On Fri, 28 Aug 2009, Tim Verhoeven wrote: > We have a project for which we are could use some help. Its about > creating a link between our Mantis bugtracker (bugs.centos.org) and > upstreams Bugzilla (bugzilla.redhat.com). At the moment we need to > manually track all upstreams bugs reported in our own bugtracker. This > is a time-consuming process and it something that can be automated. So > why not do it. I'd say "why do it at all". > My idea was to create a daemon of some sort that at regular intervals > checks all bugs in our systems that have a upstream bugzilla number > entered (Mantis support customs fields that could be used for this). > If there is a update in the bugzilla since the last run it is copied > into the bug in Mantis. That why we can get notified automatically and > can react to the updates more quickly and spend less time needing to > follow up. Firstly I'd say that CentOS should respond as quickly as possible to all upstream updates, and there is no need to monitor their bug tracker to do that. In fact, doing anything at all other than quickly building and verifying upstream updates will actually slow down how quickly CentOS can respond to upstream updates (given finite resources). However, if you really did insist on doing this, then you want something which is event driven, not done by polling. The upstream bug tracker (do we really need to play this "upstream" mystery game?) allows registration of email addresses for each relevant bug. Create a contribs.org email address for a bugzilla notification robot, and then when things happen upstream, the robot can attach new information to Mantis. Or just attach a bugzilla URL to mantis, then close the bug, with status "Upstream". Anyone visiting Mantis who wishes to know status can then just click the link, and get the latest news straight from the horse's mouth. And no need for CentOS worker ants to spend any time at all on followup. --- Charlie