Am Montag, den 04.10.2010, 19:59 -0400 schrieb Devin Heitmueller: > On Mon, Oct 4, 2010 at 7:21 PM, hermann pitton <hermann-pitton@xxxxxxxx> wrote: > > thanks for the report and pointing to the details again. > > > > We can see, that my testings on four different machines and Dmitri's > > tests have not been enough. Mauro had the Dual card=78 version from me > > too at least for analog TV testing. > > > > And, that was on hg with most backward compat as possible. > > > > How good are our chances, to run in such and similar troubles in the > > future, in fact staying only on latest -rc, rc-git and in best case on > > -next stuff previously? > > > > It will all come down to the distros and such a bug fix might take just > > a year in the future regularly ... > > > > So, if the quality control was not even sufficient on hg, what will > > happen on latest -rc git stuff for that? > > > > Obviously zillions of people do much more prefer to crash around there > > than on hg ... ;) > > I think it's been made pretty clear: we don't give a crap about > whether users' PCs crash. Getting the code into the bleeding edge > kernel is the most important thing. Reducing maintainership overhead > is clearly more important than whether the code actually works. > > Forget about the hg backport system. We would rather get crap code > into the bleeding edge kernel where almost zero users will test it > than to put it into HG where there is actually a chance for users to > see the problems before it goes into the mainline kernel (except for > the 0.1% of users who are willing to install the latest bleeding edge > kernel and make it work with all their other hardware). > > Yes, we should all be prepared for lots of regressions being > introduced, and nobody notices them until the code is already in the > distros and has reached the masses. And then maybe if the users are > lucky the distro maintainers will backport fixes. > > It's been made pretty clear that reducing merge overhead is more > important than delivering a quality product. > > I'll stop hijacking the thread now. > > Devin Devin, you are always very welcome! I know for sure, that you know what you are talking about. Thanks, Hermann -- To unsubscribe from this list: send the line "unsubscribe linux-media" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
