On 4/10/23 19:35, Thomas Huth wrote:
> On 04/10/2023 19.23, Richard Henderson wrote:
>> On 10/4/23 03:05, Philippe Mathieu-Daudé wrote:
>>> Hi,
>>>
>>> I'm getting a bunch of errors for 'optarg' declared in <unistd.h>:
>>
>> I thought things like this is why we were trying -Wshadow=local.
>>
>> I think it's unlikely that we'll be able to prevent all such cases.
>
> Given the broad range of operating systems and libraries that we support
> in QEMU, I agree with Richard - it will likely be impossible to enable
> that option without =local by default without risking that compilation
> breaks on some exotic systems or new versions of various libraries.
-Wshadow=local doesn't seem to work here which is why I switched
to -Wshadow. I probably misunderstood something from Markus cover
letter. My setup is:
C compiler for the host machine: clang (clang 14.0.3 "Apple clang
version 14.0.3 (clang-1403.0.22.14.1)")
I had trouble with -Wshadow=local with clang too.
In general I agree not wanting it by default... but for globals defined by the standard, we'd definitely want to fix.
Warner
I suppose we'll figure that out when eventually enabling -Wshadow=local
on CI. Meanwhile I already cleaned the 'optarg' warnings that were
bugging me, see:
20231004120019.93101-1-philmd@linaro.org/" rel="noreferrer noreferrer" target="_blank">https://lore.kernel.org/qemu-devel/20231004120019.93101-1-philmd@linaro.org/
I'll try to get -Wshadow=local, but the other series still seems a
good cleanup, as I used more meaningful variable names.
Regards,
Phil.