version bumps

[shayne-fletcher]

per the title

[shayne-fletcher]

@dylant-da this looks stuck

[dylant-da]

Yes, sorry, I've been meaning to take a proper look at this. Let's start with a simple restart, but I'm guessing

D:\a\ghc-lib\ghc-lib\dist-newstyle\build\x86_64-windows\ghc-8.8.1\ghc-lib-parser-8.10.1.20250825\build\GHCi\FFI.hs:1:1: error:
Error: File name does not match module name:
Saw: ‘Main’
Expected: ‘GHCi.FFI’
Error: [Cabal-7125]
Failed to build ghc-lib-parser-8.10.1.20250825 (which is required by exe:ghc-lib-test-mini-hlint from ghc-lib-test-mini-hlint-8.10.1.20250825 and exe:ghc-lib-test-mini-compile from ghc-lib-test-mini-compile-8.10.1.20250825).

Is the real culprit

[shayne-fletcher]

Yes, sorry, I've been meaning to take a proper look at this. Let's start with a simple restart, but I'm guessing

D:\a\ghc-lib\ghc-lib\dist-newstyle\build\x86_64-windows\ghc-8.8.1\ghc-lib-parser-8.10.1.20250825\build\GHCi\FFI.hs:1:1: error:
Error: File name does not match module name:
Saw: ‘Main’
Expected: ‘GHCi.FFI’
Error: [Cabal-7125]
Failed to build ghc-lib-parser-8.10.1.20250825 (which is required by exe:ghc-lib-test-mini-hlint from ghc-lib-test-mini-hlint-8.10.1.20250825 and exe:ghc-lib-test-mini-compile from ghc-lib-test-mini-compile-8.10.1.20250825).

Is the real culprit

@dylant-da switched to windows-2025 in the relevant workflow. seems to pass with that (https://github.com/shayne-fletcher/ghc-lib/actions/runs/17347071208/job/49248444742)

[dylant-da]

Ah, I didn't realize you were saying that you'd switched it on this PR, so I was getting very confused when trying to apply the changes to this branch myself. I've rerun workflows 🤞

[dylant-da]

Success!

[dylant-da]

Do you have any theories as to why windows-2025 would succeed where windows-latest doesn't?

[shayne-fletcher]

Do you have any theories as to why windows-2025 would succeed where windows-latest doesn't?

not really 😔

[shayne-fletcher]

@dylant-da the next diff is more significant #607. hopefully will fixes the current breakage

[shayne-fletcher]

@dylant-da it just means something changed on the runner in a way that's bad for this one failing config. it happens, not a big deal. another approach could be just don't run that workflow for windows but as it succeeds with this change that seems inferior. i wouldn't be surprised if circled back in a month, reenabled windows-latest and it worked. anyway, can this land please so that we can get on with the next one?