Re: [boost] [beast] Chunking example

Subject: Re: [boost] [beast] Chunking example
From: Vinnie Falco (vinnie.falco_at_[hidden])
Date: 2017-07-02 18:11:03

On Sun, Jul 2, 2017 at 10:54 AM, Bjorn Reese via Boost
<boost_at_[hidden]> wrote:
> If the above is used to read a chunked transfer, what happens to
> chunk-ext fields? Are they inserted as header fields or are they
> lost?

Chunk extensions are not valid HTTP headers. `beast::http::parser`
does not store them in the `basic_fields`. It doesn't store them at
all, they are simply discarded.

>From <https://tools.ietf.org/html/rfc7230#section-4.1.1>
"The chunked encoding is specific to each connection and is likely to
be removed or recoded by each recipient (including intermediaries)
before any higher-level application would have a chance to inspect the
extensions. Hence, use of chunk extensions is generally limited to
specialized HTTP services such as "long polling" (where client and
server can have shared expectations regarding the use of chunk
extensions) or for padding within an end-to-end secured connection.

To my understanding, chunk-extensions are a rare niche use-case with
meaning only to applications using a custom interpretation at each end
of the connection. In fact 5 years ago the IETF almost deprecated
them:
<https://trac.ietf.org/trac/httpbis/ticket/343>

Beast doesn't go out of its way to help you get at the extensions, but
it also doesn't make it impossible.

On the other hand, I do not have significant expertise with HTTP
servers; if a compelling use-case presents itself this is an aspect of
the library which may be improved, in a backward-compatible way.

> Given that this is an example of incremental reading, why does it
> use read() rather than read_some()?

The contracts for those functions are as follows

`read` continues until there's an error or the message is complete

`read_some` continues until it gets at least one byte, an error
occurs, or the message is complete

In the example I posted the behaviors are very similar since the
buffer has a 512-byte limit. Either would work.