Anaconda, manual partitioning

[Clyde E. Kunkel]

On Tue, 25 Mar 2014 09:04:08 -0600
Chris Murphy <lists at colorremedies.com> wrote:

> 
> After aggressively defending the Anaconda team much of the weekend
> from people on users@ suggesting the installer was somewhere between
> useless crap and willfully designed to piss users off, I read this
> revelation:
> 
> https://bugzilla.redhat.com/show_bug.cgi?id=1079604#c6
> 
> If only it were Anaconda's design goal to intentionally piss users
> off, because that would be an upgrade from David Cantrell's
> philosophy of "giving people just enough rope to hang themselves."
> 
> That sucks. It's not what I signed up for, and it's incongruent with
> my motivations for testing the installer. Now I'd like those hundreds
> of hours back.
> 
> As this relates to Fedora QA, I think QA should suspend mandatory
> test cases, test matrix entries, and release criteria related to
> anything in Manual Partitioning. That is, any test case related to
> Manual Partitioning would be an optional test. And if it hasn't been
> done, or if it isn't passing, even if there's a crash, the release is
> not blocked. Fedora QA can retask freed up resources to other
> packages/projects that could use QA's help.
> 
> I'd just let Manual Partitioning be a free for all between the
> Anaconda team and its volunteer user base to decide both features and
> quality level, and leave Fedora QA completely out of it.
> 
> 
> Chris Murphy

Lots of unhappy manual partitioning users?  Well, then, maybe that
points to something that needs fixing or at least re-engineering with
well-defined specification/limits.  But, don't cut QA out of the loop on
any aspect of the installer if that is what you are suggesting.
"...suspend(ing) mandatory test cases, test matrix entries, and release
criteria related to anything in Manual Partitioning..." and making them
optional is the same as the tests not being done.  I believe I have seen
some recommendations on limited manual partitioning testing and I think
that should be persued rather than, in essence, abandoning user
testing of any part of anaconda.  I suspect BZs would still be filed
and would receive little attention since the release criteria are gone
for manual partitioning.