update some testcases

Adam Williamson awilliam at redhat.com
Tue Jan 8 20:22:18 UTC 2013

On Tue, 2013-01-08 at 03:02 -0500, Kamil Paral wrote:
> > >I think we should create a separate wiki page "QA:Anaconda
> > >partitioning", that explains the differences between the "guided
> > >partitioning" and "custom partitioning", and also between
> > >>"autopart" and "manual partitioning". Ideally with screenshots.
> > >Then we can link that page and simply ask the tester to encrypt the
> > >disk using method A or B (or any of them). In >this case, I think
> > >we only mandate "autopart", nothing else.
> >    Hi Kamil,
> >    I feel a little confused,in"guided partitioning mode",we do"auto
> >    partitiong",in"custom partitioning mode",we do"manual
> >    partitiong",
> >    actually,there is no "guided partitioning mode" or"custom
> >    partitioning mode",we do default install then we can
> >    call it"guided partitiong mode" if we change something,we can say
> >    it in"custom partitiong mode", As for "auto
> >    partitioning",we only need to click"continue".Am I right?
> Guided mode is the screen where you can only preserve or delete
> partitions, nothing else. Custom mode is the screen where you can set
> up the partitions to your will.
> Automatic partitioning ("autopart") is an approach of letting Anaconda
> create the partition setup automatically for you. It is used in guided
> mode and _can_ be used in custom mode (by clicking the blue text
> "Create partition layout automatically"). Manual partitioning is the
> approach of doing everything by yourself, basically it is a custom
> mode without using autopart.
> I'm using the terms in the sense I see them most often used. But maybe
> different terms could be less confusing. Adam, do you want to comment
> on terminology?

That's how I've been doing it so far. It feels like we could somehow
come up with better names, but no-one has yet :)

The dialog which pops up after you select disks for installation - and
either says 'You have enough space, you're all set!' or 'You don't have
enough space, you need to free some up!' - is the branch point for
'Guided' vs. 'Custom' install. Both versions of that dialog have a
checkbox whose label says something about 'Customize the layout'. If you
check that checkbox and then proceed from the dialog, you are now in the
'custom partitioning' flow. If you leave the checkbox unchecked and
proceed from the dialog, you are now in the 'guided partitioning' flow.
If you had enough space, and you don't check the box, then you are now
done with partitioning, there is no further screen in the flow - but we
still refer to it as the 'guided partitioning' flow.

As kparal says, the 'autopart' algorithm can actually be invoked from
each path. 'Guided partitioning' always uses the 'autopart' algorithm to
actually create the final partition layout. In 'custom partitioning' you
can choose to use it, by clicking the 'Create partition layout
automatically' text Kamil mentioned. Or you can choose not to use it,
and to actually create each partition manually. I probably wouldn't want
to try and use the term 'manual partitioning' to mean 'custom
partitioning without using the 'autopart' link', that seems
unnecessarily confusing - really, 'custom partitioning' is still 'custom
partitioning' whether you clicked the 'autopart' link or not.

> > >In the original text there is also:
> > >#  Repeat the test, selecting a non-English keyboard map and
> > >entering a passphrase which would not be input the same on an
> > >English keyboard map
> > >Personally I would erase this one, I'm not a fan on
> > >"repeat-several-times" test cases. We can explicitly say the
> > >password _might not_ be set using an English keymap. It's then up
> > >to >people to choose. In several test runs we will receive the same
> > >result (multiple keymaps tested) with less time invested.
> > 
> >  For this one ,I'm a little disagree with you.If we do not
> >  highlight"selecting a non-English keyboard map",we QA may tends to
> >  use English keyboard map only.
> >  for example,I am an Chinese,but I would just do the default English
> >  install for all the testcases,if there is no special saying.
> If something is not explicitly stated in the test case, people should
> be free to do it as they see fit. We might set up some "testing
> introduction" page and describe these principles. But I agree with
> you, in order to encourage using different keymaps, we can explicitly
> say it's allowed. That's why I proposed:
> "We can explicitly say the password _might not_ be set using an
> English keymap."

So the reason this case is as it is right now is that when the issue of
keymap problems first came up, I considered writing separate 'keymap'
test cases, but it seemed a bit artificial: the 'non-US keymap encrypted
partitioning test case' would have been a copy/paste of the encryption
test case, but with 'use a non-US keymap' added. Seemed a bit silly, and
it didn't seem that hard just to add it into the existing test case.
While you're testing encryption, you may as well make sure it works with
a non-US keymap too. The reason for saying to run the test twice, once
with a US keymap and once without, is to make it easier to tell whether
a failure is due to the use of a non-US keymap or not: if you just use a
non-US keymap, and you get a failure, it's hard to know whether it's a
general failure, or because you're using a non-US keymap. If you run the
test twice, it should make it relatively clear what causes any failure.
There are more efficient ways of doing things, but they're harder to
explain in a test case.
Adam Williamson
Fedora QA Community Monkey
IRC: adamw | Twitter: AdamW_Fedora | identi.ca: adamwfedora

More information about the test mailing list