NathanHowell (Nathan Howell)
User

Projects

User does not belong to any projects.

User Details

User Since
Aug 20 2014, 4:19 PM (260 w, 3 d)

Recent Activity

May 2 2016

NathanHowell abandoned D2157: Attempt to fix the __builtin_clz version of log_2.
May 2 2016, 2:06 PM

Apr 29 2016

NathanHowell added a comment to D2157: Attempt to fix the __builtin_clz version of log_2.

What about comments like this? Are these relevant?

Apr 29 2016, 12:53 PM
NathanHowell added a comment to D2157: Attempt to fix the __builtin_clz version of log_2.

Would you rather just have asserts for both invariants?

Apr 29 2016, 12:34 PM

Apr 28 2016

NathanHowell added a reviewer for D2157: Attempt to fix the __builtin_clz version of log_2: simonmar.
Apr 28 2016, 4:19 PM
NathanHowell retitled D2157: Attempt to fix the __builtin_clz version of log_2 from Revert "Revert "Use __builtin_clz() to implement log_2()"" to Attempt to fix the __builtin_clz version of log_2.
Apr 28 2016, 4:17 PM
NathanHowell retitled D2157: Attempt to fix the __builtin_clz version of log_2 from to Revert "Revert "Use __builtin_clz() to implement log_2()"".
Apr 28 2016, 4:06 PM

Oct 22 2014

NathanHowell added a reviewer for D155: llvmGen: Compatibility with LLVM 3.5: NathanHowell.
Oct 22 2014, 3:59 PM
NathanHowell updated subscribers of D155: llvmGen: Compatibility with LLVM 3.5.
Oct 22 2014, 3:56 PM

Oct 6 2014

NathanHowell added a comment to D252: Support C structure as return value of ccall in cmm.

Does this support structures with non-standard packing? This is somewhat common in practice... at least the IPv6 imports on OSX use them, as do many Win32 headers.

Oct 6 2014, 7:20 PM · GHC

Sep 5 2014

NathanHowell added a comment to D13: Make Applicative a superclass of Monad.
In D13#38, @austin wrote:

Yes, the macro was more of a joke on the diff itself (and the comment about not making the optimizer 'work so hard', hence 'it is a mystery'), not a response to the find itself.

Sep 5 2014, 6:04 PM
NathanHowell added a comment to D13: Make Applicative a superclass of Monad.

This Applicative instance defines (*>) as (>>), and the default Monad instance in the AMP patch points (>>) at (*>), hence the loop. I'd assume bos added this for some reason though, and it's probably worth digging through Hackage to see how common 'override default to subclass' implementations are.

Sep 5 2014, 5:52 PM
NathanHowell added a comment to D13: Make Applicative a superclass of Monad.
In D13#33, @austin wrote:

No, doesn't look like it. There's no real error; it just never finishes.

Sep 5 2014, 3:48 AM