Change: avr-gcc@14 formula to use avr-libc 2.2.1

[retsplines]

avr-libc 2.2.1 adds support for many new Microchip devices.

The changes to io.h give an idea of the number of new chips supported.

Another user opened a PR for this which ended up becoming stale due to some outstanding style changes and the desire to not change multiple formulae per PR, so this is re-raised as a separate PR per GCC version.

[retsplines]

If this is all good, I'll open PRs for the other version-specific avr-gcc formulae.

[ladislas]

Thanks @retsplines! I've started the CI

[ladislas]

@retsplines ci failed. If you look at the logs it should be some easy fixes. You can run the commands locally to check for errors

[retsplines]

Fixed the style/lint error. There was one other which relates to a missing dependency on isl@0.18 which is pinned in avr-gcc@5 but that doesn't relate to this change and should probably be fixed in a different PR I guess.

[ladislas]

I'm going to deprecate gcc@5 and gcc@6 as homebrew did Homebrew/homebrew-core#106902

I'll let you know when I'm done so you can rebase on main

[ladislas]

@retsplines done, you can rebase on main and the issue should be fixed

I've updated it myself

[ladislas]

@fauxpark on macos-15 it fails because

==> Running Formulae#formula!(osx-cross/avr/avr-gcc@14)
==> brew deps --tree --annotate --include-build --include-test osx-cross/avr/avr-gcc@14
Warning: avr-gcc@14 has unbottled dependencies, so a bottle will not be built.
Warning: No bottle built.
==> SKIPPED osx-cross/avr/avr-gcc@14

would you know why it says so? I have the same error if I run the process on my machine

[ladislas]

ah I think it's because of binutils, i'll bump the formula

[fauxpark]

Yeah, dependencies always have to be bottled first. Was gonna get that PRed in a bit.

[ladislas]

I hope #347 does the trick

I'll be out for the rest of the day, if you have time feel free to edit my PR to get it working if needed :)

[ladislas]

@retsplines @fauxpark I've done the bottle upload manually as CI was failing.
Not sure why...

[fauxpark]

The pr-pull failed because the release already exists - this PR needed to also add revision 1. Now the existing bottles have been overwritten...