complicated downgrade
Valentin Nechayev
netch at lucky.net
Tue Jul 22 08:38:09 PDT 2003
Tue, Jul 22, 2003 at 09:20:30, bmilekic wrote about "Re: complicated downgrade":
> This sounds like the same symptoms as the latest USB problem...
> when/if you track -current or even run one of the 5.x releases, it's
> key to realize that this is very active code that you're running; it's
> not the same thing as running 4.x, for example. The code in 5.x is
> constantly actively changing, whereas the code in 4.x only receives
> comparatively well-regulated merges from 5.x, for the most part.
> Therefore, one of the things to always try is to update to the latest
> -current, rebuild, and see if you can reproduce. Chances are, your
> problem may have been fixed and, if not, at least we can be confident
> that it's reproducable on your hardware with the latest sources.
Well, I can do such mad rides on home machine, but not on remote collocation
in another country. Running fresh -current I can get a bunch of some other
problems which effectively will prevent system from running ;(
The most problem I see preventing having much more wide testbase for -current
is continuous nature of its committing. If it were developed, e.g., on
week pulse, with only fixing bugs since Thu till Mon, and providing
semi-stable snapshot on Mon, it can be more attractive to many users
which want to track -current but have no will to deal with permanent panics...
(Pulse iteration length can be arbitrary. One week is one averagely
reasonable value. 3 months, as in RELENG_4, is too long.)
-netch-
More information about the freebsd-hackers
mailing list