Re: [cpp-netlib-devel] Http Server comment

[Glyn M. <gly...@gm...>]

Hi Erik,

2010/1/22 Nelson, Erik - 2 <eri...@ba...>

> Thanks for the help... It's working now, and I'm playing with it some
> more.
>
> I'm not sure if it matters, but in connection.hpp, in the connection
> constructor, there's a bit of code like
>
>            try {
>                socket_.set_option(tcp::no_delay(true)); // Don't delay
> writing
>            } catch (system::system_error & e) {
>                handler_.log(e.what());
>            }
>
> And the set_option call *always* seems to throw
> boost::asio::error::bad_descriptor.  Not sure if it matters.
>
> One other thing I noticed (and don't personally care about) is the
> cpp-netlib coding style, like curly brace placement.  I'm not sure if
> there are boost coding guidelines or not, or if they are enforced or
> not, but cpp-netlib looks a little different than other boost source
> code.  Again, I'm perfectly happy with however it is, but if the goal is
> to move cpp-netlib into boost and boost cares (I don't know that either
> of those assumptions are true), it's probably easier to start doing it
> "the boost way" sooner rather than later.
>
>
There isn't really a "boost way", I think, the closest it gets is in the
guidelines:

http://www.boost.org/development/requirements.html

But that stops short of specifying how the source code should be formatted,
but focuses on portability and license issues.  Different boost libraries
seem to be formatted differently, depending on each author's preferences.

The coding style sometimes bothers me too (not so much curly braces
placement as indenting), but I get that all the time when I look at other
people's code ;)  If you think its a going to be a hindrance then perhaps we
could open a discussion about guidelines.

Maybe consistency would be important at least, because this project could
potentially be large and the nature of the design doesn't lend itself to
highly readable code.


Glyn