[ https://issues.jboss.org/browse/JBIDE-13671?page=com.atlassian.jira.plugi... ] Max Rydahl Andersen commented on JBIDE-13671: --------------------------------------------- so i'm all for having reliable and reproducible build qualifiers but got some questions to these claims: {quote} we can trust that the timestamp shows when the plugin was ACTUALLY last changed in Github.{quote} yes, but we will also now be generating identical filenames for *different* content, especially since we still have the parent pom pointing to snapshot's and uses overrides in mvn commands from jenkins. We'll now no longer be able to trust which binary contains what. i.e. what happens when we've rebuilt hibernate tools multiple times using the same SHA1 but against changing parent pom/TP's making different binaries under same name be available in a cmompsite on the updatesite. How do we know which binary users *actually* have when their name is the same ? {quote} So yes, it's worth it. We've been trying to make this change for 3 years. We tested it for Locus and it works great there. No reason not to roll it out to the rest of the team.{quote} No, in Locus we don't have an ever cahnging parent pom thus this is not comparable IMO. Also Locus does not have branches - it is just one stream so you don't get builds that mixes up branches as part of the daily work. Thus i'm struggling to see what has changed since we tried last time to apply this ? What makes it better to use this if we still generate different content ? -- This message was sent by Atlassian JIRA (v6.4.11#64026)