[PATCH] Handle the case of rpmostreepayload + GRUB2

Brian C. Lane bcl at redhat.com
Thu Oct 16 23:15:39 UTC 2014


On Thu, Oct 16, 2014 at 07:05:52PM -0400, Colin Walters wrote:
> On Thu, Oct 16, 2014, at 06:31 PM, Brian C. Lane wrote:
> >
> > Take a look at the errorHandler code used in the other payloads:
> > 
> > if errorHandler.cb(exn) == ERROR_RAISE:
> 
> From what I can tell that's used when the code path could occur at
> runtime as a result of dynamic conditions: erroneous user input,
> corrupted packages, unreachable network addresses, etc.
> 
> This assertion could only occur if ostree was buggy - it should always
> write /boot/loader regardless of whatever the user input.  I could just
> remove the assertions I guess, there's lots of other equivalent bits
> that would also traceback.
> 
> Right?

Isn't that the same as if we hit corrupt packages from a repo?  Maybe
there needs to be a ostree sanity check someplace in the process that
would present a nice error.

I know we've got piles of places where we traceback, I'd just like to
head off any new ones when possible.

-- 
Brian C. Lane | Anaconda Team | IRC: bcl #anaconda | Port Orchard, WA (PST8PDT)


More information about the anaconda-patches mailing list