Mirror of the gdb-patches mailing list
 help / color / mirror / Atom feed
From: Tom Tromey <tromey@redhat.com>
To: Siva Chandra <sivachandra@google.com>
Cc: Doug Evans <dje@google.com>, gdb-patches <gdb-patches@sourceware.org>
Subject: Re: [RFC - Python Scripting] New method gdb.Architecture.disassemble
Date: Tue, 12 Feb 2013 21:18:00 -0000	[thread overview]
Message-ID: <87txphmdt3.fsf@fleche.redhat.com> (raw)
In-Reply-To: <CAGyQ6gweW=9L4mVCkFetd+LvvpjSYzYKM78mYDu5HfG1LpVwiQ@mail.gmail.com>	(Siva Chandra's message of "Tue, 12 Feb 2013 06:56:51 -0800")

>>>>> "Siva" == Siva Chandra <sivachandra@google.com> writes:

Siva> Thanks for your detailed reply. I now have a patch which does not use
Siva> ui_out. Instead of calling gdb_disassembly, it essentially
Siva> re-implements disasm.c:dump_insns. The patch is attached.

I don't have any problem with this approach.
I do have some nits on the patch though.

Siva> +/* Implementation of gdb.Architecture.disassemble (self, low, high) -> List.
Siva> +   Returns a list of instructions in a memory address range.  Each instruction
Siva> +   in the list is a dict object with the following string keys:
Siva> +

You can write a shorter comment here and just put all the informative
stuff into the docs.

Siva> +  static char unknown_str[] = { '<', 'u', 'n', 'k', 'n', 'o', 'w', 'n', '>', '\0' };

This line looks too long.

I think it isn't too hard to reorganize so that this can be an ordinary
const char *.

Siva> +  result_list = PyList_New (0);
Siva> +  if (!result_list)
Siva> +    {
Siva> +      PyErr_SetString (PyExc_MemoryError,
Siva> +                       _("Unable to create a list of disassembled "
Siva> +                         "instructions."));
Siva> +      return NULL;

You don't need PyErr_SetString here.
PyList_New will have done that already.
So just return NULL.

Siva> +      struct ui_file *memfile = mem_fileopen ();
Siva> +      PyObject *insn_dict = PyDict_New ();
Siva> +      volatile struct gdb_exception except;
Siva> +
Siva> +      if (!insn_dict)
Siva> +        {
Siva> +          Py_DECREF (result_list);
Siva> +          PyErr_SetString (PyExc_MemoryError,
Siva> +                           _("Unable to create a dict for instruction."));
Siva> +
Siva> +          return NULL;

Here too.

This leaks memfile on failure.

Siva> +        }
Siva> +      if (PyList_Append (result_list, insn_dict))
Siva> +        {
Siva> +          Py_DECREF (result_list);
Siva> +          Py_DECREF (insn_dict);
Siva> +
Siva> +          return NULL;  /* PyList_Append Sets the exception.  */

This leaks it too.

Siva> +          if (funcname)
Siva> +            xfree (funcname);
Siva> +          if (filename)
Siva> +            xfree (filename);
Siva> +          if (asm_code)
Siva> +            xfree (asm_code);

xfree handles NULL arguments fine, so remove the "if"s.

Siva> +      if (PyDict_SetItemString (insn_dict, "addr",
Siva> +                                gdb_py_long_from_ulongest (pc))
Siva> +          || PyDict_SetItemString (insn_dict, "asm",
Siva> +                                   PyString_FromString (asm_code))

I think PyString_FromString handles const char * arguments fine.
So you can make this:

    asm_code == NULL ? unknown_str : asm_code

and avoid some hair above and below, and fix up unknown_str as well.

Siva> +        {
Siva> +          Py_DECREF (result_list);
Siva> +          PyErr_SetString (PyExc_MemoryError,
Siva> +                           _("Unable to add fields to instruction dict."));
Siva> +
Siva> +          return NULL;

You don't need the PyErr_SetString.
And this leaks some memory.

Siva> +      if (funcname && funcname != unknown_str)
Siva> +        xfree (funcname);
Siva> +      if (filename && filename != unknown_str)
Siva> +        xfree (filename);
Siva> +      if (asm_code && asm_code != unknown_str)
Siva> +        xfree (asm_code);

No need to check nullity; but with the other changes this could all be
unconditional.

Tom


  reply	other threads:[~2013-02-12 21:18 UTC|newest]

Thread overview: 39+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2013-02-04 14:09 Siva Chandra
2013-02-05 23:28 ` Doug Evans
2013-02-06  1:53   ` Siva Chandra
2013-02-06 20:00     ` Tom Tromey
2013-02-08 18:05     ` Doug Evans
2013-02-09 17:55       ` Matt Rice
2013-02-12 14:56       ` Siva Chandra
2013-02-12 21:18         ` Tom Tromey [this message]
2013-02-13 14:37           ` Siva Chandra
2013-02-13 17:52             ` Eli Zaretskii
2013-02-13 18:03             ` Tom Tromey
2013-02-13 19:50               ` Siva Chandra
2013-02-13 20:42                 ` Doug Evans
2013-02-14 22:46                   ` Siva Chandra
2013-02-15  6:43                     ` Doug Evans
2013-02-15 17:32                       ` Doug Evans
2013-02-15 17:40                         ` Siva Chandra
2013-02-15 17:41                           ` Siva Chandra
2013-02-15 18:57                           ` Doug Evans
2013-02-15 20:36                       ` Siva Chandra
2013-02-15 21:01                         ` Siva Chandra
2013-02-16  5:30                           ` Doug Evans
2013-02-16  8:47                           ` Eli Zaretskii
2013-02-19  5:36                             ` Siva Chandra
2013-02-19 15:51                               ` Paul_Koning
2013-02-19 16:35                                 ` Eli Zaretskii
2013-02-19 16:38                               ` Eli Zaretskii
2013-02-20 12:34                                 ` Siva Chandra
2013-02-20 18:44                                   ` Eli Zaretskii
2013-02-21  1:49                                     ` Siva Chandra
2013-02-06 19:58 ` Tom Tromey
2013-02-06 20:31   ` Phil Muldoon
2013-02-06 22:31   ` Matt Rice
2013-02-06 23:19     ` Siva Chandra
2013-02-07  1:11       ` Siva Chandra
2013-02-07 23:03         ` Matt Rice
     [not found]       ` <20130206235707.GA2353@klara.mpi.htwm.de>
2013-02-07  1:18         ` Siva Chandra
2013-02-07 14:14   ` Siva Chandra
2013-02-07 16:42     ` 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=87txphmdt3.fsf@fleche.redhat.com \
    --to=tromey@redhat.com \
    --cc=dje@google.com \
    --cc=gdb-patches@sourceware.org \
    --cc=sivachandra@google.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