Mirror of the gdb-patches mailing list
 help / color / mirror / Atom feed
From: Pedro Alves <palves@redhat.com>
To: Tom Tromey <tom@tromey.com>, gdb-patches@sourceware.org
Subject: Re: [RFA 15/23] Use std::vector to avoid cleanups
Date: Fri, 02 Jun 2017 19:22:00 -0000	[thread overview]
Message-ID: <51da6cfd-42c9-536b-17b0-dbd9998e0ff4@redhat.com> (raw)
In-Reply-To: <20170503224626.2818-16-tom@tromey.com>

On 05/03/2017 11:46 PM, Tom Tromey wrote:
> This patch introduces std::vector in several spots in gdb that were
> using xmalloc and a cleanup.
> 
> The simple_search_memory change warrants a bit of extra attention.  It
> was previously using malloc to try to handle memory allocation
> failures.  I don't know whether this was actually important.

(
It still pains me a little to see std::vector used when we're going to
fill in the data anyway, because std::vector value/zero-initializes,
which is pointless in these cases.  :-)  I think I'll need to accept
that people dislike the idea of unique_ptr<gdb_byte[]>, so I'll stop
mentioning that, though at some point I think I may propose an
allocator that default-initializes, and then I'll nag people
to use _that_, like:

 typedef std::vector<gdb_byte, default_init_allocator> gdb_byte_vec;

 gdb_byte_vec buf (size);
)

> @@ -601,24 +599,18 @@ elf_rel_plt_read (minimal_symbol_reader &reader,
>  	 OBJFILE the symbol is undefined and the objfile having NAME defined
>  	 may not yet have been loaded.  */
>  
> -      if (string_buffer_size < name_len + got_suffix_len + 1)
> -	{
> -	  string_buffer_size = 2 * (name_len + got_suffix_len);
> -	  string_buffer = (char *) xrealloc (string_buffer, string_buffer_size);
> -	}
> -      memcpy (string_buffer, name, name_len);
> -      memcpy (&string_buffer[name_len], SYMBOL_GOT_PLT_SUFFIX,
> -	      got_suffix_len + 1);
> +      string_buffer.clear ();
> +      string_buffer.insert (string_buffer.end (), name, name + name_len);
> +      string_buffer.insert (string_buffer.end (), got_suffix,
> +			    got_suffix + got_suffix_len + 1);

So the patch all looks good to me, but it seems to me that in this
particular case, making this buffer be a std::string would clarify
the code:

      string_buffer.assign (name, name_len);
      string_buffer.append (got_suffix, got_suffix_len);

>  
> -      msym = record_minimal_symbol (reader, string_buffer,
> +      msym = record_minimal_symbol (reader, string_buffer.data (),
>  				    name_len + got_suffix_len,

and here maybe use string_buffer.size ()

(std::string is guaranteed to be null terminated.)

>                                      true, address, mst_slot_got_plt, got_plt,
>  				    objfile);

Thanks,
Pedro Alves


  reply	other threads:[~2017-06-02 19:22 UTC|newest]

Thread overview: 58+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2017-05-03 22:46 [RFA 00/23] More miscellaneous C++-ification Tom Tromey
2017-05-03 22:46 ` [RFA 04/23] Use gdb_file_up in fbsd-nat.c Tom Tromey
2017-05-03 23:52   ` John Baldwin
2017-05-03 22:46 ` [RFA 07/23] Remove make_cleanup_fclose Tom Tromey
2017-06-02 17:32   ` Pedro Alves
2017-05-03 22:46 ` [RFA 14/23] Use unique_xmalloc_ptr in jit.c Tom Tromey
2017-06-02 18:42   ` Pedro Alves
2017-06-05 13:09     ` Tom Tromey
2017-05-03 22:46 ` [RFA 19/23] Replace do_restore_instream_cleanup with scoped_restore Tom Tromey
2017-06-05 13:49   ` Pedro Alves
2017-05-03 22:46 ` [RFA 08/23] Remove an unlink cleanup Tom Tromey
2017-06-02 17:37   ` Pedro Alves
2017-05-03 22:46 ` [RFA 05/23] Use gdb_file_up in source.c Tom Tromey
2017-06-02 17:27   ` Pedro Alves
2017-05-03 22:46 ` [RFA 13/23] Replace tui_restore_gdbout with scoped_restore Tom Tromey
2017-06-02 18:34   ` Pedro Alves
2017-05-03 22:46 ` [RFA 16/23] Remove in_user_command Tom Tromey
2017-06-05 13:27   ` Pedro Alves
2017-05-03 22:46 ` [RFA 20/23] Avoid some manual memory management in Python Tom Tromey
2017-06-05 13:55   ` Pedro Alves
2017-05-03 22:46 ` [RFA 09/23] Remove close cleanup Tom Tromey
2017-06-02 18:08   ` Pedro Alves
2017-07-19 22:52     ` Tom Tromey
2017-07-31 19:08       ` Simon Marchi
2017-08-01 21:52         ` Tom Tromey
2017-05-03 22:46 ` [RFA 23/23] Use gdb_argv_up in Python Tom Tromey
     [not found]   ` <32dc20d7-52ef-1599-d1fa-1ec299fe291c@redhat.com>
2017-07-19 22:52     ` Tom Tromey
2017-07-19 22:51       ` Tom Tromey
2017-05-03 22:46 ` [RFA 06/23] Change open_terminal_stream to return a gdb_file_up Tom Tromey
2017-06-02 17:31   ` Pedro Alves
2017-05-03 22:46 ` [RFA 02/23] Introduce and use gdb_file_up Tom Tromey
2017-06-02 17:08   ` Pedro Alves
2017-05-03 22:46 ` [RFA 11/23] Remove make_cleanup_free_so Tom Tromey
2017-06-02 18:25   ` Pedro Alves
2017-05-03 22:46 ` [RFA 12/23] More uses of scoped_restore Tom Tromey
2017-06-02 18:31   ` Pedro Alves
2017-05-03 22:46 ` [RFA 10/23] Remove make_cleanup_restore_current_language Tom Tromey
2017-06-02 18:18   ` Pedro Alves
2017-06-05 13:09     ` Tom Tromey
2017-05-03 22:46 ` [RFA 17/23] Use a scoped_restore for user_call_depth Tom Tromey
2017-06-05 13:32   ` Pedro Alves
2017-06-05 16:56     ` Tom Tromey
2017-05-03 22:46 ` [RFA 22/23] Make gdb_buildargv return a unique pointer Tom Tromey
2017-06-05 16:21   ` Pedro Alves
2017-07-23 17:26     ` Tom Tromey
2017-05-03 22:46 ` [RFA 03/23] Use gdb_file_up in find_and_open_script Tom Tromey
2017-06-02 17:24   ` Pedro Alves
2017-07-23 16:08     ` Tom Tromey
2017-05-03 22:46 ` [RFA 21/23] Remove a cleanup in Python Tom Tromey
2017-06-05 13:56   ` Pedro Alves
2017-05-03 22:46 ` [RFA 01/23] Introduce and use ui_out_emit_table Tom Tromey
2017-06-02 17:04   ` Pedro Alves
2017-05-03 22:46 ` [RFA 18/23] Use a scoped_restore for command_nest_depth Tom Tromey
2017-06-05 13:38   ` Pedro Alves
2017-05-03 22:46 ` [RFA 15/23] Use std::vector to avoid cleanups Tom Tromey
2017-06-02 19:22   ` Pedro Alves [this message]
2017-07-23 16:06     ` Tom Tromey
2017-05-29 17:31 ` [RFA 00/23] More miscellaneous C++-ification 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=51da6cfd-42c9-536b-17b0-dbd9998e0ff4@redhat.com \
    --to=palves@redhat.com \
    --cc=gdb-patches@sourceware.org \
    --cc=tom@tromey.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