The state of the unit tests

[Esben Mose Hansen]

On Sunday 30 May 2010 13:49:53 Andreas Pakulat wrote:
> On 30.05.10 11:31:45, Esben Mose Hansen wrote:
> > In my humble opinion, failing unit tests and especially completely broken
> > unit tests are worse than no unit test at all, since it discourages
> > further testing.
> 
> FACK. 

I don't think I know that word, and wikipedia doesn't either :o)


> > 1. disabling and/or expect-fail failing tests
> 
> I don't see how that helps, the tests are still failing/broken.

It would help in this way: If I run make test and something fails, I probably 
screwed up something. That is a big help. If a lot of tests are failing before 
I begin, you'll have to dig to see if I broke the test, or if it was already 
broken.

> 
> > 2. some sort of automatic fingering of which checkin breaks previously
> > passing tests.
> 
> That requires writing a tool which updates kdevplatform and kdevelop
> clones to roughly the same revision. Also

wouldn't git pull do that? :) I was more worried about the getting the test 
result up on the web bit. 

> > 3. some effort to make tests that depends on externalities (I'm guessing
> > at least vcs + gdb depends heavily on such) to check for such
> > externalities and replaces the test with an expect fail.
> 
> Gdb depends on a gdb, vcs doesn't depend on anything, svn-tests depend
> on the svn libs. This are all dependencies we also have for the full app
> so are not a problem. The one depending on user-interaction has a fix in
> a pending merge-request.

That is good, then :) And now we are down to 2 failing tests in kdevelop, gdb 
and qtprinters.

-- 
Kind regards, Esben