http://en.wikipedia.org/wiki/Unusual_software_bug#Heisenbug
Let us know if you can catch it in action. On Wed, Jan 19, 2011 at 8:19 PM, Thomas Rampelberg <[email protected]> wrote: > For some reason today, couch isn't acting up. I'm gonna get a curl > session next time I see this happening. > > On Wed, Jan 19, 2011 at 8:37 AM, Paul Davis <[email protected]> > wrote: >> Most odd. Think you could put together a script or curl session that >> can reproduce the issue? >> >> On Wed, Jan 19, 2011 at 11:35 AM, Thomas Rampelberg >> <[email protected]> wrote: >>> I'm talking directly to couch and it's internal (talking over >>> localhost:5984). >>> >>> On Tue, Jan 18, 2011 at 4:35 PM, Paul Davis <[email protected]> >>> wrote: >>>> On Tue, Jan 18, 2011 at 7:29 PM, Thomas Rampelberg <[email protected]> >>>> wrote: >>>>> I'm on trunk and having some weirdness with views and doc revisions. >>>>> I've got a simple view (map + reduce) that is emit("key", null). My >>>>> URL looks along these lines: >>>>> >>>>> /database/_design/design/_view/my_view?key="key"&reduce=false&include_docs=true >>>>> >>>>> When I do this, after updating all the docs with an _update handler, I >>>>> get old revisions. This persists after doing a db compact and view >>>>> cleanup (and restart of the server). If I add revs=true to the url, >>>>> the view start returning the latest. I'm interested in whether this is >>>>> a design decision or just a bug? >>>>> >>>> >>>> That's sounds like its either a bug of some sort. My first guess would >>>> be that its a caching layer somewhere that doesn't get busted till you >>>> change the URL because I don't think anything in the view system >>>> actually uses the revs query string parameter. >>>> >>>> Hard to say without more knowledge on what sort of caching >>>> intermediaries you have setup whether that's internal or external to >>>> couch (if it is indeed caching). >>>> >>> >> >
