On 24-06-2005 15:25, "Stefan Bodewig" <[EMAIL PROTECTED]> wrote: > On Fri, 24 Jun 2005, Adam R. B. Jack <[EMAIL PROTECTED]> wrote: > >>> Yes, I look for module failures and even worse for "module success >>> but with warnings". The later usually means stuff has been moved >>> in svn, we are now unable to check it out, but Gump doesn't >>> consider it a failure. >> >> Want this to go away now that SF.net have their CVS act together & >> we are doing a lot of SVN migrations? > > No, it doesn't go away.
FWIW, the algorithm I have in my head (come to apachecon! We can talk about it! :-)) would mean that a migrated project would be "*-ed", ie built and linked against but considered to have an error, and that the corresponding module would be considered failed, and that the failure cause for the project would point to the module. The following e-mails would be sent: - to the module "owner" about the problem (every day) - to the project "owner" about the problem if it doesn't go (after n days...in the event module "owner" and project "owner" are the "same" its just one e-mail listing two problems) - to the gump list in a summary e-mail (every day) that lists the problems with causes, the owners of the affected modules, and the duration in that state Basically, we make "modules" (and repositories and other parts of the object model) first-class citizens in the gump world, and support notification about them just as well or just as bad as for "projects". Cheers! Leo --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
