couchdb-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Robert Samuel Newson <>
Subject Re: Implementation questions
Date Mon, 16 Jan 2017 23:45:54 GMT
The duplication is temporary, we'd like to address it (indirectly) in 3.0. users should focus
on the clustered interface, the use cases for the node-local interfaces are few and efforts
will be made to reduce that to zero.

I can't comment on the peruser bits except to confirm that the plugin is broken in 2.0.


> On 15 Jan 2017, at 15:29, Thomas GUILLET <> wrote:
> Hi all,
> Sorry the previous message left the draft state too quickly.
> Thanks a lot for the hard work already put into CouchDB!
> As explained in a message on the user mailing list, I wanted to use
> couchdb-peruser on CouchDB 2.0.0 so I submitted a patch on github
> (
> I am currently looking at deploying a CouchDB instance in the wild so
> I have been looking into securing the endpoints.
> * Is there a reason why _all_dbs is available to everyone?
> * I looked into the source code of the different building blocks,
> there seem to be a lot of duplication between chttp and couch_httpd.
> (My understanding is that the former is for the Cluster endpoint and
> the latter being for the node-only endpoint but I am surprised to see
> that much duplication).
> * Current peruser set the user as the admin of his database, would
> anyone like to have the ability to switch off that setting (user is
> only a member of his database) or add some granularity?
> Sorry, searching through the mailing list archives is a bit tricky, I
> only went a few months back to see if those questions were answered
> previously.
> Thanks a lot,
> Thomas

View raw message