On Mon, Nov 16, 2009 at 1:00 PM, Chris Anderson <[email protected]> wrote:
> On Thu, Nov 12, 2009 at 7:04 PM, Noah Slater <[email protected]> wrote:
>>
>> On 12 Nov 2009, at 23:05, Chris Anderson wrote:
>>
>>> Using a different engine ought to be trivial. Thanks for keeping us in the
>>> loop.
>>
>> What about the curl bindings?
>>
>
> They aren't strictly necessary unless you want to run the headless
> test suite. I think they still count as mostly trivial, the HTTP
> interface they expose is pretty simple, so doing it with something
> other than curl would work.
>
>
> --
> Chris Anderson
> http://jchrisa.net
> http://couch.io
>

As soon as I commit the couchjs refactor it should be extremely easy
(relatively speaking) to replace the HTTP interface with any sort of
bindings that are desired. I made sure to have the HTTP interface
enabled by a single function call so that the command line switch to
enable it would be simpler. A couple #ifdef's and that could easily be
swapped to a ./configure defined implementation. There's also the
possibility of having a compile time disabling of the HTTP support
which I contemplated for making the cURL dependency optional.

Also, the new HTTP interface mirrors the synchronous XHR Object's API
so it should hopefully be a bit more natural for JS people to use.

Paul Davis

Reply via email to