Yes, this is a bug in etag generation, not a concerted effort to make them weak.
Sane solution sounds sane.
On Wed, Jun 29, 2011 at 2:54 PM, Robert Newson wrote:
> I think the sanest fix is to make view etags for include_docs=true use
> the original algorithm, so that they always change if the d
On Jun 29, 2011, at 11:54 AM, Robert Newson wrote:
> I think the sanest fix is to make view etags for include_docs=true use
> the original algorithm, so that they always change if the database
> changes. If you file a ticket, I can fix this in the morning.
OK! Filed https://issues.apache.org/jir
I think the sanest fix is to make view etags for include_docs=true use
the original algorithm, so that they always change if the database
changes. If you file a ticket, I can fix this in the morning.
B.
On 29 June 2011 19:22, Jens Alfke wrote:
>
> On Jun 29, 2011, at 11:13 AM, Jens Alfke wrote:
On Jun 29, 2011, at 11:13 AM, Jens Alfke wrote:
> So for example, I do this:
>
> $ curl -I http://localhost:5984/demo-addresses/_all_docs
Sorry, to be a good test the URL should have been like
http://127.0.0.1:5984/demo-addresses/_changes?include_docs=true
because without include_docs,
On Jun 29, 2011, at 8:02 AM, Robert Newson wrote:
> This is deliberate ("COUCHDB-799 - More granular ETags for views." --
> "ETags for views now only change when their underlying view index
> changes due to indexing or purges. ETags are also specific to each
> view.")
>
> The view is updated if
This is deliberate ("COUCHDB-799 - More granular ETags for views." --
"ETags for views now only change when their underlying view index
changes due to indexing or purges. ETags are also specific to each
view.")
The view is updated if and only if the emitted data actually changes.
B.
On 29 June 2
Hi Behrad,
can you open an issue for this? http://issues.apache.org/jira/browse/COUCHDB
Cheers
Jan
--
On 29 Jun 2011, at 13:02, Behrad Zari wrote:
> Dear Couchees,
> Consider a view on a CouchDB 1.1 private DB which emits: ( doc._id, 1 )
> 1) call the view with reduce=false&include_docs=true b
Dear Couchees,
Consider a view on a CouchDB 1.1 private DB which emits: ( doc._id, 1 )
1) call the view with reduce=false&include_docs=true by your browser!
2) update a document (with futon)
3) recall that view
You will see a cached-inconsistent document included in view results!
(according to _r