On Wed, 2012-11-07 at 07:51 +0000, "Jóhann B. Guðmundsson" wrote: > We can still test this and installing Fedora, and arguably we should be > doing that along other OS other than just Microsoft as well. > > We would just not have the Fedora release dependent upon the result from > those tests... Really. So if it were known that Anaconda WOULD hose an existing Windows install you would be ok with releasing? Or if Anaconda simply puked and failed to install a working, booting Fedora in the presence of a Windows partition? That is what a release grade product does? And not just Windows, if Anaconda failed to do it's job and install a working Fedora without destroying any other OS install it is a serious problem. If the newly installed Fedora doesn't work it is a fail. If it causes damage beyond itself it is a disaster. Testing for and preventing such a PR disaster SHOULD be a release criteria. Failing to even test means that not only would a broken Fedora be released upon a unsuspecting world, there wouldn't even be a warning in the release notes to point to. Is it even possible to test every corner case out there in the complex world we live in? Probably not. Doesn't mean that making a respectable attempt is a bad idea. We have suffered since day one with Microsoft's refusal to admit other products exist and merrily destroy other boot loaders without so much as an "I see something else in the MBR, should I overwrite it?" prompt. We are supposed to be the White Hats and care about users and all that, lets live up to our own book of rules.
Attachment:
signature.asc
Description: This is a digitally signed message part
-- test mailing list test@xxxxxxxxxxxxxxxxxxxxxxx To unsubscribe: https://admin.fedoraproject.org/mailman/listinfo/test
