Nathan Howell
Nathan Howell
Commits like f7e310856540241ebb68f98f5c451d1ad070f9f5 churn a huge amount of code due to formatting changes. Could you rebase this pull request down to a smaller number of commits against the HEAD and...
Yes, I believe so. I'm not much of a fan of the split either... any chance you could put together a patch?
I added derived `MonadFix` instances in 0b042455cdab68b92e12db6485a8bb4a436953b8... if you have a chance to try them out that'd be great.
Can you fix llvm-base as well? It doesn't make much sense to remove it from one location but not the other.
I'm starting to think this is a bug in GHC, since regular C/C++ code uses `i1` as well for `_Bool` and `bool`...
Just spoke to Igloo... the `Storable Bool` instance used by ffi doesn't match up with the `i1` `bool` used by C/C++, it should be `CChar` or `Word8` instead... so it's...
If you see any such corruptions can you send a patch? I agree it's the wrong thing and LLVM definitely treats its Bool (i1) type as a single bit.
Patches would be appreciated :thumbsup:
We're aware of the problem. If you don't have the time or desire to roll the patch back I'll try to get it done soon.
`Bool` better captures the semantics of the API but it needs to be done correctly.