Mirror of the gdb-patches mailing list
 help / color / mirror / Atom feed
From: Joel Brobecker <brobecker@adacore.com>
To: Thiago Jung Bauermann <bauerman@br.ibm.com>
Cc: Daniel Jacobowitz <drow@false.org>,
		gdb-patches ml <gdb-patches@sourceware.org>
Subject: Re: [RFA] Add la_getstr member to language_defn
Date: Tue, 03 Feb 2009 00:23:00 -0000	[thread overview]
Message-ID: <20090203002336.GB3964@adacore.com> (raw)
In-Reply-To: <1230949603.8380.143.camel@localhost.localdomain>

> The patch doesn't apply anymore. This is the same patch, refreshed
> against HEAD as of Dec 28th. Ok?

Just a few comments and questions in addition to Tom's comments...

> +      if ((TYPE_NFIELDS (type) == 1)
> +	  && TYPE_CODE (TYPE_FIELD_TYPE (type, 0)) == TYPE_CODE_RANGE)

The extra parens around "TYPE_NFIELDS (type) == 1" shouldn't be
necessary, right? In this case, it's pretty harmless, but a little
bit below, this really starts making it hard to read a condition...

> +  if (((VALUE_LVAL (value) == not_lval)
> +      || (VALUE_LVAL (value) == lval_internalvar)) && (fetchlimit != UINT_MAX))

Can this be formatted as follow:

  if ((VALUE_LVAL (value) == not_lval
       || VALUE_LVAL (value) == lval_internalvar)
      && fetchlimit != UINT_MAX)

? (assuming my reading is correct!) The reason I like my suggestion
is because there are less parentheses, so it's easier to match them
without using my favorite editor; also, the formatting makes it clear
at which level the || and the && operators are.

> +  if ((TYPE_CODE (element_type) != TYPE_CODE_INT)
> +      && (TYPE_CODE (element_type) != TYPE_CODE_CHAR))

Same here.

> @@ -511,6 +626,7 @@ const struct language_defn minimal_language_defn =
>    c_language_arch_info,
>    default_print_array_index,
>    default_pass_by_reference,
> +  default_get_string,
>    LANG_MAGIC
>  };

I was wondering if it wouldn't be more useful to use the c_get_string
function as the default rather than the default_get_string stub.
What do you guys think? Granted, I know that there are cases for Ada
where this isn't going to be enough, but I also know that it's going
to handle some cases fine. I don't know how other languages such as
Fortran or Pascal encode their strings, but chances are that if we
call this function with a value whose type is an array of chars/ints
or a char/int pointer, the c_get_string value would probably work.

As for the Ada implementation, it's probably going to look like this:

  if (value_type is char/int_array or char/int_pointer)
    {
      /* A string a-la-C: Call the C get_string routine...  */
      c_get_string (...);
      return;
    }

  /* Ada-specific strings handled here.  */

So setting the la_get_string method to c_get_string would be a good
starting point for Ada.  The implementation is robust enough that
we'd get an error if the value type is not handled properly, which
is exactly what the default_get_string routine always does.

-- 
Joel


  parent reply	other threads:[~2009-02-03  0:23 UTC|newest]

Thread overview: 33+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2008-11-24 13:24 Thiago Jung Bauermann
2008-11-24 13:32 ` Daniel Jacobowitz
2008-11-24 14:59   ` Thiago Jung Bauermann
2008-11-24 16:19     ` Daniel Jacobowitz
2008-11-24 20:22       ` Thiago Jung Bauermann
2008-11-25  2:16         ` Daniel Jacobowitz
2008-11-25  8:53           ` Thiago Jung Bauermann
2009-01-03  2:27             ` Thiago Jung Bauermann
2009-02-02 18:42               ` Tom Tromey
2009-02-03 12:51                 ` Thiago Jung Bauermann
2009-02-03 17:44                   ` Tom Tromey
2009-02-04 12:37                     ` Thiago Jung Bauermann
2009-02-04 19:19                       ` Tom Tromey
2009-02-04 22:26                         ` Thiago Jung Bauermann
2009-02-05  0:55                           ` Tom Tromey
2009-02-05 12:21                             ` Thiago Jung Bauermann
2009-02-05 16:01                               ` Pierre Muller
2009-02-05 16:30                                 ` Thiago Jung Bauermann
2009-02-05 16:46                                 ` Tom Tromey
2009-02-05 15:55                           ` Tom Tromey
2009-02-05 16:07                             ` Pierre Muller
2009-02-05 16:33                               ` Thiago Jung Bauermann
2009-02-05 16:46                                 ` Tom Tromey
2009-02-03  0:23               ` Joel Brobecker [this message]
2009-02-03 13:02                 ` Thiago Jung Bauermann
2009-02-03 17:01                   ` Joel Brobecker
2009-02-03 17:40                     ` Tom Tromey
2009-02-05 17:01                     ` Thiago Jung Bauermann
2009-02-05 23:51                       ` Joel Brobecker
2008-11-24 20:03 ` Tom Tromey
2008-11-24 21:19   ` Thiago Jung Bauermann
2008-11-25  0:55     ` Tom Tromey
2008-11-25 11:27       ` Thiago Jung Bauermann

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=20090203002336.GB3964@adacore.com \
    --to=brobecker@adacore.com \
    --cc=bauerman@br.ibm.com \
    --cc=drow@false.org \
    --cc=gdb-patches@sourceware.org \
    /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