Gaël Deest

@tchoutri Modulo the small comments for myself, yes, it is !

The `HasServer` instance for `Auth` does depend on cookie settings being present in the context anyhow: https://hackage.haskell.org/package/servant-auth-server-0.4.7.0/docs/Servant-Auth-Server.html#t:HasServer To be clear, I do think it is a design wart: the instance...

@tomsmalley Thanks for opening an issue ; as the author of the change I can attest that this indeed wasn't intentional (and I am truly sorry for any problem it...

It is extremely surprising to me that `NamedRoutes` would bring performance gains on _flat_ APIs (some gains on nested routes would have been less surprising IMHO). I am guessing `NamedRoutes`...

The `HasServer` instance for `NamedRoutes` still relies on `:` under the hood, see here: https://hackage.haskell.org/package/servant-server-0.19/docs/src/Servant.Server.Internal.html#line-929 We piggy-back on the instance of `ToServantApi api` which is a vanilla servant API.

@tfausak The only obvious difference between your two benchmarks is that in @Vlix's, all handlers are the same (`foo`). But maybe does it actually make a difference ?

Also, I notice that compilation times for the `:` routes do not seem quadratic at all in the number of routes in your benchmarks.

> tl;dr MY BAD! It's not NamedRoutes! It's the latest release of servant-(server-)0.19! The plot thickens :-) Off the top of my head, I can't think of any change in...

@Vlix I am not able to reproduce your results either. Compilation times for a module of 64 anonymous routes seems pretty much the same on `HEAD` and on `0.18.3`: `de923fc887f9e0197c8795fcb0e6967b196e0e17`...

Just tried with GHC 8.10.7 in a slightly modified servant repository — still can't reproduce the huge gains witnessed by some, so obviously I am not able to `git bissect`....