public inbox for gdb-patches@sourceware.org
 help / color / mirror / Atom feed
From: Andrew Burgess <aburgess@redhat.com>
To: Tom Tromey <tromey@adacore.com>, gdb-patches@sourceware.org
Subject: Re: [PATCH v2 5/9] Introduce "static constructor" for mi_parse
Date: Fri, 11 Aug 2023 12:02:34 +0100	[thread overview]
Message-ID: <875y5let9x.fsf@redhat.com> (raw)
In-Reply-To: <20230404-dap-loaded-sources-v2-5-93f229095e03@adacore.com>

Tom Tromey via Gdb-patches <gdb-patches@sourceware.org> writes:

> Change the mi_parse function to be a static method of mi_parse.  This
> lets us remove the 'set_args' setter function.

Hi Tom,

Sorry to respond to such an old thread, but I ran into the
mi_parse::make code today.  From your commit message I don't see the
connection for why we need a static mi_parse::make function in order to
get rid of the set_args method.

For reference, I switched the code back to using constructors (but
didn't add back set_args), and I'm not seeing any regressions in
gdb.mi/* and gdb.python/*.

Could you expand on why this change is needed?

Thanks,
Andrew


> ---
>  gdb/mi/mi-main.c  |  2 +-
>  gdb/mi/mi-parse.c |  6 +++---
>  gdb/mi/mi-parse.h | 28 +++++++++++++---------------
>  3 files changed, 17 insertions(+), 19 deletions(-)
>
> diff --git a/gdb/mi/mi-main.c b/gdb/mi/mi-main.c
> index aaebce40fac..b430115daea 100644
> --- a/gdb/mi/mi-main.c
> +++ b/gdb/mi/mi-main.c
> @@ -1919,7 +1919,7 @@ mi_execute_command (const char *cmd, int from_tty)
>  
>    try
>      {
> -      command = mi_parse (cmd, &token);
> +      command = mi_parse::make (cmd, &token);
>      }
>    catch (const gdb_exception &exception)
>      {
> diff --git a/gdb/mi/mi-parse.c b/gdb/mi/mi-parse.c
> index f077eb36a7c..b7c5a8cecdf 100644
> --- a/gdb/mi/mi-parse.c
> +++ b/gdb/mi/mi-parse.c
> @@ -109,7 +109,7 @@ mi_parse_escape (const char **string_ptr)
>  void
>  mi_parse::parse_argv ()
>  {
> -  const char *chp = m_args.get ();
> +  const char *chp = m_args.c_str ();
>    int argc = 0;
>    char **argv = XNEWVEC (char *, argc + 1);
>  
> @@ -216,7 +216,7 @@ mi_parse::~mi_parse ()
>  }
>  
>  std::unique_ptr<struct mi_parse>
> -mi_parse (const char *cmd, char **token)
> +mi_parse::make (const char *cmd, char **token)
>  {
>    const char *chp;
>  
> @@ -345,7 +345,7 @@ mi_parse (const char *cmd, char **token)
>      }
>  
>    /* Save the rest of the arguments for the command.  */
> -  parse->set_args (chp);
> +  parse->m_args = chp;
>  
>    /* Fully parsed, flag as an MI command.  */
>    parse->op = MI_COMMAND;
> diff --git a/gdb/mi/mi-parse.h b/gdb/mi/mi-parse.h
> index edb61547354..6f1da6e6eb5 100644
> --- a/gdb/mi/mi-parse.h
> +++ b/gdb/mi/mi-parse.h
> @@ -41,7 +41,17 @@ enum mi_command_type
>  
>  struct mi_parse
>    {
> -    mi_parse () = default;
> +    /* Attempts to parse CMD returning a ``struct mi_parse''.  If CMD is
> +       invalid, an exception is thrown.  For an MI_COMMAND COMMAND, ARGS
> +       and OP are initialized.  Un-initialized fields are zero.  *TOKEN is
> +       set to the token, even if an exception is thrown.  It is allocated
> +       with xmalloc; it must either be freed with xfree, or assigned to
> +       the TOKEN field of the resultant mi_parse object, to be freed by
> +       mi_parse_free.  */
> +
> +    static std::unique_ptr<struct mi_parse> make (const char *cmd,
> +						  char **token);
> +
>      ~mi_parse ();
>  
>      DISABLE_COPY_AND_ASSIGN (mi_parse);
> @@ -54,9 +64,6 @@ struct mi_parse
>      const char *args () const
>      { return m_args.c_str (); }
>  
> -    void set_args (const char *args)
> -    { m_args = args; }
> -
>      enum mi_command_type op = MI_COMMAND;
>      char *command = nullptr;
>      char *token = nullptr;
> @@ -75,20 +82,11 @@ struct mi_parse
>  
>    private:
>  
> +    mi_parse () = default;
> +
>      std::string m_args;
>    };
>  
> -/* Attempts to parse CMD returning a ``struct mi_parse''.  If CMD is
> -   invalid, an exception is thrown.  For an MI_COMMAND COMMAND, ARGS
> -   and OP are initialized.  Un-initialized fields are zero.  *TOKEN is
> -   set to the token, even if an exception is thrown.  It is allocated
> -   with xmalloc; it must either be freed with xfree, or assigned to
> -   the TOKEN field of the resultant mi_parse object, to be freed by
> -   mi_parse_free.  */
> -
> -extern std::unique_ptr<struct mi_parse> mi_parse (const char *cmd,
> -						  char **token);
> -
>  /* Parse a string argument into a print_values value.  */
>  
>  enum print_values mi_parse_print_values (const char *name);
>
> -- 
> 2.40.0


  reply	other threads:[~2023-08-11 11:02 UTC|newest]

Thread overview: 18+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2023-05-18 20:18 [PATCH v2 0/9] Implement the DAP "loadedSources" request Tom Tromey
2023-05-18 20:18 ` [PATCH v2 1/9] Use field_signed from Python MI commands Tom Tromey
2023-05-18 20:18 ` [PATCH v2 2/9] Use member initializers in mi_parse Tom Tromey
2023-05-18 20:18 ` [PATCH v2 3/9] Use accessor for mi_parse::args Tom Tromey
2023-05-18 20:18 ` [PATCH v2 4/9] Change mi_parse_argv to a method Tom Tromey
2023-05-18 20:18 ` [PATCH v2 5/9] Introduce "static constructor" for mi_parse Tom Tromey
2023-08-11 11:02   ` Andrew Burgess [this message]
2023-08-11 13:10     ` Tom Tromey
2023-08-11 15:06       ` Andrew Burgess
2023-05-18 20:18 ` [PATCH v2 6/9] Introduce mi_parse helper methods Tom Tromey
2023-05-18 20:18 ` [PATCH v2 7/9] Add second mi_parse constructor Tom Tromey
2023-05-18 20:18 ` [PATCH v2 8/9] Implement gdb.execute_mi Tom Tromey
2023-05-19  5:37   ` Eli Zaretskii
2023-05-29 15:54   ` Simon Marchi
2023-06-02 19:19     ` Tom Tromey
2023-06-08 20:16     ` Tom Tromey
2023-05-18 20:18 ` [PATCH v2 9/9] Implement DAP loadedSources request Tom Tromey
2023-05-23 19:46 ` [PATCH v2 0/9] Implement the DAP "loadedSources" request Tom Tromey

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=875y5let9x.fsf@redhat.com \
    --to=aburgess@redhat.com \
    --cc=gdb-patches@sourceware.org \
    --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;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).