System testing

[William Muriithi]

>Also a few years back (at the time of 2.4 and early 2.6) compiling the
> kernel was considered to be a good hardware test. The code was considered
> to be so clean that it would always compile on a properly working system.
> So if you see compilation errors - it's the hardware. On the other hand
> I tried to compile a kernel recently (about half a year ago) and it grew
> into a X00MB monster during the compilation and for the reason I don't
> remember now it didn't compile and I dropped the idea.
>
Apology as this is a little off topic as it has nothing to do with system
testing but ...

Hmm, you implying the kernel quality has gone down now? My impression is
the opposite.  That impression is arrived at by listening to Linux
symposium recordings,  so may be wrong. But I doubt it, most kernel dev are
now old guys with decades of kernel development experience, so that should
at least improve the code quality

> BTW, then was the last time you compiled the kernel for an i386/x84
system? ;)
>
Long ago.  I still remember the reason I stopped.  It was a paragraph on
RedHat documents that mentioned using self compiled kernel risk exposing
you to bugs that are triggered by some compiler flags but not in redhat
provided kernel. That mean most of the Google search would not help as your
problem would be unique

That being said, may be good to compile kernel as long as its not used on
production system.  That way, we can catch and resport those bugs.
> Regards,
> Anthony
> --
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://gtalug.org/pipermail/legacy/attachments/20121125/ec8a9262/attachment.html>