Comments on: Making Strong Types Hashable

By: Daniel Houck

Daniel Houck — Thu, 08 Jun 2017 01:16:00 +0000

In reply to Jonathan Boccara. Ah, that makes sense, and explains why checkIfHashable would work. I wonder if there's a reason beyond just typedefs being easier to compile than functions? Too bad the standard doesn't usually explain the reasoning behind that kind of decision.

By: Jonathan Boccara

Jonathan Boccara — Wed, 07 Jun 2017 19:53:00 +0000

In reply to Daniel Houck. Good question, Daniel. I made the test, and while a function isn't instantiated if unused, the using directive is.

By: Daniel Houck

Daniel Houck — Tue, 30 May 2017 22:04:00 +0000

I’m surprised this works; I would have expected SFINAE would mean that as long as you didn’t try to use checkIfHashable the struct and operator() would be found. Does using the struct as a template parameter automatically attempt to fully instantiate it, or do I need to re-read when SFINAE applies?

By: Jonathan Boccara

Jonathan Boccara — Tue, 30 May 2017 21:31:00 +0000

In reply to Romain Létendart. You did very well to nitpick about this Romain, it was confusing :) It's fixed now. Thanks!

By: Romain Létendart

Romain Létendart — Tue, 30 May 2017 20:27:00 +0000

Always interesting to see how to benefit from the standard library with custom types!
Just to nitpick, the last example should output 20 not 10.