From mboxrd@z Thu Jan 1 00:00:00 1970 From: Robert Schwebel Date: Tue, 25 Oct 2011 08:34:44 +0200 Subject: [Buildroot] Some topics for the Buildroot Developer Day In-Reply-To: <87vcrdy5g4.fsf@macbook.be.48ers.dk> References: <20111024173815.2277c6dd@skate> <20111024165904.GB9199@pengutronix.de> <20111024214747.279c6e72@skate> <4EA5C53B.5000406@eukrea.com> <87vcrdy5g4.fsf@macbook.be.48ers.dk> Message-ID: <20111025063444.GI9199@pengutronix.de> List-Id: MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: buildroot@busybox.net On Tue, Oct 25, 2011 at 08:02:19AM +0200, Peter Korsgaard wrote: > Yes, I'm using buildbot at work for the defconfigs we use, and it > works very well - But the problem is how to get good test coverage of > our 700+ package / toolchain / kernel combinations? > > What I'm currently doing is simply lots of 'make randpackageconfig' > builds - But it generates lots of data, so I haven't found a good way > of putting this information online. When we started with tests, we quickly found out that you don't only have to be able to test, but also have to cope with what you'll find :) So if you find lots of broken things, you have to fix them, and that quickly becomes a lot of work. So our current strategy is to have a set of "real-life" configs (customer BSPs), plus corner cases (a config with all-yes). We build that for the most important platforms (x86/powerpc/arm) and on the kernels which have most relevance for us (linus-git, linux-next) on a nightly base. That doesn't cover everything, but seems to be a good compromise. rsc -- Pengutronix e.K. | | Industrial Linux Solutions | http://www.pengutronix.de/ | Peiner Str. 6-8, 31137 Hildesheim, Germany | Phone: +49-5121-206917-0 | Amtsgericht Hildesheim, HRA 2686 | Fax: +49-5121-206917-5555 |