From thewoolleyman at gmail.com Sun Jan 4 02:45:12 2009 From: thewoolleyman at gmail.com (Chad Woolley) Date: Sun, 4 Jan 2009 00:45:12 -0700 Subject: [Cruisecontrolrb-developers] Git project building over and over even though there are no commits In-Reply-To: References: Message-ID: Update: Just committed some changes, I think everything is working now - I'm running it on a couple of live svn and git servers. To recap: * Changes here: http://github.com/thewoolleyman/cruisecontrol.rb/tree/master * Contains fixes for these bugs: * http://cruisecontrolrb.lighthouseapp.com/projects/9150/tickets/221-git-error-short-commit-ids-are-not-always-unique * http://cruisecontrolrb.lighthouseapp.com/projects/9150/tickets/212-builder-failure-if-svn-log-fails-with-credentials-error Note my last comment on the second ticket - I don't think the current approach is the best, but I didn't want to make drastic changes. We should really check the return code instead of just relying on the existence of stderr. Jeremy/Alexey, can one of you look at these and apply if they look good, or let me know what needs to be changed if not? Thanks, -- Chad On Fri, Dec 26, 2008 at 10:39 PM, Chad Woolley wrote: > ffmike stepped up and patched this one, seems to be working for me in > my branch, which also includes an (imperfect) fix for the bad return > code detection: > > http://github.com/thewoolleyman/cruisecontrol.rb/tree/master > > > On Wed, Nov 19, 2008 at 3:09 AM, Chad Woolley wrote: >> Found the root cause of this (short commit IDs are not unique, and ccrb >> still doesn't detect bad return codes from source control commands): >> >> http://cruisecontrolrb.lighthouseapp.com/projects/9150-cruise-control-rb/tickets/221-git-error-short-commit-ids-are-not-always-unique >> >> On Wed, Nov 12, 2008 at 1:28 PM, Chad Woolley >> wrote: >>> >>> Updating to latest ccrb trunk didn't help, restarting cruise/box didn't >>> work - it just builds every 5 minutes. Working copy looks clean. >>> >>> Anyone seen this? >>> >>> This is on the new Rails CI server, if anyone wants to help debug it I can >>> give access. I don't have time right now, waiting to see if it goes away on >>> the next commit. >>> >>> -- Chad >> >> >