No, I think I understand well the difficulty of tracking down those kernel bugs. I only object to the habit among developers of asking you to try a later revision, even when there's no reason to believe that the newer version fixes the problem. For example, in my recent experience where I was asked to try the kernel.org vanilla kernel, there was absolutely no delta whatsoever between the Ubuntu version and the kernel.org version of the driver in question. So there was no possibility that the newer version even fixed my problem.
Anyway, my complaint is not the availability of the new software. I think it's fantastic that users will have the option to easily switch to the kernel.org tree. What I _am_ worried about is the ability to track the number of defects in the Ubuntu kernel will be degraded, because devs will ask for spurious tests of other revisions and then set bugs to Incomplete.
Posted Mar 3, 2009 7:20 UTC (Tue) by khim (subscriber, #9252)
[Link]
For example, in my recent experience where I was asked to try
the kernel.org vanilla kernel, there was absolutely no delta whatsoever
between the Ubuntu version and the kernel.org version of the driver in
question. So there was no possibility that the newer version even fixed my
problem.
Driver can easily trigger bug in stock kernel and if this bug will be
fixed problem will disappear. For this reason it was long policy of kernel
guys to reject any and all bug-reports with nVidia-tainted kernel, for
example. I see nothing wrong with asking guys to test vanilla kernel -
packages should just make it easier...
What I _am_ worried about is the ability to track the number of
defects in the Ubuntu kernel will be degraded, because devs will ask for
spurious tests of other revisions and then set bugs to
Incomplete.
Or they can just add one more item to submission form: "what happend
with Linus stock kernel"... This way submitter will know what to do from
the start...