On 09/08/2019 12:34, Jan Beulich wrote:
> On 09.08.2019 13:05, Ian Jackson wrote:
>> Jan Beulich writes ("Re: [Xen-devel] [PATCH 2/2]
>> tools/tests/cpu-policy: disable -Wformat-overflow"):
>>> Would you mind clarifying which 12 you mean to change to 13?
>>> The one in "%.12s" would, if changed and afaict, then
>>> legitimately trigger the warning. And we've already objected
>>> to the array to get grown.
>>
>> I meant the array. I missed that objection. I just went and read the
>> thread
>> tests/cpu-policy: fix format-overflow warning by null terminating
>> strings
>> and it did conclude that the compiler was wrong to complain.
>>
>> But for me it doesn't follow that the original code is necessarily the
>> best way of doing things, and I didn't see anyone giving an argument
>> why simply increasing the array was a bad idea.
>>
>> C "prefers" null-terminated strings in that they work somewhat better
>> with a variety of primitives.
>
> Right, but the %.<num>s specification exits precisely to allow
> to deal with potentially not nul-terminated strings. ACPI code
> in the hypervisor makes quite a bit of use of this, for example,
> without triggering any compiler warnings with 9.1.0.
I also wonder whether
https://gcc.gnu.org/onlinedocs/gcc/Common-Variable-Attributes.html#index-nonstring-variable-attribute
might be a way of fixing this, seeing as it exists specifically for the
purpose.
~Andrew
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel
