Redoing PR for Travis

[rgrinberg]

@vbmithr

[rgrinberg]

Why do we have the ipaddr to string commit? Seems like it's breaking compatibility.

[vbmithr]

The ability to use hostnames (that are resolved by Async) was negated by the old interface. The Ipaddr.t provided was anyway converted as a string. In practice, if you want to connect to an URL in text form, how do you get an IP address from it with Ipaddr.t with Async?

This is indeed breaking API change but I think this is better. What do you think? If you don't, do you know any straightforward way to get an Ipaddr.t from a hostname (I'm not using Lwt) ?

[rgrinberg]

I see. That's indeed annoying. What about introducing a new set of
constructors that allow people to specify the host names directly?

Since this is a breaking change, I'd like to hear what other maintainers
think.

@dsheets @avsm @samoht

On 06/12, Vincent Bernardoff wrote:

The ability to use hostnames (that are resolved by Async) was negated by the old interface. The Ipaddr.t provided was anyway converted as a string. In practice, if you want to connect to an URL in text form, how do you get an IP address from it with Ipaddr.t with Async?

This is indeed breaking API change but I think this is better. What do you think? If you don't, do you know any straightforward way to get an Ipaddr.t from a hostname (I'm not using Lwt) ?

---

You are receiving this because you authored the thread.
Reply to this email directly or view it on GitHub:
#133 (comment)

[rgrinberg]

@vbmithr OK I've thought over this more and without more feedback I'm in favor of this change.

A couple of things still bug me however:

1. Ipaddr.t isn't used in any of the constructors. I suppose that's alright.
2. Our interface still isn't exposing the "async way" of constructing an endpoint which is the Tcp.where_to_connect type. Does it make sense to expose that in another constructor as well perhaps?

[vbmithr]

@rgrinberg Yes definitely we need the same constructor as in Async. I would even remove the existing one to leave only this one. For the Ipaddr.t not used, I don't see why it would be a problem.

[dinosaure]

Due to #311, this PR will be close. I think we fixed on #311 what we tried to do in this PR. Thanks!