Mailing-List: contact user-help@couchdb.apache.org; run by ezmlm
Precedence: bulk
Reply-To: user@couchdb.apache.org
Received-SPF: neutral (athena.apache.org: local policy)
Sender: J Chris Anderson <jchris@couch.io>
Content-Type: text/plain; charset=us-ascii
Mime-Version: 1.0 (Apple Message framework v1081)
Subject: Re: Query parameters in rewrites.json not handled as default values
From: J Chris Anderson <jchris@apache.org>
In-Reply-To: <AANLkTikY9xqFyO0Oo3P7TK960xSzF=RcNSpkdm-Udvo5@mail.gmail.com>
Date: Sun, 12 Sep 2010 10:16:35 -0700
Content-Transfer-Encoding: quoted-printable
Message-Id: <2714FE12-54B0-4CBF-8F9A-2F66FD944991@apache.org>
References: <AANLkTim8a9CuPu+Eu9fG39FJgpu3S+1j6GXrKgWQ1CfX@mail.gmail.com>
 <F5DAC449-5346-4194-A6A3-0ABD9A3998AD@apache.org>
 <AANLkTimvYk6UMVS0pNNoNDAtqZEXHjTtM261aMXsRBs6@mail.gmail.com>
 <AANLkTi=a8+yVQ11gDDtckHfcOeNFaTQyzK9NiOdoYp3E@mail.gmail.com>
 <AANLkTikY9xqFyO0Oo3P7TK960xSzF=RcNSpkdm-Udvo5@mail.gmail.com>
To: user@couchdb.apache.org


On Sep 12, 2010, at 9:00 AM, Benoit Chesneau wrote:

> On Sun, Sep 12, 2010 at 5:43 PM, Henrik Skupin <hskupin@gmail.com> =
wrote:
>> On Sun, Sep 12, 2010 at 4:08 AM, Benoit Chesneau =
<bchesneau@gmail.com>wrote:
>>=20
>>>=20
>>> Well it will be a regression then. It has worked like this, and =
after
>>> your request, we discussed about changed it . The point is that you
>>> didn't want that someone can override values defined in the ddoc and =
I
>>> think you had a case for this.
>>>=20
>>> I'm not sure what is the right way to handle it right now and we
>>> should consider all the usecases before changing anything.
>>>=20
>>=20
>> I'm happy to file a bug report but I agree. If that behavior has been
>> changed on purpose, it would be good to know the reasons. Do you have =
the
>> bug # or the changeset, which has been changed this behavior, so we =
have a
>> reference point?
>>=20
>> --
>> Henrik Skupin
>> QA Engineer
>> Mozilla Corporation
>>=20
>=20
> I thas been discussed on pre-release cycle unfortunatly :/ But If I
> remember  the question about using query as a default or a way to fix
> variable was around the question : "How do we make sure the
> startkey/endkey for this path aren't rewritten?". Which is I think a
> good question. =46rom time to time you want to make sure that the path
> /key is always the patch routing the key for this list. or /about go
> to a doc with some id.
>=20
> Instead of using query for the defaults, I would suggest to introduce
> a `default` member to the rule or maybe a way to set as a default
> value in the query like :
>=20
> "query": {
>  "key": { "default": 1, "value": ":var"}
> }
>=20
> I've a preference for a `default`member though, it eases the code imo.
>=20

yes I agree default is easier (and also backwards compatible to just =
leave the existing query definition alone).

Chris

> - benoit