At Wed, 31 Aug 2011 21:19:52 +0300, Jukka Ruohonen wrote: > > On Wed, Aug 31, 2011 at 07:24:26PM +0200, Marc Balmer wrote: > > If only more developers would develop a sense of quality and at least > > test-compile their stuff, or run a relase build, before committing these > > figures would already look a lot nicer... > > Sense of quality has very little to do with build breakages. Statistically > speaking, most of the failures are due trivial things (prime example: set > lists and other oddities in the build system). (Also: not all of us have > 12-core machines that would be required to do release builds before each > commit, possibly on several different architectures.) > > Regressions on the contrary are a real issue practically in every operating > system project. Fortunately NetBSD has made some good progess at this front. > > Loganaden and others who may want to try and track -current, a good way to > help is to watch the automated test results and report possible regressions. > For daily runs see: http://releng.netbsd.org/. It might also be a good idea to automate integration testing before committing and reject failed integrations. There is an IMNSHO excellent tool in Aegis (aegis.sf.net). Best regards, Marko
Attachment:
pgp1v24CUv6VD.pgp
Description: PGP signature