Removing JS extractor support from Search

Ryan Zezeski rzezeski at basho.com
Fri Aug 12 00:39:38 EDT 2011


Riak Users,

Recently it has come to our attention, thanks to Greg Pascale, that the JS
extractor support for Search is broken.  As I was reviewing a patch to
rectify this I came to the conclusion that this support should be removed
altogether.  A few reasons being:

1. This feature has never worked. This alone wouldn't be such a good reason
but my point is that removing this feature wouldn't break any existing users
because it never worked to being with.

2. The docs on how to use this feature aren't entirely clear and the code to
support it is a little ugly IMO.

3. Like any other JS support in Riak it has to pay the penalty of shipping
bytes back and forth between VMs.  There is also potential for contention on
the JS VM side if the pool is not big enough.

4. Finally, the main reason I can imagine anyone using JS extractors is to
pull apart incoming JSON objects.  This is already handled by Search via an
Erlang extractor that is invoked any time it sees an 'application/json'
content-type header.  This will be much more efficient than using a JS
extractor.

If anyone can think of any reasons that they feel strongly that this
functionality should not be removed then please speak now.

-Ryan
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.basho.com/pipermail/riak-users_lists.basho.com/attachments/20110812/58ae4883/attachment.html>


More information about the riak-users mailing list