Re: [boost] [beast] message bodies

Subject: Re: [boost] [beast] message bodies
From: Vinnie Falco (vinnie.falco_at_[hidden])
Date: 2017-07-02 19:34:46

On Sun, Jul 2, 2017 at 11:52 AM, Mike Gresens via Boost
<boost_at_[hidden]> wrote:
> Am 02.07.2017 um 20:34 schrieb Vinnie Falco via Boost:
>     response<string_body> req;"
>
> nothing to change here since
>
>     using string_body = beast::http::sequence_body<std::string>; //
> convenience decl in beast

Phew, okay, so we are on the same page. I don't want users to have to
write `beast::http::const_sequence_body<std::string>`

What you are asking for is a feature request, which requires no change
in interface to any algorithm. That's a much easier ask. You already
proposed two containers which were merged to the master branch in the
example/common directory (thanks for your contribution!)

<https://github.com/vinniefalco/Beast/blob/78a065ba39836d91d7e70d93de7f9140f518083b/example/common/const_body.hpp>

<https://github.com/vinniefalco/Beast/blob/78a065ba39836d91d7e70d93de7f9140f518083b/example/common/mutable_body.hpp>

These are not public interfaces yet, but I could be convinced to make
them public interfaces if:

1. There is established practice that is either standardized or close to it [1]

2. The names and implementation of the concept checks either use
existing standard library facilities if they exist, or provide an
implementation that tracks the standard (or proposal)

3. The documentation (javadoc comments and exposition in the
documentation) meets the quality requirements of the library

4. There is general support (which there very likely already is)

I don't mind doing this work myself but I am not in a hurry to do so
since existing use of Beast will be unaffected and there are other
priorities. On the other hand if someone else wants to submit a pull
request, I wouldn't turn it away!

[1] This might be related
http://www.open-std.org/jtc1/sc22/wg21/docs/papers/2014/n3884.pdf