Caching issue - Githubissues

buddycloud / buddycloud-server-java

The buddycloud server written in Java.

http://buddycloud.com

Apache License 2.0

67 stars 40 forks source link

Caching issue #45

Open lloydwatkin opened 11 years ago

lloydwatkin commented 11 years ago

With the lazy way that caching works on the java server the following situation causes an issue:

Server started
Remote data cleaned out
Notification of item comes in from external server
User loads channel
Channel seen as cached therefore we don't go remote to get additional data

Infinite scroll from the webclient should solve this but it won't kick in if there's no scroll bar it won't kick in. However if I reduce the size of the browser and kick in the scroll bar it kicks in RSM and retrieves remote posts.

Maybe the java server should have a flag to say it hasn't retrieved posts from a remote node after start up and go off and get posts if required when first requested.

lloydwatkin commented 11 years ago

<you-missed-something/> not handled either

imaginator commented 11 years ago

What's the benefit of not getting the posts at startup vs getting them at the first time a client calls the /sync endpoint?

On 21 August 2013 15:11, Lloyd Watkin notifications@github.com wrote:

not handled either

— Reply to this email directly or view it on GitHubhttps://github.com/buddycloud/buddycloud-server-java/issues/45#issuecomment-23015723 .

Simon Tennant | buddycloud.com | +49 17 8545 0880 | office hours: goo.gl/tQgxP

lloydwatkin commented 11 years ago

/sync wouldn't trigger a retrieval of remote data (at least unless the API server does something special).

I'm not saying there's a benefit at all, this is just how things are implemented at present. I've just come off a call with @ashward about this saying we should all have a chat (@abmargb included) about what to do with this. As far as I understand the nodejs server implementation of the behaviour isn't idea either?