[RFA,3/4] Remove TYPE_TAG_NAME

Message ID 20180417195125.14200-4-tom@tromey.com
State New
Headers show
Series
  • Remove TYPE_TAG_NAME
Related show

Commit Message

Tom Tromey April 17, 2018, 7:51 p.m.
TYPE_TAG_NAME has been an occasional source of confusion and bugs.  It
seems to me that it is only useful for C and C++ -- but even there,
not so much, because at least with DWARF there doesn't seem to be any
way to wind up with a type where the name and the tag name are both
non-NULL and different.

So, this patch removes TYPE_TAG_NAME entirely.  This should save a
little memory, but more importantly, it simplifies this part of gdb.

A few minor test suite adjustments were needed.  In some situations
the new code does not yield identical output to the old code.

ChangeLog
2018-04-17  Tom Tromey  <tom@tromey.com>

	* valops.c (enum_constant_from_type, value_namespace_elt)
	(value_maybe_namespace_elt): Update.
	* valarith.c (find_size_for_pointer_math): Update.
	* target-descriptions.c (make_gdb_type): Update.
	* symmisc.c (print_symbol): Update.
	* stabsread.c (define_symbol, read_type)
	(complain_about_struct_wipeout, add_undefined_type)
	(cleanup_undefined_types_1): Update.
	* rust-lang.c (rust_tuple_type_p, rust_slice_type_p)
	(rust_range_type_p, val_print_struct, rust_print_struct_def)
	(rust_internal_print_type, rust_composite_type)
	(rust_evaluate_funcall, rust_evaluate_subexp): Update.
	* python/py-type.c (typy_get_tag): Update.
	* p-typeprint.c (pascal_type_print_base): Update.
	* mdebugread.c (parse_symbol, parse_type): Update.
	* m2-typeprint.c (m2_long_set, m2_record_fields, m2_enum):
	Update.
	* guile/scm-type.c (gdbscm_type_tag): Update.
	* go-lang.c (sixg_string_p): Update.
	* gnu-v3-abi.c (build_gdb_vtable_type, build_std_type_info_type):
	Update.
	* gdbtypes.h (struct main_type) <tag_name>: Remove.
	(TYPE_TAG_NAME): Remove.
	* gdbtypes.c (type_name_no_tag): Simplify.
	(check_typedef, check_types_equal, recursive_dump_type)
	(copy_type_recursive, arch_composite_type): Update.
	* f-typeprint.c (f_type_print_base): Update.  Print "Type" prefix
	in summary mode when needed.
	* eval.c (evaluate_funcall): Update.
	* dwarf2read.c (fixup_go_packaging, read_structure_type)
	(process_structure_scope, read_enumeration_type)
	(read_namespace_type, read_module_type, determine_prefix): Update.
	* cp-support.c (inspect_type): Update.
	* coffread.c (process_coff_symbol, decode_base_type): Update.
	* c-varobj.c (c_is_path_expr_parent): Update.
	* c-typeprint.c (c_type_print_base_struct_union): Update.
	(c_type_print_base_1): Update.  Print struct/class/union/enum in
	summary when using C language.
	* ax-gdb.c (gen_struct_ref, gen_namespace_elt)
	(gen_maybe_namespace_elt): Update.
	* ada-lang.c (ada_type_name): Simplify.
	(empty_record, ada_template_to_fixed_record_type_1)
	(template_to_static_fixed_type)
	(to_record_with_fixed_variant_part, ada_check_typedef): Update.

testsuite/ChangeLog
2018-04-17  Tom Tromey  <tom@tromey.com>

	* gdb.xml/tdesc-regs.exp (load_description): Update expected
	results.
	* gdb.dwarf2/method-ptr.exp: Set language to C++.
	* gdb.dwarf2/member-ptr-forwardref.exp: Set language to C++.
	* gdb.cp/typeid.exp (do_typeid_tests): Update type_re.
	* gdb.base/maint.exp (maint_pass_if): Update.
---
 gdb/ChangeLog                                      | 47 +++++++++++++++++++
 gdb/ada-lang.c                                     | 17 ++-----
 gdb/ax-gdb.c                                       |  6 +--
 gdb/c-typeprint.c                                  | 46 +++++++++++++------
 gdb/c-varobj.c                                     |  3 +-
 gdb/coffread.c                                     | 16 +------
 gdb/cp-support.c                                   |  2 +-
 gdb/dwarf2read.c                                   | 31 ++++---------
 gdb/eval.c                                         |  4 +-
 gdb/f-typeprint.c                                  | 13 ++++--
 gdb/gdbtypes.c                                     | 29 ++----------
 gdb/gdbtypes.h                                     | 21 ++-------
 gdb/gnu-v3-abi.c                                   |  4 +-
 gdb/go-lang.c                                      |  4 +-
 gdb/guile/scm-type.c                               | 10 +++-
 gdb/m2-typeprint.c                                 | 22 +++------
 gdb/mdebugread.c                                   | 14 +++---
 gdb/p-typeprint.c                                  | 20 ++++----
 gdb/python/py-type.c                               | 10 +++-
 gdb/rust-lang.c                                    | 53 +++++++++++-----------
 gdb/stabsread.c                                    | 20 +++-----
 gdb/symmisc.c                                      |  2 +-
 gdb/target-descriptions.c                          |  1 -
 gdb/testsuite/ChangeLog                            |  9 ++++
 gdb/testsuite/gdb.base/maint.exp                   |  2 +-
 gdb/testsuite/gdb.cp/typeid.exp                    |  2 +-
 gdb/testsuite/gdb.dwarf2/member-ptr-forwardref.exp |  1 +
 gdb/testsuite/gdb.dwarf2/method-ptr.exp            |  1 +
 gdb/testsuite/gdb.xml/tdesc-regs.exp               |  4 +-
 gdb/valarith.c                                     |  2 -
 gdb/valops.c                                       |  6 +--
 31 files changed, 215 insertions(+), 207 deletions(-)

-- 
2.13.6

Comments

Keith Seitz April 18, 2018, 4:57 p.m. | #1
On 04/17/2018 12:51 PM, Tom Tromey wrote:
> ChangeLog

> 2018-04-17  Tom Tromey  <tom@tromey.com>

> 

> 	* valops.c (enum_constant_from_type, value_namespace_elt)

> 	(value_maybe_namespace_elt): Update.

> 	* valarith.c (find_size_for_pointer_math): Update.

> 	* target-descriptions.c (make_gdb_type): Update.

> 	* symmisc.c (print_symbol): Update.

> 	* stabsread.c (define_symbol, read_type)

> 	(complain_about_struct_wipeout, add_undefined_type)

> 	(cleanup_undefined_types_1): Update.

> 	* rust-lang.c (rust_tuple_type_p, rust_slice_type_p)

> 	(rust_range_type_p, val_print_struct, rust_print_struct_def)

> 	(rust_internal_print_type, rust_composite_type)

> 	(rust_evaluate_funcall, rust_evaluate_subexp): Update.

> 	* python/py-type.c (typy_get_tag): Update.

> 	* p-typeprint.c (pascal_type_print_base): Update.

> 	* mdebugread.c (parse_symbol, parse_type): Update.

> 	* m2-typeprint.c (m2_long_set, m2_record_fields, m2_enum):

> 	Update.

> 	* guile/scm-type.c (gdbscm_type_tag): Update.

> 	* go-lang.c (sixg_string_p): Update.

> 	* gnu-v3-abi.c (build_gdb_vtable_type, build_std_type_info_type):

> 	Update.

> 	* gdbtypes.h (struct main_type) <tag_name>: Remove.

> 	(TYPE_TAG_NAME): Remove.

> 	* gdbtypes.c (type_name_no_tag): Simplify.

> 	(check_typedef, check_types_equal, recursive_dump_type)

> 	(copy_type_recursive, arch_composite_type): Update.

> 	* f-typeprint.c (f_type_print_base): Update.  Print "Type" prefix

> 	in summary mode when needed.

> 	* eval.c (evaluate_funcall): Update.

> 	* dwarf2read.c (fixup_go_packaging, read_structure_type)

> 	(process_structure_scope, read_enumeration_type)

> 	(read_namespace_type, read_module_type, determine_prefix): Update.

> 	* cp-support.c (inspect_type): Update.

> 	* coffread.c (process_coff_symbol, decode_base_type): Update.

> 	* c-varobj.c (c_is_path_expr_parent): Update.

> 	* c-typeprint.c (c_type_print_base_struct_union): Update.

> 	(c_type_print_base_1): Update.  Print struct/class/union/enum in

> 	summary when using C language.

> 	* ax-gdb.c (gen_struct_ref, gen_namespace_elt)

> 	(gen_maybe_namespace_elt): Update.

> 	* ada-lang.c (ada_type_name): Simplify.

> 	(empty_record, ada_template_to_fixed_record_type_1)

> 	(template_to_static_fixed_type)

> 	(to_record_with_fixed_variant_part, ada_check_typedef): Update.

> 

> testsuite/ChangeLog

> 2018-04-17  Tom Tromey  <tom@tromey.com>

> 

> 	* gdb.xml/tdesc-regs.exp (load_description): Update expected

> 	results.

> 	* gdb.dwarf2/method-ptr.exp: Set language to C++.

> 	* gdb.dwarf2/member-ptr-forwardref.exp: Set language to C++.

> 	* gdb.cp/typeid.exp (do_typeid_tests): Update type_re.

> 	* gdb.base/maint.exp (maint_pass_if): Update.


Oh, ChangeLogs... Once again I wonder if "update all callers" would have saved you quite a bit of busywork?

> diff --git a/gdb/c-typeprint.c b/gdb/c-typeprint.c

> index a53331aefc..846727f298 100644

> --- a/gdb/c-typeprint.c

> +++ b/gdb/c-typeprint.c

> @@ -1602,15 +1602,31 @@ c_type_print_base_1 (struct type *type, struct ui_file *stream,

>  

>    /* When SHOW is zero or less, and there is a valid type name, then

>       always just print the type name directly from the type.  */

> -  /* If we have "typedef struct foo {. . .} bar;" do we want to print

> -     it as "struct foo" or as "bar"?  Pick the latter, because C++

> -     folk tend to expect things like "class5 *foo" rather than "struct

> -     class5 *foo".  */

>  

>    if (show <= 0

>        && TYPE_NAME (type) != NULL)

>      {

>        c_type_print_modifier (type, stream, 0, 1);

> +

> +      /* If we have "typedef struct foo {. . .} bar;" do we want to

> +	 print it as "struct foo" or as "bar"?  Pick the latter for

> +	 C++, because C++ folk tend to expect things like "class5

> +	 *foo" rather than "struct class5 *foo".  */

> +      if (language == language_c || language == language_minimal)

> +	{

> +	  if (TYPE_CODE (type) == TYPE_CODE_UNION)

> +	    fprintf_filtered (stream, "union ");

> +	  else if (TYPE_CODE (type) == TYPE_CODE_STRUCT)

> +	    {

> +	      if (TYPE_DECLARED_CLASS (type))

> +		fprintf_filtered (stream, "class ");

> +	      else

> +		fprintf_filtered (stream, "struct ");

> +	    }

> +	  else if (TYPE_CODE (type) == TYPE_CODE_ENUM)

> +	    fprintf_filtered (stream, "enum ");

> +	}

> +

>        print_name_maybe_canonical (TYPE_NAME (type), flags, stream);

>        return;

>      }


I'm almost afraid to ask, but why was language_minimal necessary here? A small comment might be appropriate?
[I think I can already guess the heinous reason...] Do you know if there is a test case that specifically covers this block with language_minimal?

> diff --git a/gdb/gdbtypes.h b/gdb/gdbtypes.h

> index 552a2a2a16..70dac75855 100644

> --- a/gdb/gdbtypes.h

> +++ b/gdb/gdbtypes.h

> @@ -721,26 +721,12 @@ struct main_type

>  

>    /* * Name of this type, or NULL if none.

>  

> -     This is used for printing only, except by poorly designed C++

> -     code.  For looking up a name, look for a symbol in the

> -     VAR_DOMAIN.  This is generally allocated in the objfile's

> -     obstack.  However coffread.c uses malloc.  */

> +     This is used for printing only.  For looking up a name, look for

> +     a symbol in the VAR_DOMAIN.  This is generally allocated in the

> +     objfile's obstack.  However coffread.c uses malloc.  */


Good riddance! I never did understand the "except by poorly designed C++ code" comment.

> diff --git a/gdb/cp-support.c b/gdb/cp-support.c

> index dde417be80..c16b67171f 100644

> --- a/gdb/cp-support.c

> +++ b/gdb/cp-support.c

> @@ -199,7 +199,7 @@ inspect_type (struct demangle_parse_info *info,

>  	      && strcmp (TYPE_NAME (type), name) == 0)

>  	    return 0;

>  

> -	  is_anon = (TYPE_TAG_NAME (type) == NULL

> +	  is_anon = (TYPE_NAME (type) == NULL

>  		     && (TYPE_CODE (type) == TYPE_CODE_ENUM

>  			 || TYPE_CODE (type) == TYPE_CODE_STRUCT

>  			 || TYPE_CODE (type) == TYPE_CODE_UNION));

> diff --git a/gdb/guile/scm-type.c b/gdb/guile/scm-type.c

> index ce67695b76..949e0dca82 100644

> --- a/gdb/guile/scm-type.c

> +++ b/gdb/guile/scm-type.c

> @@ -576,10 +576,16 @@ gdbscm_type_tag (SCM self)

>    type_smob *t_smob

>      = tyscm_get_type_smob_arg_unsafe (self, SCM_ARG1, FUNC_NAME);

>    struct type *type = t_smob->type;

> +  const char *tagname = nullptr;

>  

> -  if (!TYPE_TAG_NAME (type))

> +  if (TYPE_CODE (type) == TYPE_CODE_STRUCT

> +      || TYPE_CODE (type) == TYPE_CODE_UNION

> +      || TYPE_CODE (type) == TYPE_CODE_ENUM)

> +    tagname = TYPE_NAME (type);

> +

> +  if (tagname == nullptr)

>      return SCM_BOOL_F;

> -  return gdbscm_scm_from_c_string (TYPE_TAG_NAME (type));

> +  return gdbscm_scm_from_c_string (tagname);

>  }

>  

>  /* (type-name <gdb:type>) -> string

> diff --git a/gdb/python/py-type.c b/gdb/python/py-type.c

> index 8a948eeaa6..6cbb1b663d 100644

> --- a/gdb/python/py-type.c

> +++ b/gdb/python/py-type.c

> @@ -416,10 +416,16 @@ static PyObject *

>  typy_get_tag (PyObject *self, void *closure)

>  {

>    struct type *type = ((type_object *) self)->type;

> +  const char *tagname = nullptr;

>  

> -  if (!TYPE_TAG_NAME (type))

> +  if (TYPE_CODE (type) == TYPE_CODE_STRUCT

> +      || TYPE_CODE (type) == TYPE_CODE_UNION

> +      || TYPE_CODE (type) == TYPE_CODE_ENUM)

> +    tagname = TYPE_NAME (type);

> +

> +  if (tagname == nullptr)

>      Py_RETURN_NONE;

> -  return PyString_FromString (TYPE_TAG_NAME (type));

> +  return PyString_FromString (tagname);

>  }

>  

>  /* Return the type, stripped of typedefs. */


The above three code snippets suggest the introduction of a type_tag_name (struct type *) convenience function might be justified, but I am not going to request any such change.

> diff --git a/gdb/testsuite/gdb.xml/tdesc-regs.exp b/gdb/testsuite/gdb.xml/tdesc-regs.exp

> index bb7b9be16c..94d4007a89 100644

> --- a/gdb/testsuite/gdb.xml/tdesc-regs.exp

> +++ b/gdb/testsuite/gdb.xml/tdesc-regs.exp

> @@ -179,7 +179,7 @@ gdb_test "ptype \$extrareg" "type = (int|long|long long)"

>  gdb_test "ptype \$uintreg" "type = uint32_t"

>  gdb_test "ptype \$vecreg" "type = int8_t __attribute__ \\(\\(vector_size\\(4\\)\\)\\)"

>  gdb_test "ptype \$unionreg" \

> -    "type = union {\r\n *v4int8 v4;\r\n *v2int16 v2;\r\n}"

> +    "type = union vecint {\r\n *v4int8 v4;\r\n *v2int16 v2;\r\n}"

>  gdb_test "ptype \$unionreg.v4" "type = int8_t __attribute__ \\(\\(vector_size\\(4\\)\\)\\)"

>  gdb_test "ptype \$structreg" \

>      "type = struct struct1 {\r\n *v4int8 v4;\r\n *v2int16 v2;\r\n}"


<rhetorical>Was the tag name never printed until now?</rhetorical> Wow!

> @@ -189,7 +189,7 @@ gdb_test "ptype \$bitfields" \

>  gdb_test "ptype \$flags" \

>      "type = flag flags {\r\n *bool X @0;\r\n *uint32_t Y @2;\r\n}"

>  gdb_test "ptype \$mixed_flags" \

> -    "type = flag mixed_flags {\r\n *bool A @0;\r\n *uint32_t B @1-3;\r\n *bool C @4;\r\n *uint32_t D @5;\r\n *uint32_t @6-7;\r\n *enum {yes = 1, no = 0, maybe = 2, so} Z @8-9;\r\n}"

> +    "type = flag mixed_flags {\r\n *bool A @0;\r\n *uint32_t B @1-3;\r\n *bool C @4;\r\n *uint32_t D @5;\r\n *uint32_t @6-7;\r\n *enum Z_values {yes = 1, no = 0, maybe = 2, so} Z @8-9;\r\n}"

>  # Reggroups should have at least general and the extra foo group

>  gdb_test "maintenance print reggroups" \

>      " Group\[ \t\]+Type\[ \t\]+\r\n.* general\[ \t\]+user\[ \t\]+\r\n.* foo\[ \t\]+user\[ \t\]+"


Warning: I am not a maintainer, ... You know the rest of that drill. :-)

Keith
Tom Tromey April 18, 2018, 5:57 p.m. | #2
>>>>> "Keith" == Keith Seitz <keiths@redhat.com> writes:


>> +      /* If we have "typedef struct foo {. . .} bar;" do we want to

>> +	 print it as "struct foo" or as "bar"?  Pick the latter for

>> +	 C++, because C++ folk tend to expect things like "class5

>> +	 *foo" rather than "struct class5 *foo".  */

>> +      if (language == language_c || language == language_minimal)


Keith> I'm almost afraid to ask, but why was language_minimal necessary here?
Keith> A small comment might be appropriate?
Keith> [I think I can already guess the heinous reason...] Do you know if
Keith> there is a test case that specifically covers this block with
Keith> language_minimal?

I think in this case, I added language_minimal on the theory that it is
"C-like".  I don't think there's a test case.

I can add a comment, and a test case wouldn't be too hard either.

Keith> Good riddance! I never did understand the "except by poorly
Keith> designed C++ code" comment.

Haha, me neither.  I always considered it a critique of some other code
elsewhere in gdb (perhaps long since removed) but who knows.

>> gdb_test "ptype \$structreg" \

>> "type = struct struct1 {\r\n *v4int8 v4;\r\n *v2int16 v2;\r\n}"


Keith> <rhetorical>Was the tag name never printed until now?</rhetorical> Wow!

Maybe I should have investigated this one more deeply?
I dunno.  The change seemed to make sense to me so I just moved on.

Tom
Keith Seitz April 18, 2018, 6:06 p.m. | #3
On 04/18/2018 10:57 AM, Tom Tromey wrote:
>>>>>> "Keith" == Keith Seitz <keiths@redhat.com> writes:

> 

>>> +      /* If we have "typedef struct foo {. . .} bar;" do we want to

>>> +	 print it as "struct foo" or as "bar"?  Pick the latter for

>>> +	 C++, because C++ folk tend to expect things like "class5

>>> +	 *foo" rather than "struct class5 *foo".  */

>>> +      if (language == language_c || language == language_minimal)

> 

> Keith> I'm almost afraid to ask, but why was language_minimal necessary here?

> Keith> A small comment might be appropriate?

> Keith> [I think I can already guess the heinous reason...] Do you know if

> Keith> there is a test case that specifically covers this block with

> Keith> language_minimal?

> 

> I think in this case, I added language_minimal on the theory that it is

> "C-like".  I don't think there's a test case.

> 

> I can add a comment, and a test case wouldn't be too hard either.


Seeing "language_minimal" caught me off guard. A comment would more than satisfy me.

> 

>>> gdb_test "ptype \$structreg" \

>>> "type = struct struct1 {\r\n *v4int8 v4;\r\n *v2int16 v2;\r\n}"

> 

> Keith> <rhetorical>Was the tag name never printed until now?</rhetorical> Wow!

> 

> Maybe I should have investigated this one more deeply?

> I dunno.  The change seemed to make sense to me so I just moved on.


I agree. My rhetorical "question" was really more a statement of disbelief. Things like this seem to catch me by surprise every so often -- sometimes more often than not.

Keith
Joel Brobecker May 31, 2018, 1:33 a.m. | #4
> TYPE_TAG_NAME has been an occasional source of confusion and bugs.  It

> seems to me that it is only useful for C and C++ -- but even there,

> not so much, because at least with DWARF there doesn't seem to be any

> way to wind up with a type where the name and the tag name are both

> non-NULL and different.

> 

> So, this patch removes TYPE_TAG_NAME entirely.  This should save a

> little memory, but more importantly, it simplifies this part of gdb.

> 

> A few minor test suite adjustments were needed.  In some situations

> the new code does not yield identical output to the old code.

> 

> ChangeLog

> 2018-04-17  Tom Tromey  <tom@tromey.com>

> 

> 	* valops.c (enum_constant_from_type, value_namespace_elt)

> 	(value_maybe_namespace_elt): Update.

> 	* valarith.c (find_size_for_pointer_math): Update.

> 	* target-descriptions.c (make_gdb_type): Update.

> 	* symmisc.c (print_symbol): Update.

> 	* stabsread.c (define_symbol, read_type)

> 	(complain_about_struct_wipeout, add_undefined_type)

> 	(cleanup_undefined_types_1): Update.

> 	* rust-lang.c (rust_tuple_type_p, rust_slice_type_p)

> 	(rust_range_type_p, val_print_struct, rust_print_struct_def)

> 	(rust_internal_print_type, rust_composite_type)

> 	(rust_evaluate_funcall, rust_evaluate_subexp): Update.

> 	* python/py-type.c (typy_get_tag): Update.

> 	* p-typeprint.c (pascal_type_print_base): Update.

> 	* mdebugread.c (parse_symbol, parse_type): Update.

> 	* m2-typeprint.c (m2_long_set, m2_record_fields, m2_enum):

> 	Update.

> 	* guile/scm-type.c (gdbscm_type_tag): Update.

> 	* go-lang.c (sixg_string_p): Update.

> 	* gnu-v3-abi.c (build_gdb_vtable_type, build_std_type_info_type):

> 	Update.

> 	* gdbtypes.h (struct main_type) <tag_name>: Remove.

> 	(TYPE_TAG_NAME): Remove.

> 	* gdbtypes.c (type_name_no_tag): Simplify.

> 	(check_typedef, check_types_equal, recursive_dump_type)

> 	(copy_type_recursive, arch_composite_type): Update.

> 	* f-typeprint.c (f_type_print_base): Update.  Print "Type" prefix

> 	in summary mode when needed.

> 	* eval.c (evaluate_funcall): Update.

> 	* dwarf2read.c (fixup_go_packaging, read_structure_type)

> 	(process_structure_scope, read_enumeration_type)

> 	(read_namespace_type, read_module_type, determine_prefix): Update.

> 	* cp-support.c (inspect_type): Update.

> 	* coffread.c (process_coff_symbol, decode_base_type): Update.

> 	* c-varobj.c (c_is_path_expr_parent): Update.

> 	* c-typeprint.c (c_type_print_base_struct_union): Update.

> 	(c_type_print_base_1): Update.  Print struct/class/union/enum in

> 	summary when using C language.

> 	* ax-gdb.c (gen_struct_ref, gen_namespace_elt)

> 	(gen_maybe_namespace_elt): Update.

> 	* ada-lang.c (ada_type_name): Simplify.

> 	(empty_record, ada_template_to_fixed_record_type_1)

> 	(template_to_static_fixed_type)

> 	(to_record_with_fixed_variant_part, ada_check_typedef): Update.

> 

> testsuite/ChangeLog

> 2018-04-17  Tom Tromey  <tom@tromey.com>

> 

> 	* gdb.xml/tdesc-regs.exp (load_description): Update expected

> 	results.

> 	* gdb.dwarf2/method-ptr.exp: Set language to C++.

> 	* gdb.dwarf2/member-ptr-forwardref.exp: Set language to C++.

> 	* gdb.cp/typeid.exp (do_typeid_tests): Update type_re.

> 	* gdb.base/maint.exp (maint_pass_if): Update.


This looks good to me. Perhaps add the comment that Keith mentioned
about the use of language_minimal in c_type_print_base_1. Sometimes,
I wonder whether we might have over-engineered that thing.

Thanks!
-- 
Joel

Patch

diff --git a/gdb/ada-lang.c b/gdb/ada-lang.c
index de20c43bed..41235def3e 100644
--- a/gdb/ada-lang.c
+++ b/gdb/ada-lang.c
@@ -8136,18 +8136,15 @@  ada_prefer_type (struct type *type0, struct type *type1)
   return 0;
 }
 
-/* The name of TYPE, which is either its TYPE_NAME, or, if that is
-   null, its TYPE_TAG_NAME.  Null if TYPE is null.  */
+/* The name of TYPE, which is its TYPE_NAME.  Null if TYPE is
+   null.  */
 
 const char *
 ada_type_name (struct type *type)
 {
   if (type == NULL)
     return NULL;
-  else if (TYPE_NAME (type) != NULL)
-    return TYPE_NAME (type);
-  else
-    return TYPE_TAG_NAME (type);
+  return TYPE_NAME (type);
 }
 
 /* Search the list of "descriptive" types associated to TYPE for a type
@@ -8316,7 +8313,6 @@  empty_record (struct type *templ)
   TYPE_FIELDS (type) = NULL;
   INIT_CPLUS_SPECIFIC (type);
   TYPE_NAME (type) = "<empty>";
-  TYPE_TAG_NAME (type) = NULL;
   TYPE_LENGTH (type) = 0;
   return type;
 }
@@ -8375,7 +8371,6 @@  ada_template_to_fixed_record_type_1 (struct type *type,
     TYPE_ALLOC (rtype, nfields * sizeof (struct field));
   memset (TYPE_FIELDS (rtype), 0, sizeof (struct field) * nfields);
   TYPE_NAME (rtype) = ada_type_name (type);
-  TYPE_TAG_NAME (rtype) = NULL;
   TYPE_FIXED_INSTANCE (rtype) = 1;
 
   off = 0;
@@ -8652,7 +8647,6 @@  template_to_static_fixed_type (struct type *type0)
 	      memcpy (TYPE_FIELDS (type), TYPE_FIELDS (type0),
 		      sizeof (struct field) * nfields);
 	      TYPE_NAME (type) = ada_type_name (type0);
-	      TYPE_TAG_NAME (type) = NULL;
 	      TYPE_FIXED_INSTANCE (type) = 1;
 	      TYPE_LENGTH (type) = 0;
 	    }
@@ -8702,7 +8696,6 @@  to_record_with_fixed_variant_part (struct type *type, const gdb_byte *valaddr,
   memcpy (TYPE_FIELDS (rtype), TYPE_FIELDS (type),
           sizeof (struct field) * nfields);
   TYPE_NAME (rtype) = ada_type_name (type);
-  TYPE_TAG_NAME (rtype) = NULL;
   TYPE_FIXED_INSTANCE (rtype) = 1;
   TYPE_LENGTH (rtype) = TYPE_LENGTH (type);
 
@@ -9313,11 +9306,11 @@  ada_check_typedef (struct type *type)
   type = check_typedef (type);
   if (type == NULL || TYPE_CODE (type) != TYPE_CODE_ENUM
       || !TYPE_STUB (type)
-      || TYPE_TAG_NAME (type) == NULL)
+      || TYPE_NAME (type) == NULL)
     return type;
   else
     {
-      const char *name = TYPE_TAG_NAME (type);
+      const char *name = TYPE_NAME (type);
       struct type *type1 = ada_find_any_type (name);
 
       if (type1 == NULL)
diff --git a/gdb/ax-gdb.c b/gdb/ax-gdb.c
index 9a44d4b750..2468061aad 100644
--- a/gdb/ax-gdb.c
+++ b/gdb/ax-gdb.c
@@ -1533,7 +1533,7 @@  gen_struct_ref (struct agent_expr *ax, struct axs_value *value,
   
   if (!found)
     error (_("Couldn't find member named `%s' in struct/union/class `%s'"),
-	   field, TYPE_TAG_NAME (type));
+	   field, TYPE_NAME (type));
 }
 
 static int
@@ -1629,7 +1629,7 @@  gen_namespace_elt (struct agent_expr *ax, struct axs_value *value,
 
   if (!found)
     error (_("No symbol \"%s\" in namespace \"%s\"."), 
-	   name, TYPE_TAG_NAME (curtype));
+	   name, TYPE_NAME (curtype));
 
   return found;
 }
@@ -1644,7 +1644,7 @@  static int
 gen_maybe_namespace_elt (struct agent_expr *ax, struct axs_value *value,
 			 const struct type *curtype, char *name)
 {
-  const char *namespace_name = TYPE_TAG_NAME (curtype);
+  const char *namespace_name = TYPE_NAME (curtype);
   struct block_symbol sym;
 
   sym = cp_lookup_symbol_namespace (namespace_name, name,
diff --git a/gdb/c-typeprint.c b/gdb/c-typeprint.c
index a53331aefc..846727f298 100644
--- a/gdb/c-typeprint.c
+++ b/gdb/c-typeprint.c
@@ -1174,13 +1174,13 @@  c_type_print_base_struct_union (struct type *type, struct ui_file *stream,
      spurious "{unnamed struct}"/"{unnamed union}"/"{unnamed
      enum}" tag for unnamed struct/union/enum's, which we don't
      want to print.  */
-  if (TYPE_TAG_NAME (type) != NULL
-      && !startswith (TYPE_TAG_NAME (type), "{unnamed"))
+  if (TYPE_NAME (type) != NULL
+      && !startswith (TYPE_NAME (type), "{unnamed"))
     {
       /* When printing the tag name, we are still effectively
 	 printing in the outer context, hence the use of FLAGS
 	 here.  */
-      print_name_maybe_canonical (TYPE_TAG_NAME (type), flags, stream);
+      print_name_maybe_canonical (TYPE_NAME (type), flags, stream);
       if (show > 0)
 	fputs_filtered (" ", stream);
     }
@@ -1189,10 +1189,10 @@  c_type_print_base_struct_union (struct type *type, struct ui_file *stream,
     {
       /* If we just printed a tag name, no need to print anything
 	 else.  */
-      if (TYPE_TAG_NAME (type) == NULL)
+      if (TYPE_NAME (type) == NULL)
 	fprintf_filtered (stream, "{...}");
     }
-  else if (show > 0 || TYPE_TAG_NAME (type) == NULL)
+  else if (show > 0 || TYPE_NAME (type) == NULL)
     {
       struct type *basetype;
       int vptr_fieldno;
@@ -1602,15 +1602,31 @@  c_type_print_base_1 (struct type *type, struct ui_file *stream,
 
   /* When SHOW is zero or less, and there is a valid type name, then
      always just print the type name directly from the type.  */
-  /* If we have "typedef struct foo {. . .} bar;" do we want to print
-     it as "struct foo" or as "bar"?  Pick the latter, because C++
-     folk tend to expect things like "class5 *foo" rather than "struct
-     class5 *foo".  */
 
   if (show <= 0
       && TYPE_NAME (type) != NULL)
     {
       c_type_print_modifier (type, stream, 0, 1);
+
+      /* If we have "typedef struct foo {. . .} bar;" do we want to
+	 print it as "struct foo" or as "bar"?  Pick the latter for
+	 C++, because C++ folk tend to expect things like "class5
+	 *foo" rather than "struct class5 *foo".  */
+      if (language == language_c || language == language_minimal)
+	{
+	  if (TYPE_CODE (type) == TYPE_CODE_UNION)
+	    fprintf_filtered (stream, "union ");
+	  else if (TYPE_CODE (type) == TYPE_CODE_STRUCT)
+	    {
+	      if (TYPE_DECLARED_CLASS (type))
+		fprintf_filtered (stream, "class ");
+	      else
+		fprintf_filtered (stream, "struct ");
+	    }
+	  else if (TYPE_CODE (type) == TYPE_CODE_ENUM)
+	    fprintf_filtered (stream, "enum ");
+	}
+
       print_name_maybe_canonical (TYPE_NAME (type), flags, stream);
       return;
     }
@@ -1661,10 +1677,10 @@  c_type_print_base_1 (struct type *type, struct ui_file *stream,
          "{unnamed struct}"/"{unnamed union}"/"{unnamed enum}"
          tag for unnamed struct/union/enum's, which we don't
          want to print.  */
-      if (TYPE_TAG_NAME (type) != NULL
-	  && !startswith (TYPE_TAG_NAME (type), "{unnamed"))
+      if (TYPE_NAME (type) != NULL
+	  && !startswith (TYPE_NAME (type), "{unnamed"))
 	{
-	  print_name_maybe_canonical (TYPE_TAG_NAME (type), flags, stream);
+	  print_name_maybe_canonical (TYPE_NAME (type), flags, stream);
 	  if (show > 0)
 	    fputs_filtered (" ", stream);
 	}
@@ -1674,10 +1690,10 @@  c_type_print_base_1 (struct type *type, struct ui_file *stream,
 	{
 	  /* If we just printed a tag name, no need to print anything
 	     else.  */
-	  if (TYPE_TAG_NAME (type) == NULL)
+	  if (TYPE_NAME (type) == NULL)
 	    fprintf_filtered (stream, "{...}");
 	}
-      else if (show > 0 || TYPE_TAG_NAME (type) == NULL)
+      else if (show > 0 || TYPE_NAME (type) == NULL)
 	{
 	  LONGEST lastval = 0;
 
@@ -1786,7 +1802,7 @@  c_type_print_base_1 (struct type *type, struct ui_file *stream,
 
     case TYPE_CODE_NAMESPACE:
       fputs_filtered ("namespace ", stream);
-      fputs_filtered (TYPE_TAG_NAME (type), stream);
+      fputs_filtered (TYPE_NAME (type), stream);
       break;
 
     default:
diff --git a/gdb/c-varobj.c b/gdb/c-varobj.c
index 03656c0a81..7f12989979 100644
--- a/gdb/c-varobj.c
+++ b/gdb/c-varobj.c
@@ -145,8 +145,7 @@  c_is_path_expr_parent (const struct varobj *var)
   /* Anonymous unions and structs are also not path_expr parents.  */
   if ((TYPE_CODE (type) == TYPE_CODE_STRUCT
        || TYPE_CODE (type) == TYPE_CODE_UNION)
-      && TYPE_NAME (type) == NULL
-      && TYPE_TAG_NAME (type) == NULL)
+      && TYPE_NAME (type) == NULL)
     {
       const struct varobj *parent = var->parent;
 
diff --git a/gdb/coffread.c b/gdb/coffread.c
index cad3e7e2f1..e07a440443 100644
--- a/gdb/coffread.c
+++ b/gdb/coffread.c
@@ -1787,11 +1787,11 @@  process_coff_symbol (struct coff_symbol *cs,
 	  /* Some compilers try to be helpful by inventing "fake"
 	     names for anonymous enums, structures, and unions, like
 	     "~0fake" or ".0fake".  Thanks, but no thanks...  */
-	  if (TYPE_TAG_NAME (SYMBOL_TYPE (sym)) == 0)
+	  if (TYPE_NAME (SYMBOL_TYPE (sym)) == 0)
 	    if (SYMBOL_LINKAGE_NAME (sym) != NULL
 		&& *SYMBOL_LINKAGE_NAME (sym) != '~'
 		&& *SYMBOL_LINKAGE_NAME (sym) != '.')
-	      TYPE_TAG_NAME (SYMBOL_TYPE (sym)) =
+	      TYPE_NAME (SYMBOL_TYPE (sym)) =
 		concat (SYMBOL_LINKAGE_NAME (sym), (char *)NULL);
 
 	  add_symbol_to_list (sym, &file_symbols);
@@ -1957,10 +1957,6 @@  decode_base_type (struct coff_symbol *cs,
 	  type = coff_alloc_type (cs->c_symnum);
 	  TYPE_CODE (type) = TYPE_CODE_STRUCT;
 	  TYPE_NAME (type) = NULL;
-	  /* This used to set the tag to "<opaque>".  But I think
-	     setting it to NULL is right, and the printing code can
-	     print it as "struct {...}".  */
-	  TYPE_TAG_NAME (type) = NULL;
 	  INIT_CPLUS_SPECIFIC (type);
 	  TYPE_LENGTH (type) = 0;
 	  TYPE_FIELDS (type) = 0;
@@ -1981,10 +1977,6 @@  decode_base_type (struct coff_symbol *cs,
 	  /* Anonymous union type.  */
 	  type = coff_alloc_type (cs->c_symnum);
 	  TYPE_NAME (type) = NULL;
-	  /* This used to set the tag to "<opaque>".  But I think
-	     setting it to NULL is right, and the printing code can
-	     print it as "union {...}".  */
-	  TYPE_TAG_NAME (type) = NULL;
 	  INIT_CPLUS_SPECIFIC (type);
 	  TYPE_LENGTH (type) = 0;
 	  TYPE_FIELDS (type) = 0;
@@ -2007,10 +1999,6 @@  decode_base_type (struct coff_symbol *cs,
 	  type = coff_alloc_type (cs->c_symnum);
 	  TYPE_CODE (type) = TYPE_CODE_ENUM;
 	  TYPE_NAME (type) = NULL;
-	  /* This used to set the tag to "<opaque>".  But I think
-	     setting it to NULL is right, and the printing code can
-	     print it as "enum {...}".  */
-	  TYPE_TAG_NAME (type) = NULL;
 	  TYPE_LENGTH (type) = 0;
 	  TYPE_FIELDS (type) = 0;
 	  TYPE_NFIELDS (type) = 0;
diff --git a/gdb/cp-support.c b/gdb/cp-support.c
index dde417be80..c16b67171f 100644
--- a/gdb/cp-support.c
+++ b/gdb/cp-support.c
@@ -199,7 +199,7 @@  inspect_type (struct demangle_parse_info *info,
 	      && strcmp (TYPE_NAME (type), name) == 0)
 	    return 0;
 
-	  is_anon = (TYPE_TAG_NAME (type) == NULL
+	  is_anon = (TYPE_NAME (type) == NULL
 		     && (TYPE_CODE (type) == TYPE_CODE_ENUM
 			 || TYPE_CODE (type) == TYPE_CODE_STRUCT
 			 || TYPE_CODE (type) == TYPE_CODE_UNION));
diff --git a/gdb/dwarf2read.c b/gdb/dwarf2read.c
index 8dbe12db02..78c36887ee 100644
--- a/gdb/dwarf2read.c
+++ b/gdb/dwarf2read.c
@@ -9778,8 +9778,6 @@  fixup_go_packaging (struct dwarf2_cu *cu)
 				     saved_package_name);
       struct symbol *sym;
 
-      TYPE_TAG_NAME (type) = TYPE_NAME (type);
-
       sym = allocate_symbol (objfile);
       SYMBOL_SET_LANGUAGE (sym, language_go, &objfile->objfile_obstack);
       SYMBOL_SET_NAMES (sym, saved_package_name,
@@ -15621,18 +15619,13 @@  read_structure_type (struct die_info *die, struct dwarf2_cu *cu)
 	  if (get_die_type (die, cu) != NULL)
 	    return get_die_type (die, cu);
 
-	  TYPE_TAG_NAME (type) = full_name;
-	  if (die->tag == DW_TAG_structure_type
-	      || die->tag == DW_TAG_class_type)
-	    TYPE_NAME (type) = TYPE_TAG_NAME (type);
+	  TYPE_NAME (type) = full_name;
 	}
       else
 	{
 	  /* The name is already allocated along with this objfile, so
 	     we don't need to duplicate it for the type.  */
-	  TYPE_TAG_NAME (type) = name;
-	  if (die->tag == DW_TAG_class_type)
-	    TYPE_NAME (type) = TYPE_TAG_NAME (type);
+	  TYPE_NAME (type) = name;
 	}
     }
 
@@ -15913,7 +15906,7 @@  process_structure_scope (struct die_info *die, struct dwarf2_cu *cu)
 		    complaint (&symfile_complaints,
 			       _("virtual function table pointer "
 				 "not found when defining class '%s'"),
-			       TYPE_TAG_NAME (type) ? TYPE_TAG_NAME (type) :
+			       TYPE_NAME (type) ? TYPE_NAME (type) :
 			       "");
 		}
 	      else
@@ -16104,7 +16097,7 @@  read_enumeration_type (struct die_info *die, struct dwarf2_cu *cu)
   TYPE_CODE (type) = TYPE_CODE_ENUM;
   name = dwarf2_full_name (NULL, die, cu);
   if (name != NULL)
-    TYPE_TAG_NAME (type) = name;
+    TYPE_NAME (type) = name;
 
   attr = dwarf2_attr (die, DW_AT_type, cu);
   if (attr != NULL)
@@ -16649,7 +16642,6 @@  read_namespace_type (struct die_info *die, struct dwarf2_cu *cu)
 
   /* Create the type.  */
   type = init_type (objfile, TYPE_CODE_NAMESPACE, 0, name);
-  TYPE_TAG_NAME (type) = TYPE_NAME (type);
 
   return set_die_type (die, type, cu);
 }
@@ -16715,9 +16707,6 @@  read_module_type (struct die_info *die, struct dwarf2_cu *cu)
                sect_offset_str (die->sect_off));
   type = init_type (objfile, TYPE_CODE_MODULE, 0, module_name);
 
-  /* determine_prefix uses TYPE_TAG_NAME.  */
-  TYPE_TAG_NAME (type) = TYPE_NAME (type);
-
   return set_die_type (die, type, cu);
 }
 
@@ -22155,18 +22144,18 @@  determine_prefix (struct die_info *die, struct dwarf2_cu *cu)
 	   DW_TAG_namespace DIEs with a name of "::" for the global namespace.
 	   Work around this problem here.  */
 	if (cu->language == language_cplus
-	    && strcmp (TYPE_TAG_NAME (parent_type), "::") == 0)
+	    && strcmp (TYPE_NAME (parent_type), "::") == 0)
 	  return "";
 	/* We give a name to even anonymous namespaces.  */
-	return TYPE_TAG_NAME (parent_type);
+	return TYPE_NAME (parent_type);
       case DW_TAG_class_type:
       case DW_TAG_interface_type:
       case DW_TAG_structure_type:
       case DW_TAG_union_type:
       case DW_TAG_module:
 	parent_type = read_type_die (parent, cu);
-	if (TYPE_TAG_NAME (parent_type) != NULL)
-	  return TYPE_TAG_NAME (parent_type);
+	if (TYPE_NAME (parent_type) != NULL)
+	  return TYPE_NAME (parent_type);
 	else
 	  /* An anonymous structure is only allowed non-static data
 	     members; no typedefs, no member functions, et cetera.
@@ -22191,8 +22180,8 @@  determine_prefix (struct die_info *die, struct dwarf2_cu *cu)
 	parent_type = read_type_die (parent, cu);
 	if (TYPE_DECLARED_CLASS (parent_type))
 	  {
-	    if (TYPE_TAG_NAME (parent_type) != NULL)
-	      return TYPE_TAG_NAME (parent_type);
+	    if (TYPE_NAME (parent_type) != NULL)
+	      return TYPE_NAME (parent_type);
 	    return "";
 	  }
 	/* Fall through.  */
diff --git a/gdb/eval.c b/gdb/eval.c
index b6fbfcf6c9..0bed4aa3fc 100644
--- a/gdb/eval.c
+++ b/gdb/eval.c
@@ -984,13 +984,13 @@  evaluate_funcall (type *expect_type, expression *exp, int *pos,
       function_name = NULL;
       if (TYPE_CODE (type) == TYPE_CODE_NAMESPACE)
 	{
-	  function = cp_lookup_symbol_namespace (TYPE_TAG_NAME (type),
+	  function = cp_lookup_symbol_namespace (TYPE_NAME (type),
 						 name,
 						 get_selected_block (0),
 						 VAR_DOMAIN).symbol;
 	  if (function == NULL)
 	    error (_("No symbol \"%s\" in namespace \"%s\"."),
-		   name, TYPE_TAG_NAME (type));
+		   name, TYPE_NAME (type));
 
 	  tem = 1;
 	  /* arg2 is left as NULL on purpose.  */
diff --git a/gdb/f-typeprint.c b/gdb/f-typeprint.c
index f14f1811f4..3f13c111ac 100644
--- a/gdb/f-typeprint.c
+++ b/gdb/f-typeprint.c
@@ -290,7 +290,12 @@  f_type_print_base (struct type *type, struct ui_file *stream, int show,
 
   if ((show <= 0) && (TYPE_NAME (type) != NULL))
     {
-      fprintfi_filtered (level, stream, "%s", TYPE_NAME (type));
+      const char *prefix = "";
+      if (TYPE_CODE (type) == TYPE_CODE_UNION)
+	prefix = "Type, C_Union :: ";
+      else if (TYPE_CODE (type) == TYPE_CODE_STRUCT)
+	prefix = "Type ";
+      fprintfi_filtered (level, stream, "%s%s", prefix, TYPE_NAME (type));
       return;
     }
 
@@ -370,7 +375,7 @@  f_type_print_base (struct type *type, struct ui_file *stream, int show,
 	fprintfi_filtered (level, stream, "Type, C_Union :: ");
       else
 	fprintfi_filtered (level, stream, "Type ");
-      fputs_filtered (TYPE_TAG_NAME (type), stream);
+      fputs_filtered (TYPE_NAME (type), stream);
       /* According to the definition,
          we only print structure elements in case show > 0.  */
       if (show > 0)
@@ -387,12 +392,12 @@  f_type_print_base (struct type *type, struct ui_file *stream, int show,
 	      fputs_filtered ("\n", stream);
 	    }
 	  fprintfi_filtered (level, stream, "End Type ");
-	  fputs_filtered (TYPE_TAG_NAME (type), stream);
+	  fputs_filtered (TYPE_NAME (type), stream);
 	}
       break;
 
     case TYPE_CODE_MODULE:
-      fprintfi_filtered (level, stream, "module %s", TYPE_TAG_NAME (type));
+      fprintfi_filtered (level, stream, "module %s", TYPE_NAME (type));
       break;
 
     default_case:
diff --git a/gdb/gdbtypes.c b/gdb/gdbtypes.c
index b3a037971e..b3a388656e 100644
--- a/gdb/gdbtypes.c
+++ b/gdb/gdbtypes.c
@@ -1469,12 +1469,6 @@  smash_to_method_type (struct type *type, struct type *self_type,
 const char *
 type_name_no_tag (const struct type *type)
 {
-  if (TYPE_TAG_NAME (type) != NULL)
-    return TYPE_TAG_NAME (type);
-
-  /* Is there code which expects this to return the name if there is
-     no tag name?  My guess is that this is mainly used for C++ in
-     cases where the two will always be the same.  */
   return TYPE_NAME (type);
 }
 
@@ -2445,10 +2439,8 @@  check_typedef (struct type *type)
 	    return make_qualified_type (type, instance_flags, NULL);
 
 	  name = type_name_no_tag (type);
-	  /* FIXME: shouldn't we separately check the TYPE_NAME and
-	     the TYPE_TAG_NAME, and look in STRUCT_DOMAIN and/or
-	     VAR_DOMAIN as appropriate?  (this code was written before
-	     TYPE_NAME and TYPE_TAG_NAME were separate).  */
+	  /* FIXME: shouldn't we look in STRUCT_DOMAIN and/or
+	     VAR_DOMAIN as appropriate?  */
 	  if (name == NULL)
 	    {
 	      stub_noname_complaint ();
@@ -2534,10 +2526,8 @@  check_typedef (struct type *type)
   else if (TYPE_STUB (type) && !currently_reading_symtab)
     {
       const char *name = type_name_no_tag (type);
-      /* FIXME: shouldn't we separately check the TYPE_NAME and the
-         TYPE_TAG_NAME, and look in STRUCT_DOMAIN and/or VAR_DOMAIN
-         as appropriate?  (this code was written before TYPE_NAME and
-         TYPE_TAG_NAME were separate).  */
+      /* FIXME: shouldn't we look in STRUCT_DOMAIN and/or VAR_DOMAIN
+         as appropriate?  */
       struct symbol *sym;
 
       if (name == NULL)
@@ -3547,9 +3537,6 @@  check_types_equal (struct type *type1, struct type *type2,
       || TYPE_NFIELDS (type1) != TYPE_NFIELDS (type2))
     return 0;
 
-  if (!compare_maybe_null_strings (TYPE_TAG_NAME (type1),
-				   TYPE_TAG_NAME (type2)))
-    return 0;
   if (!compare_maybe_null_strings (TYPE_NAME (type1), TYPE_NAME (type2)))
     return 0;
 
@@ -4419,10 +4406,6 @@  recursive_dump_type (struct type *type, int spaces)
 		    TYPE_NAME (type) ? TYPE_NAME (type) : "<NULL>");
   gdb_print_host_address (TYPE_NAME (type), gdb_stdout);
   printf_filtered (")\n");
-  printfi_filtered (spaces, "tagname '%s' (",
-		    TYPE_TAG_NAME (type) ? TYPE_TAG_NAME (type) : "<NULL>");
-  gdb_print_host_address (TYPE_TAG_NAME (type), gdb_stdout);
-  printf_filtered (")\n");
   printfi_filtered (spaces, "code 0x%x ", TYPE_CODE (type));
   switch (TYPE_CODE (type))
     {
@@ -4806,8 +4789,6 @@  copy_type_recursive (struct objfile *objfile,
 
   if (TYPE_NAME (type))
     TYPE_NAME (new_type) = xstrdup (TYPE_NAME (type));
-  if (TYPE_TAG_NAME (type))
-    TYPE_TAG_NAME (new_type) = xstrdup (TYPE_TAG_NAME (type));
 
   TYPE_INSTANCE_FLAGS (new_type) = TYPE_INSTANCE_FLAGS (type);
   TYPE_LENGTH (new_type) = TYPE_LENGTH (type);
@@ -5145,7 +5126,7 @@  arch_composite_type (struct gdbarch *gdbarch, const char *name,
 
   gdb_assert (code == TYPE_CODE_STRUCT || code == TYPE_CODE_UNION);
   t = arch_type (gdbarch, code, 0, NULL);
-  TYPE_TAG_NAME (t) = name;
+  TYPE_NAME (t) = name;
   INIT_CPLUS_SPECIFIC (t);
   return t;
 }
diff --git a/gdb/gdbtypes.h b/gdb/gdbtypes.h
index 552a2a2a16..70dac75855 100644
--- a/gdb/gdbtypes.h
+++ b/gdb/gdbtypes.h
@@ -721,26 +721,12 @@  struct main_type
 
   /* * Name of this type, or NULL if none.
 
-     This is used for printing only, except by poorly designed C++
-     code.  For looking up a name, look for a symbol in the
-     VAR_DOMAIN.  This is generally allocated in the objfile's
-     obstack.  However coffread.c uses malloc.  */
+     This is used for printing only.  For looking up a name, look for
+     a symbol in the VAR_DOMAIN.  This is generally allocated in the
+     objfile's obstack.  However coffread.c uses malloc.  */
 
   const char *name;
 
-  /* * Tag name for this type, or NULL if none.  This means that the
-     name of the type consists of a keyword followed by the tag name.
-     Which keyword is determined by the type code ("struct" for
-     TYPE_CODE_STRUCT, etc.).  As far as I know C/C++ are the only
-     languages with this feature.
-
-     This is used for printing only, except by poorly designed C++ code.
-     For looking up a name, look for a symbol in the STRUCT_DOMAIN.
-     One more legitimate use is that if TYPE_STUB is set, this is
-     the name to use to look for definitions in other files.  */
-
-  const char *tag_name;
-
   /* * Every type is now associated with a particular objfile, and the
      type is allocated on the objfile_obstack for that objfile.  One
      problem however, is that there are times when gdb allocates new
@@ -1281,7 +1267,6 @@  extern void allocate_gnat_aux_type (struct type *);
 #define TYPE_INSTANCE_FLAGS(thistype) (thistype)->instance_flags
 #define TYPE_MAIN_TYPE(thistype) (thistype)->main_type
 #define TYPE_NAME(thistype) TYPE_MAIN_TYPE(thistype)->name
-#define TYPE_TAG_NAME(type) TYPE_MAIN_TYPE(type)->tag_name
 #define TYPE_TARGET_TYPE(thistype) TYPE_MAIN_TYPE(thistype)->target_type
 #define TYPE_POINTER_TYPE(thistype) (thistype)->pointer_type
 #define TYPE_REFERENCE_TYPE(thistype) (thistype)->reference_type
diff --git a/gdb/gnu-v3-abi.c b/gdb/gnu-v3-abi.c
index 859187f2e9..781acaa409 100644
--- a/gdb/gnu-v3-abi.c
+++ b/gdb/gnu-v3-abi.c
@@ -164,7 +164,7 @@  build_gdb_vtable_type (struct gdbarch *arch)
   t = arch_type (arch, TYPE_CODE_STRUCT, offset * TARGET_CHAR_BIT, NULL);
   TYPE_NFIELDS (t) = field - field_list;
   TYPE_FIELDS (t) = field_list;
-  TYPE_TAG_NAME (t) = "gdb_gnu_v3_abi_vtable";
+  TYPE_NAME (t) = "gdb_gnu_v3_abi_vtable";
   INIT_CPLUS_SPECIFIC (t);
 
   return make_type_with_address_space (t, TYPE_INSTANCE_FLAG_CODE_SPACE);
@@ -1028,7 +1028,7 @@  build_std_type_info_type (struct gdbarch *arch)
   t = arch_type (arch, TYPE_CODE_STRUCT, offset * TARGET_CHAR_BIT, NULL);
   TYPE_NFIELDS (t) = field - field_list;
   TYPE_FIELDS (t) = field_list;
-  TYPE_TAG_NAME (t) = "gdb_gnu_v3_type_info";
+  TYPE_NAME (t) = "gdb_gnu_v3_type_info";
   INIT_CPLUS_SPECIFIC (t);
 
   return t;
diff --git a/gdb/go-lang.c b/gdb/go-lang.c
index e9cba7713c..60ab2f3dc3 100644
--- a/gdb/go-lang.c
+++ b/gdb/go-lang.c
@@ -106,8 +106,8 @@  static int
 sixg_string_p (struct type *type)
 {
   if (TYPE_NFIELDS (type) == 2
-      && TYPE_TAG_NAME (type) != NULL
-      && strcmp (TYPE_TAG_NAME (type), "string") == 0)
+      && TYPE_NAME (type) != NULL
+      && strcmp (TYPE_NAME (type), "string") == 0)
     return 1;
 
   return 0;
diff --git a/gdb/guile/scm-type.c b/gdb/guile/scm-type.c
index ce67695b76..949e0dca82 100644
--- a/gdb/guile/scm-type.c
+++ b/gdb/guile/scm-type.c
@@ -576,10 +576,16 @@  gdbscm_type_tag (SCM self)
   type_smob *t_smob
     = tyscm_get_type_smob_arg_unsafe (self, SCM_ARG1, FUNC_NAME);
   struct type *type = t_smob->type;
+  const char *tagname = nullptr;
 
-  if (!TYPE_TAG_NAME (type))
+  if (TYPE_CODE (type) == TYPE_CODE_STRUCT
+      || TYPE_CODE (type) == TYPE_CODE_UNION
+      || TYPE_CODE (type) == TYPE_CODE_ENUM)
+    tagname = TYPE_NAME (type);
+
+  if (tagname == nullptr)
     return SCM_BOOL_F;
-  return gdbscm_scm_from_c_string (TYPE_TAG_NAME (type));
+  return gdbscm_scm_from_c_string (tagname);
 }
 
 /* (type-name <gdb:type>) -> string
diff --git a/gdb/m2-typeprint.c b/gdb/m2-typeprint.c
index 27e9429211..7279ab5b54 100644
--- a/gdb/m2-typeprint.c
+++ b/gdb/m2-typeprint.c
@@ -439,22 +439,14 @@  m2_long_set (struct type *type, struct ui_file *stream, int show, int level,
 
   if (m2_is_long_set (type))
     {
-      if (TYPE_TAG_NAME (type) != NULL)
-	{
-	  fputs_filtered (TYPE_TAG_NAME (type), stream);
-	  if (show == 0)
-	    return 1;
-	}
-      else if (TYPE_NAME (type) != NULL)
+      if (TYPE_NAME (type) != NULL)
 	{
 	  fputs_filtered (TYPE_NAME (type), stream);
 	  if (show == 0)
 	    return 1;
+	  fputs_filtered (" = ", stream);
 	}
 
-      if (TYPE_TAG_NAME (type) != NULL || TYPE_NAME (type) != NULL)
-	fputs_filtered (" = ", stream);
-
       if (get_long_set_bounds (type, &low, &high))
 	{
 	  fprintf_filtered(stream, "SET OF ");
@@ -537,11 +529,11 @@  m2_record_fields (struct type *type, struct ui_file *stream, int show,
 		  int level, const struct type_print_options *flags)
 {
   /* Print the tag if it exists.  */
-  if (TYPE_TAG_NAME (type) != NULL)
+  if (TYPE_NAME (type) != NULL)
     {
-      if (!startswith (TYPE_TAG_NAME (type), "$$"))
+      if (!startswith (TYPE_NAME (type), "$$"))
 	{
-	  fputs_filtered (TYPE_TAG_NAME (type), stream);
+	  fputs_filtered (TYPE_NAME (type), stream);
 	  if (show > 0)
 	    fprintf_filtered (stream, " = ");
 	}
@@ -601,10 +593,10 @@  m2_enum (struct type *type, struct ui_file *stream, int show, int level)
   if (show < 0)
     {
       /* If we just printed a tag name, no need to print anything else.  */
-      if (TYPE_TAG_NAME (type) == NULL)
+      if (TYPE_NAME (type) == NULL)
 	fprintf_filtered (stream, "(...)");
     }
-  else if (show > 0 || TYPE_TAG_NAME (type) == NULL)
+  else if (show > 0 || TYPE_NAME (type) == NULL)
     {
       fprintf_filtered (stream, "(");
       len = TYPE_NFIELDS (type);
diff --git a/gdb/mdebugread.c b/gdb/mdebugread.c
index c0bce55148..a760fd9692 100644
--- a/gdb/mdebugread.c
+++ b/gdb/mdebugread.c
@@ -1031,10 +1031,10 @@  parse_symbol (SYMR *sh, union aux_ext *ax, char *ext_sh, int bigend,
 	   (.Fxx or .xxfake or empty) for unnamed struct/union/enums.
 	   Alpha cc puts out an sh->iss of zero for those.  */
 	if (sh->iss == 0 || name[0] == '.' || name[0] == '\0')
-	  TYPE_TAG_NAME (t) = NULL;
+	  TYPE_NAME (t) = NULL;
 	else
-	  TYPE_TAG_NAME (t) = obconcat (&mdebugread_objfile->objfile_obstack,
-					name, (char *) NULL);
+	  TYPE_NAME (t) = obconcat (&mdebugread_objfile->objfile_obstack,
+				    name, (char *) NULL);
 
 	TYPE_CODE (t) = type_code;
 	TYPE_LENGTH (t) = sh->value;
@@ -1699,10 +1699,10 @@  parse_type (int fd, union aux_ext *ax, unsigned int aux_index, int *bs,
 	  /* Do not set the tag name if it is a compiler generated tag name
 	     (.Fxx or .xxfake or empty) for unnamed struct/union/enums.  */
 	  if (name[0] == '.' || name[0] == '\0')
-	    TYPE_TAG_NAME (tp) = NULL;
-	  else if (TYPE_TAG_NAME (tp) == NULL
-		   || strcmp (TYPE_TAG_NAME (tp), name) != 0)
-	    TYPE_TAG_NAME (tp)
+	    TYPE_NAME (tp) = NULL;
+	  else if (TYPE_NAME (tp) == NULL
+		   || strcmp (TYPE_NAME (tp), name) != 0)
+	    TYPE_NAME (tp)
 	      = ((const char *)
 		 obstack_copy0 (&mdebugread_objfile->objfile_obstack,
 				name, strlen (name)));
diff --git a/gdb/p-typeprint.c b/gdb/p-typeprint.c
index e35e0fbd80..f82a2bc981 100644
--- a/gdb/p-typeprint.c
+++ b/gdb/p-typeprint.c
@@ -522,9 +522,9 @@  pascal_type_print_base (struct type *type, struct ui_file *stream, int show,
          only after args !!  */
       break;
     case TYPE_CODE_STRUCT:
-      if (TYPE_TAG_NAME (type) != NULL)
+      if (TYPE_NAME (type) != NULL)
 	{
-	  fputs_filtered (TYPE_TAG_NAME (type), stream);
+	  fputs_filtered (TYPE_NAME (type), stream);
 	  fputs_filtered (" = ", stream);
 	}
       if (HAVE_CPLUS_STRUCT (type))
@@ -538,9 +538,9 @@  pascal_type_print_base (struct type *type, struct ui_file *stream, int show,
       goto struct_union;
 
     case TYPE_CODE_UNION:
-      if (TYPE_TAG_NAME (type) != NULL)
+      if (TYPE_NAME (type) != NULL)
 	{
-	  fputs_filtered (TYPE_TAG_NAME (type), stream);
+	  fputs_filtered (TYPE_NAME (type), stream);
 	  fputs_filtered (" = ", stream);
 	}
       fprintf_filtered (stream, "case <?> of ");
@@ -550,10 +550,10 @@  pascal_type_print_base (struct type *type, struct ui_file *stream, int show,
       if (show < 0)
 	{
 	  /* If we just printed a tag name, no need to print anything else.  */
-	  if (TYPE_TAG_NAME (type) == NULL)
+	  if (TYPE_NAME (type) == NULL)
 	    fprintf_filtered (stream, "{...}");
 	}
-      else if (show > 0 || TYPE_TAG_NAME (type) == NULL)
+      else if (show > 0 || TYPE_NAME (type) == NULL)
 	{
 	  pascal_type_print_derivation_info (stream, type);
 
@@ -740,9 +740,9 @@  pascal_type_print_base (struct type *type, struct ui_file *stream, int show,
       break;
 
     case TYPE_CODE_ENUM:
-      if (TYPE_TAG_NAME (type) != NULL)
+      if (TYPE_NAME (type) != NULL)
 	{
-	  fputs_filtered (TYPE_TAG_NAME (type), stream);
+	  fputs_filtered (TYPE_NAME (type), stream);
 	  if (show > 0)
 	    fputs_filtered (" ", stream);
 	}
@@ -753,10 +753,10 @@  pascal_type_print_base (struct type *type, struct ui_file *stream, int show,
       if (show < 0)
 	{
 	  /* If we just printed a tag name, no need to print anything else.  */
-	  if (TYPE_TAG_NAME (type) == NULL)
+	  if (TYPE_NAME (type) == NULL)
 	    fprintf_filtered (stream, "(...)");
 	}
-      else if (show > 0 || TYPE_TAG_NAME (type) == NULL)
+      else if (show > 0 || TYPE_NAME (type) == NULL)
 	{
 	  fprintf_filtered (stream, "(");
 	  len = TYPE_NFIELDS (type);
diff --git a/gdb/python/py-type.c b/gdb/python/py-type.c
index 8a948eeaa6..6cbb1b663d 100644
--- a/gdb/python/py-type.c
+++ b/gdb/python/py-type.c
@@ -416,10 +416,16 @@  static PyObject *
 typy_get_tag (PyObject *self, void *closure)
 {
   struct type *type = ((type_object *) self)->type;
+  const char *tagname = nullptr;
 
-  if (!TYPE_TAG_NAME (type))
+  if (TYPE_CODE (type) == TYPE_CODE_STRUCT
+      || TYPE_CODE (type) == TYPE_CODE_UNION
+      || TYPE_CODE (type) == TYPE_CODE_ENUM)
+    tagname = TYPE_NAME (type);
+
+  if (tagname == nullptr)
     Py_RETURN_NONE;
-  return PyString_FromString (TYPE_TAG_NAME (type));
+  return PyString_FromString (tagname);
 }
 
 /* Return the type, stripped of typedefs. */
diff --git a/gdb/rust-lang.c b/gdb/rust-lang.c
index cf8a15ee43..aa02ede52f 100644
--- a/gdb/rust-lang.c
+++ b/gdb/rust-lang.c
@@ -95,8 +95,8 @@  rust_tuple_type_p (struct type *type)
      nothing else in the debuginfo to distinguish a tuple from a
      struct.  */
   return (TYPE_CODE (type) == TYPE_CODE_STRUCT
-	  && TYPE_TAG_NAME (type) != NULL
-	  && TYPE_TAG_NAME (type)[0] == '(');
+	  && TYPE_NAME (type) != NULL
+	  && TYPE_NAME (type)[0] == '(');
 }
 
 /* Return true if all non-static fields of a structlike type are in a
@@ -143,9 +143,9 @@  static bool
 rust_slice_type_p (struct type *type)
 {
   return (TYPE_CODE (type) == TYPE_CODE_STRUCT
-	  && TYPE_TAG_NAME (type) != NULL
-	  && (strncmp (TYPE_TAG_NAME (type), "&[", 2) == 0
-	      || strcmp (TYPE_TAG_NAME (type), "&str") == 0));
+	  && TYPE_NAME (type) != NULL
+	  && (strncmp (TYPE_NAME (type), "&[", 2) == 0
+	      || strcmp (TYPE_NAME (type), "&str") == 0));
 }
 
 /* Return true if TYPE is a range type, otherwise false.  */
@@ -157,8 +157,8 @@  rust_range_type_p (struct type *type)
 
   if (TYPE_CODE (type) != TYPE_CODE_STRUCT
       || TYPE_NFIELDS (type) > 2
-      || TYPE_TAG_NAME (type) == NULL
-      || strstr (TYPE_TAG_NAME (type), "::Range") == NULL)
+      || TYPE_NAME (type) == NULL
+      || strstr (TYPE_NAME (type), "::Range") == NULL)
     return false;
 
   if (TYPE_NFIELDS (type) == 0)
@@ -342,13 +342,13 @@  val_print_struct (struct type *type, int embedded_offset,
 
   if (!is_tuple)
     {
-      if (TYPE_TAG_NAME (type) != NULL)
-        fprintf_filtered (stream, "%s", TYPE_TAG_NAME (type));
+      if (TYPE_NAME (type) != NULL)
+        fprintf_filtered (stream, "%s", TYPE_NAME (type));
 
       if (TYPE_NFIELDS (type) == 0)
         return;
 
-      if (TYPE_TAG_NAME (type) != NULL)
+      if (TYPE_NAME (type) != NULL)
         fputs_filtered (" ", stream);
     }
 
@@ -617,7 +617,7 @@  rust_print_struct_def (struct type *type, const char *varstring,
   /* Print a tuple type simply.  */
   if (rust_tuple_type_p (type))
     {
-      fputs_filtered (TYPE_TAG_NAME (type), stream);
+      fputs_filtered (TYPE_NAME (type), stream);
       return;
     }
 
@@ -627,7 +627,7 @@  rust_print_struct_def (struct type *type, const char *varstring,
 
   /* Compute properties of TYPE here because, in the enum case, the
      rest of the code ends up looking only at the variant part.  */
-  const char *tagname = TYPE_TAG_NAME (type);
+  const char *tagname = TYPE_NAME (type);
   bool is_tuple_struct = rust_tuple_struct_type_p (type);
   bool is_tuple = rust_tuple_type_p (type);
   bool is_enum = rust_enum_p (type);
@@ -808,11 +808,11 @@  rust_internal_print_type (struct type *type, const char *varstring,
 	int i, len = 0;
 
 	fputs_filtered ("enum ", stream);
-	if (TYPE_TAG_NAME (type) != NULL)
+	if (TYPE_NAME (type) != NULL)
 	  {
-	    fputs_filtered (TYPE_TAG_NAME (type), stream);
+	    fputs_filtered (TYPE_NAME (type), stream);
 	    fputs_filtered (" ", stream);
-	    len = strlen (TYPE_TAG_NAME (type));
+	    len = strlen (TYPE_NAME (type));
 	  }
 	fputs_filtered ("{\n", stream);
 
@@ -823,7 +823,7 @@  rust_internal_print_type (struct type *type, const char *varstring,
 	    QUIT;
 
 	    if (len > 0
-		&& strncmp (name, TYPE_TAG_NAME (type), len) == 0
+		&& strncmp (name, TYPE_NAME (type), len) == 0
 		&& name[len] == ':'
 		&& name[len + 1] == ':')
 	      name += len + 2;
@@ -912,7 +912,6 @@  rust_composite_type (struct type *original,
 
   TYPE_CODE (result) = TYPE_CODE_STRUCT;
   TYPE_NAME (result) = name;
-  TYPE_TAG_NAME (result) = name;
 
   TYPE_NFIELDS (result) = nfields;
   TYPE_FIELDS (result)
@@ -1090,10 +1089,10 @@  rust_evaluate_funcall (struct expression *exp, int *pos, enum noside noside)
        && TYPE_CODE (type) != TYPE_CODE_ENUM)
       || rust_tuple_type_p (type))
     error (_("Method calls only supported on struct or enum types"));
-  if (TYPE_TAG_NAME (type) == NULL)
+  if (TYPE_NAME (type) == NULL)
     error (_("Method call on nameless type"));
 
-  std::string name = std::string (TYPE_TAG_NAME (type)) + "::" + method;
+  std::string name = std::string (TYPE_NAME (type)) + "::" + method;
 
   block = get_selected_block (0);
   sym = lookup_symbol (name.c_str (), block, VAR_DOMAIN, NULL);
@@ -1606,13 +1605,13 @@  rust_evaluate_subexp (struct type *expect_type, struct expression *exp,
 		if (outer_type != NULL)
 		  error(_("Cannot access field %d of variant %s::%s, "
 			  "there are only %d fields"),
-			field_number, TYPE_TAG_NAME (outer_type),
-			rust_last_path_segment (TYPE_TAG_NAME (type)),
+			field_number, TYPE_NAME (outer_type),
+			rust_last_path_segment (TYPE_NAME (type)),
 			nfields);
 		else
 		  error(_("Cannot access field %d of %s, "
 			  "there are only %d fields"),
-			field_number, TYPE_TAG_NAME (type), nfields);
+			field_number, TYPE_NAME (type), nfields);
 	      }
 
 	    /* Tuples are tuple structs too.  */
@@ -1620,13 +1619,13 @@  rust_evaluate_subexp (struct type *expect_type, struct expression *exp,
 	      {
 		if (outer_type != NULL)
 		  error(_("Variant %s::%s is not a tuple variant"),
-			TYPE_TAG_NAME (outer_type),
-			rust_last_path_segment (TYPE_TAG_NAME (type)));
+			TYPE_NAME (outer_type),
+			rust_last_path_segment (TYPE_NAME (type)));
 		else
 		  error(_("Attempting to access anonymous field %d "
 			  "of %s, which is not a tuple, tuple struct, or "
 			  "tuple-like variant"),
-		      field_number, TYPE_TAG_NAME (type));
+		      field_number, TYPE_NAME (type));
 	      }
 
 	    result = value_primitive_field (lhs, 0, field_number, type);
@@ -1668,7 +1667,7 @@  tuple structs, and tuple-like enum variants"));
 	    if (rust_tuple_type_p (type) || rust_tuple_struct_type_p (type))
 		error (_("Attempting to access named field foo of tuple "
 			 "variant %s::%s, which has only anonymous fields"),
-		       TYPE_TAG_NAME (outer_type),
+		       TYPE_NAME (outer_type),
 		       rust_last_path_segment (TYPE_NAME (type)));
 
 	    TRY
@@ -1679,7 +1678,7 @@  tuple structs, and tuple-like enum variants"));
 	    CATCH (except, RETURN_MASK_ERROR)
 	      {
 		error (_("Could not find field %s of struct variant %s::%s"),
-		       field_name, TYPE_TAG_NAME (outer_type),
+		       field_name, TYPE_NAME (outer_type),
 		       rust_last_path_segment (TYPE_NAME (type)));
 	      }
 	    END_CATCH
diff --git a/gdb/stabsread.c b/gdb/stabsread.c
index 0017f18c35..420cdd2742 100644
--- a/gdb/stabsread.c
+++ b/gdb/stabsread.c
@@ -1344,8 +1344,8 @@  define_symbol (CORE_ADDR valu, const char *string, int desc, int type,
       SYMBOL_ACLASS_INDEX (sym) = LOC_TYPEDEF;
       SYMBOL_VALUE (sym) = valu;
       SYMBOL_DOMAIN (sym) = STRUCT_DOMAIN;
-      if (TYPE_TAG_NAME (SYMBOL_TYPE (sym)) == 0)
-	TYPE_TAG_NAME (SYMBOL_TYPE (sym))
+      if (TYPE_NAME (SYMBOL_TYPE (sym)) == 0)
+	TYPE_NAME (SYMBOL_TYPE (sym))
 	  = obconcat (&objfile->objfile_obstack,
 		      SYMBOL_LINKAGE_NAME (sym),
 		      (char *) NULL);
@@ -1702,7 +1702,7 @@  again:
 	   type.  */
 	type = dbx_alloc_type (typenums, objfile);
 	TYPE_CODE (type) = code;
-	TYPE_TAG_NAME (type) = type_name;
+	TYPE_NAME (type) = type_name;
 	INIT_CPLUS_SPECIFIC (type);
 	TYPE_STUB (type) = 1;
 
@@ -1767,7 +1767,6 @@  again:
                copies of a type otherwise.  */
 	    replace_type (type, xtype);
 	    TYPE_NAME (type) = NULL;
-	    TYPE_TAG_NAME (type) = NULL;
 	  }
 	else
 	  {
@@ -3428,9 +3427,9 @@  complain_about_struct_wipeout (struct type *type)
   const char *name = "";
   const char *kind = "";
 
-  if (TYPE_TAG_NAME (type))
+  if (TYPE_NAME (type))
     {
-      name = TYPE_TAG_NAME (type);
+      name = TYPE_NAME (type);
       switch (TYPE_CODE (type))
         {
         case TYPE_CODE_STRUCT: kind = "struct "; break;
@@ -3439,11 +3438,6 @@  complain_about_struct_wipeout (struct type *type)
         default: kind = "";
         }
     }
-  else if (TYPE_NAME (type))
-    {
-      name = TYPE_NAME (type);
-      kind = "";
-    }
   else
     {
       name = "<unknown>";
@@ -4476,7 +4470,7 @@  add_undefined_type_1 (struct type *type)
 static void
 add_undefined_type (struct type *type, int typenums[2])
 {
-  if (TYPE_TAG_NAME (type) == NULL)
+  if (TYPE_NAME (type) == NULL)
     add_undefined_type_noname (type, typenums);
   else
     add_undefined_type_1 (type);
@@ -4561,7 +4555,7 @@  cleanup_undefined_types_1 (void)
 		struct pending *ppt;
 		int i;
 		/* Name of the type, without "struct" or "union".  */
-		const char *type_name = TYPE_TAG_NAME (*type);
+		const char *type_name = TYPE_NAME (*type);
 
 		if (type_name == NULL)
 		  {
diff --git a/gdb/symmisc.c b/gdb/symmisc.c
index 9adde044cd..4d56a3e952 100644
--- a/gdb/symmisc.c
+++ b/gdb/symmisc.c
@@ -540,7 +540,7 @@  print_symbol (struct gdbarch *gdbarch, struct symbol *symbol,
 
   if (SYMBOL_DOMAIN (symbol) == STRUCT_DOMAIN)
     {
-      if (TYPE_TAG_NAME (SYMBOL_TYPE (symbol)))
+      if (TYPE_NAME (SYMBOL_TYPE (symbol)))
 	{
 	  LA_PRINT_TYPE (SYMBOL_TYPE (symbol), "", outfile, 1, depth,
 			 &type_print_raw_options);
diff --git a/gdb/target-descriptions.c b/gdb/target-descriptions.c
index 5d34e29a86..ddec6ee5d6 100644
--- a/gdb/target-descriptions.c
+++ b/gdb/target-descriptions.c
@@ -428,7 +428,6 @@  make_gdb_type (struct gdbarch *gdbarch, struct tdesc_type *ttype)
     {
       m_type = arch_composite_type (m_gdbarch, NULL, TYPE_CODE_STRUCT);
       TYPE_NAME (m_type) = xstrdup (e->name.c_str ());
-      TYPE_TAG_NAME (m_type) = TYPE_NAME (m_type);
 
       for (const tdesc_type_field &f : e->fields)
 	{
diff --git a/gdb/testsuite/gdb.base/maint.exp b/gdb/testsuite/gdb.base/maint.exp
index 9322108565..4ec80812a9 100644
--- a/gdb/testsuite/gdb.base/maint.exp
+++ b/gdb/testsuite/gdb.base/maint.exp
@@ -345,7 +345,7 @@  foreach { test_name command } $test_list {
 
 set msg "maint print type"
 gdb_test_multiple "maint print type argc" $msg {
-    -re "type node $hex\r\nname .int. \\($hex\\)\r\ntagname .<NULL>. \\($hex\\)\r\ncode $hex \\(TYPE_CODE_INT\\)\r\nlength \[24\]\r\nobjfile $hex\r\ntarget_type $hex\r\npointer_type $hex\r\nreference_type $hex\r\ntype_chain $hex\r\ninstance_flags $hex\r\nflags\r\nnfields 0 $hex\r\n$gdb_prompt $" {
+    -re "type node $hex\r\nname .int. \\($hex\\)\r\ncode $hex \\(TYPE_CODE_INT\\)\r\nlength \[24\]\r\nobjfile $hex\r\ntarget_type $hex\r\npointer_type $hex\r\nreference_type $hex\r\ntype_chain $hex\r\ninstance_flags $hex\r\nflags\r\nnfields 0 $hex\r\n$gdb_prompt $" {
         pass $msg
     }
 }
diff --git a/gdb/testsuite/gdb.cp/typeid.exp b/gdb/testsuite/gdb.cp/typeid.exp
index 4378dcbb59..1731b78b9d 100644
--- a/gdb/testsuite/gdb.cp/typeid.exp
+++ b/gdb/testsuite/gdb.cp/typeid.exp
@@ -27,7 +27,7 @@  proc do_typeid_tests {started} {
     global hex
 
     # We might see the standard type or gdb's internal type.
-    set type_re "(std::type_info|struct gdb_gnu_v3_type_info)"
+    set type_re "(std::type_info|gdb_gnu_v3_type_info)"
 
 
     foreach simple_var {i cp ccp ca b} {
diff --git a/gdb/testsuite/gdb.dwarf2/member-ptr-forwardref.exp b/gdb/testsuite/gdb.dwarf2/member-ptr-forwardref.exp
index 41eff9237c..31707bbd0c 100644
--- a/gdb/testsuite/gdb.dwarf2/member-ptr-forwardref.exp
+++ b/gdb/testsuite/gdb.dwarf2/member-ptr-forwardref.exp
@@ -38,4 +38,5 @@  gdb_test "show cp-abi" {The currently selected C\+\+ ABI is "gnu-v3".*}
 
 gdb_load ${binfile}
 
+gdb_test_no_output "set language c++"
 gdb_test "ptype c" "type = struct C {\[\r\n \t\]*private:\[\r\n \t\]*int \\(C::\\*fp\\)\\(C \\*( const)?\\);\[\r\n \t\]*}"
diff --git a/gdb/testsuite/gdb.dwarf2/method-ptr.exp b/gdb/testsuite/gdb.dwarf2/method-ptr.exp
index 5a4f05350e..7fe755cb7d 100644
--- a/gdb/testsuite/gdb.dwarf2/method-ptr.exp
+++ b/gdb/testsuite/gdb.dwarf2/method-ptr.exp
@@ -82,5 +82,6 @@  if { [prepare_for_testing "failed to prepare" ${testfile} \
     return -1
 }
 
+gdb_test_no_output "set language c++"
 gdb_test "ptype the_typedef" \
     "type = int \\(S::\\*\\)\\(S \\* const, float\\)"
diff --git a/gdb/testsuite/gdb.xml/tdesc-regs.exp b/gdb/testsuite/gdb.xml/tdesc-regs.exp
index bb7b9be16c..94d4007a89 100644
--- a/gdb/testsuite/gdb.xml/tdesc-regs.exp
+++ b/gdb/testsuite/gdb.xml/tdesc-regs.exp
@@ -179,7 +179,7 @@  gdb_test "ptype \$extrareg" "type = (int|long|long long)"
 gdb_test "ptype \$uintreg" "type = uint32_t"
 gdb_test "ptype \$vecreg" "type = int8_t __attribute__ \\(\\(vector_size\\(4\\)\\)\\)"
 gdb_test "ptype \$unionreg" \
-    "type = union {\r\n *v4int8 v4;\r\n *v2int16 v2;\r\n}"
+    "type = union vecint {\r\n *v4int8 v4;\r\n *v2int16 v2;\r\n}"
 gdb_test "ptype \$unionreg.v4" "type = int8_t __attribute__ \\(\\(vector_size\\(4\\)\\)\\)"
 gdb_test "ptype \$structreg" \
     "type = struct struct1 {\r\n *v4int8 v4;\r\n *v2int16 v2;\r\n}"
@@ -189,7 +189,7 @@  gdb_test "ptype \$bitfields" \
 gdb_test "ptype \$flags" \
     "type = flag flags {\r\n *bool X @0;\r\n *uint32_t Y @2;\r\n}"
 gdb_test "ptype \$mixed_flags" \
-    "type = flag mixed_flags {\r\n *bool A @0;\r\n *uint32_t B @1-3;\r\n *bool C @4;\r\n *uint32_t D @5;\r\n *uint32_t @6-7;\r\n *enum {yes = 1, no = 0, maybe = 2, so} Z @8-9;\r\n}"
+    "type = flag mixed_flags {\r\n *bool A @0;\r\n *uint32_t B @1-3;\r\n *bool C @4;\r\n *uint32_t D @5;\r\n *uint32_t @6-7;\r\n *enum Z_values {yes = 1, no = 0, maybe = 2, so} Z @8-9;\r\n}"
 # Reggroups should have at least general and the extra foo group
 gdb_test "maintenance print reggroups" \
     " Group\[ \t\]+Type\[ \t\]+\r\n.* general\[ \t\]+user\[ \t\]+\r\n.* foo\[ \t\]+user\[ \t\]+"
diff --git a/gdb/valarith.c b/gdb/valarith.c
index 58e3a09c37..d2dd590064 100644
--- a/gdb/valarith.c
+++ b/gdb/valarith.c
@@ -61,8 +61,6 @@  find_size_for_pointer_math (struct type *ptr_type)
 	  
 	  name = TYPE_NAME (ptr_target);
 	  if (name == NULL)
-	    name = TYPE_TAG_NAME (ptr_target);
-	  if (name == NULL)
 	    error (_("Cannot perform pointer math on incomplete types, "
 		   "try casting to a known type, or void *."));
 	  else
diff --git a/gdb/valops.c b/gdb/valops.c
index 1f655b5187..cb51a2688e 100644
--- a/gdb/valops.c
+++ b/gdb/valops.c
@@ -3243,7 +3243,7 @@  enum_constant_from_type (struct type *type, const char *name)
     }
 
   error (_("no constant named \"%s\" in enum \"%s\""),
-	 name, TYPE_TAG_NAME (type));
+	 name, TYPE_NAME (type));
 }
 
 /* C++: Given an aggregate type CURTYPE, and a member name NAME,
@@ -3581,7 +3581,7 @@  value_namespace_elt (const struct type *curtype,
 
   if (retval == NULL)
     error (_("No symbol \"%s\" in namespace \"%s\"."), 
-	   name, TYPE_TAG_NAME (curtype));
+	   name, TYPE_NAME (curtype));
 
   return retval;
 }
@@ -3597,7 +3597,7 @@  value_maybe_namespace_elt (const struct type *curtype,
 			   const char *name, int want_address,
 			   enum noside noside)
 {
-  const char *namespace_name = TYPE_TAG_NAME (curtype);
+  const char *namespace_name = TYPE_NAME (curtype);
   struct block_symbol sym;
   struct value *result;