Unbreak Perl build on legacy Darwin systems

[sevan]

On modern macOS the compiler defaults to '-Werror,-Wimplicit-function-declaration' This breaks the build on code which calls a function without a prototype (invalid in C99).
Turning the check off with '-Wno-error=implicit-function-declaration' breaks the build of Perl itself on legacy versions of Darwin with compilers which do not recognise the flags.
Removing it allows Perl to once again build on Mac OS X as old as 10.4 with GCC 4.0.1.

Resolves #18902

[Leont]

I see you've already made a PR for the MakeMaker part of change. This PR should be marked as not mergeable until that has been applied, released and synced back into blead.

Otherwise I approve of this change.

[sevan]

Moved PR into draft mode until other PR for MakeMaker is merged.

[haarg]

The changes to EUMM should be removed from this PR, since that will be handled separately.

Making this change, along with the EUMM change, is likely to break a number of XS modules on macOS. I agree that we should change this, but I'm not sure about the timing.

[sevan]

I've removed cpan/ExtUtils-MakeMaker/lib/ExtUtils/MM_Darwin.pm

[sevan]

Smoke test result with unpatched blead

[sevan]

Smoke test result with patched blead

Ignore the failures, I didn't bump the rev so t/porting/cmp_version.t failed, cpan/ExtUtils-MakeMaker/t/04-xs-rpath-darwin.t fails but the test shouldn't be running on Darwin 8.
There is an issue with cpan/Test-Simple/t/Test2/behavior/ipc_wait_timeout.t hanging which is why the smoke test took so long. I had to run 'kill -9' several times over the past few days.

[jkeenan]

Smoke test result with patched blead

Ignore the failures, I didn't bump the rev so t/porting/cmp_version.t failed, cpan/ExtUtils-MakeMaker/t/04-xs-rpath-darwin.t fails but the test shouldn't be running on Darwin 8. There is an issue with cpan/Test-Simple/t/Test2/behavior/ipc_wait_timeout.t hanging which is why the smoke test took so long. I had to run 'kill -9' several times over the past few days.

There has been no discussion in this ticket in more than a year, and in that time the code has acquired several merge conflicts. Those will have to be resolved before we can proceed.

[sevan]

I'll take a look at updating over the weekend.