[PATCH] Handle the case of rpmostreepayload + GRUB2

Colin Walters walters at verbum.org
Thu Oct 16 22:10:01 UTC 2014



On Thu, Oct 16, 2014, at 06:03 PM, Brian C. Lane wrote:
> On Thu, Oct 16, 2014 at 10:17:12AM -0400, Colin Walters wrote:
> > +        boot_extlinux = boot + '/extlinux'
> > +        if os.path.isdir(boot_extlinux):
> > +            boot_syslinux = boot + '/syslinux'
> > +            boot_loader = boot + '/loader'
> > +            assert os.path.isdir(boot_loader)
> 
> Why the assert? If that fails it's going to blow up the installation
> with a traceback. It would be better to handle the error in some clean
> way.

What would be a clean way?  I can't think of anything to do other than
to fail hard.

This assertion actually isn't the most likely to trip - what seems more
likely to break is grub2/extlinux change the location of their files and
rpmostreepayload isn't changed.

But the only way I can think of to handle that is automated testing.


More information about the anaconda-patches mailing list