The more I think about this the more useful I think it would be. I'd say if you are willing to spend the time to set it up and maintain it go for it. You might want to see if anyone else has an opinion.
david jencks On 2002.09.18 11:51:44 -0400 Michael Rettig wrote: > > ----- Original Message ----- > From: David Jencks <[EMAIL PROTECTED]> > > Aside from using cruise control, how does this differ from what chris > > kimpton already has set up(http://www.lubega.com/), which runs the > > testsuite at least once a day and I think does hourly builds? How does > > your proposed system distinguish a broken build from a lengthy > multi-part > > checkin where writing the comments on the next chunk takes 10 minutes? > > CruiseControl has a "quiet period" which means that it waits for a > developer to stop checking in code before it starts it's build. Once no > checkins have been detected for x amount of minutes, then the build > starts. > > I've been the build master for Java projects with 700k loc and 50+ > developers all actively checking in and it is extremely rare that the > build starts mid checkin. When it does, it usually fixes itself within a > few minutes. > > CruiseControl differs in that it ties the build to checkins, so you can > very easily address accountability for broken builds. The web page that > is generated will give you a list of files that were checked om. More > importantly, there are names attached to those files, so you can very > easily see who checked in Foo.java that does not compile or > SimpleTest.java that fails. > > In a big project when tests break, accountability needs to be established > and quickly - and this means when, how, and who. Broken windows cannot be > tolerated whatsoever. > > Although, http://lubega.com will tell you what is broken, and > approximately when it was broken. It can't link the broken build to a > specific checkin and ultimately to the offending party. > > > > > How would your system apportion blame when there are several checkins > at > > the same time (either interlaced or within less time than it takes to > > checkout and do a clean build) and the build breaks? > > The system cannot be perfect, and tell you exactly who broke what, > especially with interlaced checkins. However, these situations are often > rare, and continuous integration with CruiseControl will give you enough > information that you can figure out what happened and solve the problem > within a few minutes. I've been in these situations before, and its never > taken me more than 15 minutes to trace through the CruiseControl logs and > find the offending developer within a few minutes. > > Hope this helps. > > Mike > -- > __________________________________________________________ > Sign-up for your own FREE Personalized E-mail at Mail.com > http://www.mail.com/?sr=signup > > > > ------------------------------------------------------- > This SF.NET email is sponsored by: AMD - Your access to the experts > on Hammer Technology! Open Source & Linux Developers, register now > for the AMD Developer Symposium. Code: EX8664 > http://www.developwithamd.com/developerlab > _______________________________________________ > Jboss-development mailing list > [EMAIL PROTECTED] > https://lists.sourceforge.net/lists/listinfo/jboss-development > > ------------------------------------------------------- This SF.NET email is sponsored by: AMD - Your access to the experts on Hammer Technology! Open Source & Linux Developers, register now for the AMD Developer Symposium. Code: EX8664 http://www.developwithamd.com/developerlab _______________________________________________ Jboss-development mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/jboss-development
