backporting COUCHDB-1008 into 1.1.x

[Filipe David Manana <fd...@apache.org>]

Hi,

In 1.1.x we have a new configuration option named "nodelay" under the
the "httpd" section.
However, COUCHDB-1008 adds the option "socket_options" which is more
generic, allowing nodelay to be specified there amongst other socket
options.

My concern is that having "nodelay" as an isolated option in an
official release means that in 1.2 (and future releases), we would
still have to support it for 1.1 backward compatibility, which means
slightly more code.

Any objections about backporting this patch into 1.1.x?



-- 
Filipe David Manana,
fdmanana@gmail.com, fdmanana@apache.org

"Reasonable men adapt themselves to the world.
 Unreasonable men adapt the world to themselves.
 That's why all progress depends on unreasonable men."

Re: backporting COUCHDB-1008 into 1.1.x

[Jan Lehnardt <ja...@apache.org>]

On 31 Jan 2011, at 20:55, Filipe David Manana wrote:

> Hi,
> 
> In 1.1.x we have a new configuration option named "nodelay" under the
> the "httpd" section.
> However, COUCHDB-1008 adds the option "socket_options" which is more
> generic, allowing nodelay to be specified there amongst other socket
> options.
> 
> My concern is that having "nodelay" as an isolated option in an
> official release means that in 1.2 (and future releases), we would
> still have to support it for 1.1 backward compatibility, which means
> slightly more code.
> 
> Any objections about backporting this patch into 1.1.x?

No objections, I think making this solid and avoiding future work is
a good idea and the code change is not massive by any measure. Thanks
for considering the future-BC situation :)

Cheers
Jan
-- 

Re: backporting COUCHDB-1008 into 1.1.x

[Paul Davis <pa...@gmail.com>]

On Mon, Jan 31, 2011 at 2:55 PM, Filipe David Manana
<fd...@apache.org> wrote:
> Hi,
>
> In 1.1.x we have a new configuration option named "nodelay" under the
> the "httpd" section.
> However, COUCHDB-1008 adds the option "socket_options" which is more
> generic, allowing nodelay to be specified there amongst other socket
> options.
>
> My concern is that having "nodelay" as an isolated option in an
> official release means that in 1.2 (and future releases), we would
> still have to support it for 1.1 backward compatibility, which means
> slightly more code.
>
> Any objections about backporting this patch into 1.1.x?
>
>
>
> --
> Filipe David Manana,
> fdmanana@gmail.com, fdmanana@apache.org
>
> "Reasonable men adapt themselves to the world.
>  Unreasonable men adapt the world to themselves.
>  That's why all progress depends on unreasonable men."
>

+0

1.1 hasn't been released yet so I don't see too much of an issue in
back porting, especially if it saves future work.