[PATCH] debug/tst-longjmp_chk2: Make signal handler more conservative [BZ #20248]

[Florian Weimer]

On 06/13/2016 08:42 PM, Mike Frysinger wrote:
> On 13 Jun 2016 16:23, Florian Weimer wrote:
>> On 06/13/2016 04:03 PM, Mike Frysinger wrote:
>>> On 13 Jun 2016 13:11, Florian Weimer wrote:
>>>> +static void
>>>> +write_message (const char *message)
>>>> +{
>>>> +  ssize_t unused __attribute__ ((unused));
>>>> +  for (int i = 0; i < pass; ++i)
>>>> +    unused = write (STDOUT_FILENO, " ", 1);
>>>> +  unused = write (STDOUT_FILENO, message, strlen (message));
>>>
>>> are you just trying to ignore the wur markings on write ?  you can use
>>> this style instead:
>>> 	if (write (...))
>>> 	  {
>>> 	    /* ignored */
>>> 	  }
>>>
>>> which i'm sure can be macroed.
>>
>> I believe the __attribute_ ((unused)) approach is the recommended way to
>> suppress this warning.  Everything else is just gaming the GCC optimizers.
>
> it's the game that a number of GNU project rely upon quite a bit last
> i looked

Thanks.  I brought this up on the GCC list.  But I don't think there 
will be any movement unless I submit a patch …

Florian