Re: [PATCH 3/5] Improve code of multihook tests

[Kamil Paral]

----- "Will Woods"  wrote:
> On Fri, 2010-09-10 at 15:52 +0200, Kamil Páral wrote:
> > This patch fixes some issues created by adding multihook
>
> capabilities to
> > rpmlint, rpmguard and initscripts tests with patch
> > f16b2646fa397b0cd55e3e4bf9918d21541e8840. It re-enables opt-in
>
> emails.
> > It fixes problem where rpmlint's rpm dir cache wasn't cleared
>
> between
> > successive runs. It also cleans up the code a lot. A lot of recent
> > changes made the code almost unmaintainable and very hard to read
> > (especially in printing/output appending/log appending tasks). This
> > patch reworked all of that quite a lot, it should be much more
>
> readable
> > and simpler now. In short, it tries again to have these scripts
>
> ready to
> > be served as examples for other people.
> >
> > This patch also removes autotest exception throwing and uses
>
> assertions.
>
> This patch is the most complex of the 5, but it looks quite reasonable
> and cleans up the code nicely.
>
> I also really like the use of assertions to catch unexpected/buggy
> conditions - I assume that will show up as a crash/error in autotest?

Yes, the whole test throws AssertionError. I'm not sure what the color
of the field in autotest web UI is :)

As for autoqa-results, we are still missing one patch that would make
the AutoQATest class catch *any* exception and report it as CRASHED.
It's quite an important piece, but it requires changes in all our tests.
I'll try to create it today.