[ https://issues.jboss.org/browse/JBIDE-21145?page=com.atlassian.jira.plugi... ] Nick Boldt edited comment on JBIDE-21145 at 12/4/15 4:22 PM: ------------------------------------------------------------- As to the notion that it's as fast or faster to do: * p2 diff check comparing composite with [some other composite which doesn't exist, or some aggregate which is by definition different*] * JBT aggregate build * JBT aggregate p2 director install test than to do: * JBT composite p2 director install test * JBT aggregate build * p2 diff check comparing local JBT aggregate with last published JBT aggregate Yes, you're right... it would be about as much time either way. But since we don't have a composite site snapshot against which to compare (the composite site is only ever a single "latest snapshot"), I can't see how you'd perform a p2diff check BEFORE doing a build of the JBT aggregate, then comparing that against the previous aggregate. So, *TL;DR*... I can set up a p2diff check of the aggregate repo after it's built and before it's published, even though you more or less said that the SHA check was good enough. See JBIDE-16970 for future discussion on that topic. \* - it's always different because the composite includes test plugins and the aggregate does not. So we're always testing a DIFFERENT install than the p2director (everything in JBT+Central+EA / JBDS+Central+EA, via headless install) and install-grinder (everything in JBT+Central+EA / JBDS+Central+EA, via UI install). was (Author: nickboldt): As to the notion that it's as fast or faster to do: * p2 diff check comparing composite with [some other composite which doesn't exist, or some aggregate which is by definition different*] * JBT aggregate build * JBT aggregate p2 director install test than to do: * JBT composite p2 director install test * JBT aggregate build * p2 diff check comparing local JBT aggregate with last published JBT aggregate Yes, you're right... it would be about as much time either way. But since we don't have a composite site snapshot against which to compare (the composite site is only ever a single "latest snapshot"), I can't see how you'd perform a p2diff check BEFORE doing a build of the JBT aggregate, then comparing that against the previous aggregate. So, *TL;DR*... I can set up a p2diff check of the aggregate repo after it's built and before it's published, even though you more or less said that the SHA check was good enough. See JBIDE-16970 for future discussion on that topic. * - it's always different because the composite includes test plugins and the aggregate does not. So we're always testing a DIFFERENT install than the p2director (everything in JBT+Central+EA / JBDS+Central+EA, via headless install) and install-grinder (everything in JBT+Central+EA / JBDS+Central+EA, via UI install). -- This message was sent by Atlassian JIRA (v6.4.11#64026)