ubuntu-appstore-developers team mailing list archive

[James Tait <james.tait@xxxxxxxxxxxxxx>]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On 04/07/13 09:59, John Lenton wrote:
> Maybe include the info from the responseHeader in a header?

Maybe if it's requested.  But I tend to err towards the principle of
YAGNI, so unless someone has a specific need for it, I don't think we
should bother.

> Me, I wouldn't -- if we can avoid parsing the solr response
> entirely, even better.

Well, yes, ideally we'd just be able to deploy and configure a Solr
server and be done with it. ;)  As you pointed out on IRC, there is
the omitHeader request parameter to get rid of the responseHeader, but
there are still reasons to do a little work on the server side:

 - Request/response sanitisation.
 - Stable API - schema changes shouldn't affect clients.
 - Solr doesn't support embedded objects.
 - Non-semantic changes to the document (e.g. whitespace) shouldn't
   change ETags (and thus break caching).
 - Solr's pagination lacks any navigability.

AFAICT, the document transformers we have available OOTB aren't going
to help with any of this.  We *could* implement our own transformers,
but I think while we're still figuring out what we need it's easier to
rapidly iterate in Python - we can optimise later.

Cheers,

JT
- -- 
- ---------------------------------------+--------------------------------
James Tait, BSc                        |    xmpp:jayteeuk@xxxxxxxxxxxxxx
Programmer and Free Software advocate  |        Tel: +44 (0)870 490 2407
- ---------------------------------------+--------------------------------

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.12 (GNU/Linux)
Comment: Using GnuPG with undefined - http://www.enigmail.net/

iEYEARECAAYFAlHVUsEACgkQyDo4xMNTLibRWACfa+rUxJVODcnagtTE9vcezlfK
FooAn3lrcp9VrLgAKpEo1DoW+aonSoB+
=aalR
-----END PGP SIGNATURE-----