_unit_obstack, struct attribute);
       comp_dir->name = DW_AT_comp_dir;
       comp_dir->form = DW_FORM_string;
-      DW_STRING_IS_CANONICAL (comp_dir) = 0;
-      DW_STRING (comp_dir) = stub_comp_dir;
+      comp_dir->set_string_noncanonical (stub_comp_dir);
     }
 
   /* Set up for reading the DWO CU/TU.  */
@@ -19641,16 +19640,13 @@ read_attribute_reprocess (const struct die_reader_specs *reader,
       case DW_FORM_GNU_str_index:
 	{
 	  unsigned int str_index = DW_UNSND (attr);
+	  gdb_assert (!attr->canonical_string_p ());
 	  if (reader->dwo_file != NULL)
-	    {
-	      DW_STRING (attr) = read_dwo_str_index (reader, str_index);
-	      DW_STRING_IS_CANONICAL (attr) = 0;
-	    }
+	    attr->set_string_noncanonical (read_dwo_str_index (reader,
+							       str_index));
 	  else
-	    {
-	      DW_STRING (attr) = read_stub_str_index (cu, str_index);
-	      DW_STRING_IS_CANONICAL (attr) = 0;
-	    }
+	    attr->set_string_noncanonical (read_stub_str_index (cu,
+								str_index));
 	  break;
 	}
       default:
@@ -19746,17 +19742,17 @@ read_attribute_value (const struct die_reader_specs *reader,
       }
       break;
     case DW_FORM_string:
-      DW_STRING (attr) = read_direct_string (abfd, info_ptr, &bytes_read);
-      DW_STRING_IS_CANONICAL (attr) = 0;
+      attr->set_string_noncanonical (read_direct_string (abfd, info_ptr,
+							 &bytes_read));
       info_ptr += bytes_read;
       break;
     case DW_FORM_strp:
       if (!cu->per_cu->is_dwz)
 	{
-	  DW_STRING (attr) = read_indirect_string (per_objfile,
-						   abfd, info_ptr, cu_header,
-						   &bytes_read);
-	  DW_STRING_IS_CANONICAL (attr) = 0;
+	  attr->set_string_noncanonical
+	    (read_indirect_string (per_objfile,
+				   abfd, info_ptr, cu_header,
+				   &bytes_read));
 	  info_ptr += bytes_read;
 	  break;
 	}
@@ -19764,9 +19760,9 @@ read_attribute_value (const struct die_reader_specs *reader,
     case DW_FORM_line_strp:
       if (!cu->per_cu->is_dwz)
 	{
-	  DW_STRING (attr) = per_objfile->read_line_string (info_ptr, cu_header,
-							    &bytes_read);
-	  DW_STRING_IS_CANONICAL (attr) = 0;
+	  attr->set_string_noncanonical
+	    (per_objfile->read_line_string (info_ptr, cu_header,
+					    &bytes_read));
 	  info_ptr += bytes_read;
 	  break;
 	}
@@ -19777,8 +19773,8 @@ read_attribute_value (const struct die_reader_specs *reader,
 	LONGEST str_offset = cu_header->read_offset (abfd, info_ptr,
 						     &bytes_read);
 
-	DW_STRING (attr) = dwz->read_string (objfile, str_offset);
-	DW_STRING_IS_CANONICAL (attr) = 0;
+	attr->set_string_noncanonical
+	  (dwz->read_string (objfile, str_offset));
 	info_ptr += bytes_read;
       }
       break;
@@ -19944,6 +19940,7 @@ read_attribute (const struct die_reader_specs *reader,
 		const gdb_byte *info_ptr, bool *need_reprocess)
 {
   attr->name = abbrev->name;
+  attr->string_is_canonical = 0;
   return read_attribute_value (reader, attr, abbrev->form,
 			       abbrev->implicit_const, info_ptr,
 			       need_reprocess);
@@ -21861,7 +21858,7 @@ dwarf2_const_value_attr (const struct attribute *attr, struct type *type,
     case DW_FORM_strx:
     case DW_FORM_GNU_str_index:
     case DW_FORM_GNU_strp_alt:
-      /* DW_STRING is already allocated on the objfile obstack, point
+      /* The string is already allocated on the objfile obstack, point
 	 directly to it.  */
       *bytes = (const gdb_byte *) attr->as_string ();
       break;
@@ -22645,8 +22642,7 @@ dwarf2_name (struct die_info *die, struct dwarf2_cu *cu)
 	      if (demangled == nullptr)
 		return nullptr;
 
-	      DW_STRING (attr) = objfile->intern (demangled.get ());
-	      DW_STRING_IS_CANONICAL (attr) = 1;
+	      attr->set_string_canonical (objfile->intern (demangled.get ()));
 	    }
 
 	  /* Strip any leading namespaces/classes, keep only the
@@ -22665,13 +22661,8 @@ dwarf2_name (struct die_info *die, struct dwarf2_cu *cu)
     }
 
   if (!attr->canonical_string_p ())
-    {
-      DW_STRING (attr) = dwarf2_canonicalize_name (attr_name, cu,
-						   objfile);
-      DW_STRING_IS_CANONICAL (attr) = 1;
-    }
-
-  /* We might have changed it just above.  */
+    attr->set_string_canonical (dwarf2_canonicalize_name (attr_name, cu,
+							  objfile));
   return attr->as_string ();
 }
 
@@ -22784,9 +22775,9 @@ dump_die_shallow (struct ui_file *f, int indent, struct die_info *die)
 	case DW_FORM_GNU_str_index:
 	case DW_FORM_GNU_strp_alt:
 	  fprintf_unfiltered (f, "string: \"%s\" (%s canonicalized)",
-		   DW_STRING (&die->attrs[i])
-		   ? DW_STRING (&die->attrs[i]) : "",
-		   die->attrs[i].canonical_string_p () ? "is" : "not");
+			      die->attrs[i].as_string ()
+			      ? die->attrs[i].as_string () : "",
+			      die->attrs[i].canonical_string_p () ? "is" : "not");
 	  break;
 	case DW_FORM_flag:
 	  if (DW_UNSND (&die->attrs[i]))
@@ -23180,7 +23171,7 @@ dwarf2_fetch_constant_bytes (sect_offset sect_off,
     case DW_FORM_strx:
     case DW_FORM_GNU_str_index:
     case DW_FORM_GNU_strp_alt:
-      /* DW_STRING is already allocated on the objfile obstack, point
+      /* The string is already allocated on the objfile obstack, point
 	 directly to it.  */
       {
 	const char *attr_name = attr->as_string ();