Mirror of the gdb-patches mailing list
 help / color / mirror / Atom feed
From: Lancelot SIX via Gdb-patches <gdb-patches@sourceware.org>
To: Tom Tromey <tromey@adacore.com>
Cc: gdb-patches@sourceware.org
Subject: Re: [PATCH 1/7] Use .def file to stringify type codes
Date: Mon, 25 Sep 2023 22:58:10 +0100	[thread overview]
Message-ID: <20230925215810.jhxq35iznktykrzl@octopus> (raw)
In-Reply-To: <20230921-field-bits-v1-1-201285360900@adacore.com>

Hi Tom,

On Thu, Sep 21, 2023 at 12:01:28PM -0600, Tom Tromey via Gdb-patches wrote:
> This changes recursive_dump_type to reuse the type-codes.def file when
> stringifying type codes.
> ---
>  gdb/gdbtypes.c | 99 +++++++++++-----------------------------------------------
>  1 file changed, 18 insertions(+), 81 deletions(-)
> 
> diff --git a/gdb/gdbtypes.c b/gdb/gdbtypes.c
> index 7011fddd695..18aa8e5c29e 100644
> --- a/gdb/gdbtypes.c
> +++ b/gdb/gdbtypes.c
> @@ -5099,6 +5099,23 @@ dump_dynamic_prop (dynamic_prop const& prop)
>      }
>  }
>  
> +/* Return a string that represents a type code.  */
> +static const char *
> +type_code_name (type_code code)
> +{
> +  switch (code)
> +    {
> +#define OP(X) case X: return # X;
> +#include "type-codes.def"
> +#undef OP
> +
> +    case TYPE_CODE_UNDEF:
> +      return "TYPE_CODE_UNDEF";
> +    default:
> +      return "UNKNOWN TYPE CODE";

Is this default "just" to make the compiler happy?  I don't expect it
should ever executed.  If so, I usually prefer the following construct:

    static const char *
    foo (some_enume e)
    {
      switch (e)
        {
        case A:
          return "A";
        case B:
          return "B";
        [...]
        /* No "default:".  */
        }
      return "unknown";
    }

The return after the switch block avoids the "control reaches end of
non-void function" error, but since there is no "default:" the compiler
can warn if some enumerator is not covered by the switch.

That being said, given how type_code is constructed, the
"#include type-codes.def" make it really unlikely  that type_code_name
can evolve without having this function evolve as well.

Best,
Lancelot.

> +    }
> +}
> +
>  void
>  recursive_dump_type (struct type *type, int spaces)
>  {
> @@ -5136,87 +5153,7 @@ recursive_dump_type (struct type *type, int spaces)
>  	      type->name () ? type->name () : "<NULL>",
>  	      host_address_to_string (type->name ()));
>    gdb_printf ("%*scode 0x%x ", spaces, "", type->code ());
> -  switch (type->code ())
> -    {
> -    case TYPE_CODE_UNDEF:
> -      gdb_printf ("(TYPE_CODE_UNDEF)");
> -      break;
> -    case TYPE_CODE_PTR:
> -      gdb_printf ("(TYPE_CODE_PTR)");
> -      break;
> -    case TYPE_CODE_ARRAY:
> -      gdb_printf ("(TYPE_CODE_ARRAY)");
> -      break;
> -    case TYPE_CODE_STRUCT:
> -      gdb_printf ("(TYPE_CODE_STRUCT)");
> -      break;
> -    case TYPE_CODE_UNION:
> -      gdb_printf ("(TYPE_CODE_UNION)");
> -      break;
> -    case TYPE_CODE_ENUM:
> -      gdb_printf ("(TYPE_CODE_ENUM)");
> -      break;
> -    case TYPE_CODE_FLAGS:
> -      gdb_printf ("(TYPE_CODE_FLAGS)");
> -      break;
> -    case TYPE_CODE_FUNC:
> -      gdb_printf ("(TYPE_CODE_FUNC)");
> -      break;
> -    case TYPE_CODE_INT:
> -      gdb_printf ("(TYPE_CODE_INT)");
> -      break;
> -    case TYPE_CODE_FLT:
> -      gdb_printf ("(TYPE_CODE_FLT)");
> -      break;
> -    case TYPE_CODE_VOID:
> -      gdb_printf ("(TYPE_CODE_VOID)");
> -      break;
> -    case TYPE_CODE_SET:
> -      gdb_printf ("(TYPE_CODE_SET)");
> -      break;
> -    case TYPE_CODE_RANGE:
> -      gdb_printf ("(TYPE_CODE_RANGE)");
> -      break;
> -    case TYPE_CODE_STRING:
> -      gdb_printf ("(TYPE_CODE_STRING)");
> -      break;
> -    case TYPE_CODE_ERROR:
> -      gdb_printf ("(TYPE_CODE_ERROR)");
> -      break;
> -    case TYPE_CODE_MEMBERPTR:
> -      gdb_printf ("(TYPE_CODE_MEMBERPTR)");
> -      break;
> -    case TYPE_CODE_METHODPTR:
> -      gdb_printf ("(TYPE_CODE_METHODPTR)");
> -      break;
> -    case TYPE_CODE_METHOD:
> -      gdb_printf ("(TYPE_CODE_METHOD)");
> -      break;
> -    case TYPE_CODE_REF:
> -      gdb_printf ("(TYPE_CODE_REF)");
> -      break;
> -    case TYPE_CODE_CHAR:
> -      gdb_printf ("(TYPE_CODE_CHAR)");
> -      break;
> -    case TYPE_CODE_BOOL:
> -      gdb_printf ("(TYPE_CODE_BOOL)");
> -      break;
> -    case TYPE_CODE_COMPLEX:
> -      gdb_printf ("(TYPE_CODE_COMPLEX)");
> -      break;
> -    case TYPE_CODE_TYPEDEF:
> -      gdb_printf ("(TYPE_CODE_TYPEDEF)");
> -      break;
> -    case TYPE_CODE_NAMESPACE:
> -      gdb_printf ("(TYPE_CODE_NAMESPACE)");
> -      break;
> -    case TYPE_CODE_FIXED_POINT:
> -      gdb_printf ("(TYPE_CODE_FIXED_POINT)");
> -      break;
> -    default:
> -      gdb_printf ("(UNKNOWN TYPE CODE)");
> -      break;
> -    }
> +  gdb_printf ("(%s)", type_code_name (type->code ()));
>    gdb_puts ("\n");
>    gdb_printf ("%*slength %s\n", spaces, "",
>  	      pulongest (type->length ()));
> 
> -- 
> 2.40.1
> 

  reply	other threads:[~2023-09-25 21:58 UTC|newest]

Thread overview: 12+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2023-09-21 18:01 [PATCH 0/7] Remove char-based bitfield macros Tom Tromey via Gdb-patches
2023-09-21 18:01 ` [PATCH 1/7] Use .def file to stringify type codes Tom Tromey via Gdb-patches
2023-09-25 21:58   ` Lancelot SIX via Gdb-patches [this message]
2023-09-26 13:10     ` Tom Tromey via Gdb-patches
2023-09-21 18:01 ` [PATCH 2/7] Print field accessibility inline Tom Tromey via Gdb-patches
2023-09-21 18:01 ` [PATCH 3/7] Remove byte vectors from cplus_struct_type Tom Tromey via Gdb-patches
2023-09-25 22:32   ` Lancelot SIX via Gdb-patches
2023-10-27 14:20     ` Tom Tromey
2023-09-21 18:01 ` [PATCH 4/7] Add field::is_public Tom Tromey via Gdb-patches
2023-09-21 18:01 ` [PATCH 5/7] Remove some QUIT calls from need_access_label_p Tom Tromey via Gdb-patches
2023-09-21 18:01 ` [PATCH 6/7] Remove some type field accessor macros Tom Tromey via Gdb-patches
2023-09-21 18:01 ` [PATCH 7/7] Remove char-based bitfield macros Tom Tromey via Gdb-patches

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20230925215810.jhxq35iznktykrzl@octopus \
    --to=gdb-patches@sourceware.org \
    --cc=lsix@lancelotsix.com \
    --cc=tromey@adacore.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox