* Rusty Russell <rusty@rustcorp.com.au> wrote:

quoted text
> > IMHO, an infrastructure change of this magnitude should absolutely 
> > be done via the Git space. This needs a ton of testing and needs 
> > bisection, a real Git track record, etc.
> 
> Not yet.  Committing untested patches into git is the enemy of 
> bisection; if one of my patches breaks an architecture, they lose the 
> ability to bisect until its fixed.  If it's a series of patches, we 
> can go back and fix it.

while the initial series might be rebased once or twice, beyond the 1-2 
days of initial integration and testing i dont think that's true, and 
i'm doing up to 3-4 bisections a day just fine, on an append-mostly 
tree.

if you have trouble turning a Git tree into a bisectable tree then your 
testing-fu is not strong enough ;-)

[ the only plausible danger is to architectures that are not used by 
  testers all that much (so that breakages can linger a lot longer 
  unnoticed) - but why should the other 99% of Linux users be put at a 
  disadvantage for them. ]

	Ingo
--
unsubscribe notice
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/