> On 6 Jun 2024, at 12:41, Sam James via Gcc <gcc@gcc.gnu.org> wrote:
>
> Andi Kleen via Gcc <gcc@gcc.gnu.org> writes:
>
>> FX Coudert via Gcc <gcc@gcc.gnu.org> writes:
>>
>>> I am trying to reduce the number of unneeded fixincludes that are used
>>> on darwin (because fixincluded headers make it impossible to change
>>> SDK once the compiler is built, which is common practice in the macOS
>>> world, and quite useful).
>>
>> It's the same problem on Linux. You get all kinds of strange problems
>> with previously installed gccs once you upgrade the system. I'm not sure
>> in fact any modern target gcc or clang as the standard compiler needs
>> fixincludes at all.
>>
>> I usually just install with install-no-fixedincludes, but really this
>> should probably be a configure option and default to on.
>
> See --disable-fixincludes from r13-2319-gbe9dd80f933480. Long overdue
> but we got it :)
the problem here (at least for _some_ targets; Darwin for one) is that we
cannot switch off *all* fixincludes (for some platform SDKs there are real
bugs, and those SDKs are immutable and will not be fixed).
The issue here is to find a way to figure out what the minimum set is, and
in the case that additions are target-independent whether they are still
relevant.
finding a mecahnism to ensure that the set is still current would be neat.
Summarising in test summary output seems like a step to allow inspection
from posted results without asking multiple taget maintainers for specific
action.
Iain
>
>>
>> -Andi
>
> thanks,
> sam
