Re: [PATCH] Fix potential NULL pointer dereference

[Luis Machado <lgustavo@codesourcery.com>]

On 10/24/2016 05:26 PM, Pedro Alves wrote:
> On 10/24/2016 10:52 PM, Luis Machado wrote:
>
>> diff --git a/gdb/exec.c b/gdb/exec.c
>> index 67ecc63..5eeac44 100644
>> --- a/gdb/exec.c
>> +++ b/gdb/exec.c
>> @@ -221,13 +221,20 @@ exec_file_locate_attach (int pid, int defer_bp_reset, int from_tty)
>>      }
>>    CATCH (err, RETURN_MASK_ERROR)
>>      {
>> +      const char *msg;
>> +
>>        if (err.message != NULL)
>> -	warning ("%s", err.message);
>> +	{
>> +	  warning ("%s", err.message);
>> +	  msg = err.message;
>> +	}
>> +      else
>> +	msg = "";
>>
>
> This is overcomplicating a bit, IMO.
>
>>        prev_err = err;
>>
>>        /* Save message so it doesn't get trashed by the catch below.  */
>> -      prev_err.message = xstrdup (err.message);
>> +      prev_err.message = xstrdup (msg);
>
>
> I think this one liner:
>
>  +      if (err.message != NULL)
>           prev_err.message = xstrdup (err.message);
>
> would have the same effect, since exception_print_same
> handles NULL ex.message?
>
> Thanks,
> Pedro Alves
>

Fine by me. I'll push that change then.