examples: Fix incomplete field selector warnings in AOC_2021_24

[bgamari]

This warning was turned up by GHC's head.hackage testing.

[LeventErkok]

Hi Ben:

Which GHC generates these warnings? (I don't get any with 9.10.1

This particular example uses the exact format of the program as generated by the AOC folks, so I'm hesitant to add the new type. But I'd be OK with adding a pragma at the top to tell GHC "Don't worry about it."

Also, I'd like to understand why this is causing warnings in the first place.

Cheers..

[bgamari]

@LeventErkok GHC 9.12 and later will generate this warning as it has been added to -Wall as described in GHC Proposal #516. If the newtype is problematic we can instead add a {-# OPTIONS_GHC -Wno-incomplete-record-selectors #-} pragma.

[LeventErkok]

Thanks! I suppose I’d have to do a CPP trick to allow compiling against older GHC? What would be the magic incapacitation?

[LeventErkok]

@bgamari This should fix it I suppose? e7daf24

[bgamari]

For the record, it turns out that I had the version number slightly wrong here; it looks like the -Wall change will only happen in 9.14; nevertheless the flag exists in 9.12 so there is no harm done by the merged patch.

Also, I would probably use __GLASGOW_HASKELL__ >= 913 instead of MIN_VERSION_base since we care about the compiler's features, not those of the base library (although I acknowledge that they are in correspondence).

[LeventErkok]

Should that be __GLASGOW_HASKELL__ >= 914? (you said 9.14 will get the change)

[bgamari]

Should that be __GLASGOW_HASKELL__ >= 914? (you said 9.14 will get the change)

The change is already present in GHC 9.13 (the current development version).

[LeventErkok]

Fixed! Thanks..