[ https://issues.jboss.org/browse/JBIDE-15766?page=com.atlassian.jira.plugi... ] Andre Dietisheim edited comment on JBIDE-15766 at 10/24/13 6:36 AM: -------------------------------------------------------------------- [~mwigglesworth-redhat] very valid concerns, did not think about them when rushing for the rest of the UI, thanks for fiddling this out. But thinking this API through a 2nd time I pretend that the whole backend API is broken. The general rule of thumb is that you need the diff to be done as close as possible to the peer that stores it. Having to do the diff in the remote client is imho the worst possible solution, it requires the client to constantly refresh while not gaining any significant guarantee, just increasing likelyhood. The constant refreshing also forces the user program (a UI or any other) to also constantly update it's env vars. My personal opinion is that - also given the normal pattern of explicit refreshes we have - we should not refresh constantly. We could eventually refresh when there's an error, sticking to the explicit refresh. On the long run we should IMHO ask the OpenShift team to change this API and switch to an API where we submit the whole map of vars at once, letting the backend fiddling out what operations are required (removal, addition, update) was (Author: adietish): [~mwigglesworth-redhat] very valid concerns, did not think about them when rushing for the rest of the UI. But thinking this API through a 2nd time I pretend that the whole backend API is broken. The general rule of thumb is that you need the diff to be done as close as possible to the peer that stores it. Having to do the diff in the remote client is imho the worst possible solution, it requires the client to constantly refresh while not gaining any significant guarantee, just increasing likelyhood. The constant refreshing also forces the user program (a UI or any other) to also constantly update it's env vars. My personal opinion is that - also given the normal pattern of explicit refreshes we have - we should not refresh constantly. We could eventually refresh when there's an error, sticking to the explicit refresh. -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira