Messages in this thread | | | Date | Tue, 26 Feb 2008 21:14:42 +0100 | From | Sam Ravnborg <> | Subject | Re: [PATCH] Add rdc321x defconfig file |
| |
On Tue, Feb 26, 2008 at 10:44:51AM +0100, Ingo Molnar wrote: > > > > It is a simple equation: 10 additional defconfigs can give you more > > build coverage by additional people. Is it worth it? > > i dont think it's worth it on x86, because it has no real meaning so it > will just be an arbitrary thing that deteriorates over time. OK
> Subarchitectures on x86 are just a shortcut for the "0.1% of systems > that were lazy to be properly abstracted into the general PC code". We > are discouraging additional subarches and the one that got added > recently will go away soon. The rest is legacy. So you say we would waste our time doing build test of the legacy sub-archs. OK.
> > Really, we should concentrate our testing to where our _testers_ are and > where our developers are. > > And according to lkml, http://kerneloops.org and a 300,000+ sample > http://smolt.fedoraproject.org/static/stats/stats.html statistics, our > testers are distributed like this: > > - more than 90% of all kernel developers use general PC hardware > > - more than 95% of our active testers use general PC hardware > > - more than 99.1% of our distro users that are willing to send us > feedback use general PC hardware as well > So I will try to concentrate more on non-PC stuff from now on to increase the developer and tester base there. Seems we have plenty of developers and testers to hack and test PC stuff.
> (In that aspect i dont count the million(s?) of non-x86 Linux based > phones as "a million users", unless they become an active part of our > ecosystem and do things like hook into kerneloops.org. It's that simple, > really.)
The "everything on lkml argument"... At my current company I cannot name a single person that I think would be able to follow lkml - sorry!
Sam
| |