Re: [PATCH v5] attribs: Implement -Wno-attributes=vendor::attr [PR101940]

[Jason Merrill <jason@redhat.com>]

On 11/6/21 14:28, Marek Polacek wrote:
> On Sat, Nov 06, 2021 at 02:32:26AM +0100, Bernhard Reutner-Fischer wrote:
>> On 6 November 2021 01:21:43 CET, Marek Polacek via Gcc-patches <gcc-patches@gcc.gnu.org> wrote:
>>
>>>
>>> Thanks, so like this?  I'm including an incremental diff so that it's
>>> clear what changed:
>>>
>>> diff --git a/gcc/attribs.c b/gcc/attribs.c
>>> index d5fba7f4bbb..addfe6f6c80 100644
>>> --- a/gcc/attribs.c
>>> +++ b/gcc/attribs.c
>>> @@ -237,7 +237,7 @@ check_attribute_tables (void)
>>>     the end of parsing of all TUs. */
>>> static vec<attribute_spec *> ignored_attributes_table;
>>>
>>> -/* Parse arguments ARGS of -Wno-attributes=.
>>> +/* Parse arguments V of -Wno-attributes=.
>>>     Currently we accept:
>>>       vendor::attr
>>>       vendor::
>>> @@ -252,12 +252,15 @@ handle_ignored_attributes_option (vec<char *> *v)
>>>
>>>    for (auto opt : v)
>>>      {
>>> +      /* We're going to be modifying the string.  */
>>> +      opt = xstrdup (opt);
>>>        char *q = strstr (opt, "::");
>>>        /* We don't accept '::attr'.  */
>>>        if (q == nullptr || q == opt)
>>> 	{
>>> 	  error ("wrong argument to ignored attributes");
>>> 	  inform (input_location, "valid format is %<ns::attr%> or %<ns::%>");
>>> +	  free (opt);
>>> 	  continue;
>>> 	}
>>
>> Only xstrdup here, after the strstr check?
>> Should maybe strdup the rest here, not full opt..
> 
> No, I want q to point into the copy of the string, since I'm about
> to modify it.  And I'd prefer a single call to xstrdup rather than
> two.

It occurs to me that instead of calling xstrdup at all, since you're 
already passing the strings to get_identifier you could use 
get_identifier_with_length instead, and then refer to IDENTIFIER_POINTER 
of the result.

Jason