develooper Front page | perl.perl5.porters | Postings from September 2010

Re: [perl #76138] perl inadvertently destroys signal handlers asof f746176000

Thread Previous | Thread Next
Paul Johnson
September 6, 2010 15:09
Re: [perl #76138] perl inadvertently destroys signal handlers asof f746176000
Message ID:
On Mon, Sep 06, 2010 at 08:26:18AM -0700, Florian Ragwitz via RT wrote:
> On Sun Sep 05 12:48:50 2010, wrote:

> > OK, I’ve changed it. That was copied and pasted from elsewhere. But
> >    doesn’t it make it less likely that we’ll catch buggy interactions
> >    between features if we try to keep the tests too standardised?
> Yes, possibly. I've asked that question on irc, and the answer I got
> from our pumpkin was "Don::t", which I relayed to you. I'm not entirely
> what the reasoning behind it was.

Irrespective of this particular instance, the good Father does bring up
an interesting philosophical question.  Do we prefer to have our tests
standardised, homogenised and unified, which eases the testing process
in many important ways, or do we prefer a hodge-podge of differing
styles, a mélange of features, which could improve the test coverage
and, as mentioned, might catch more buggy interactions?

Historically we've leant towards the former, although the passage of
time and individual preference has also pushed us somewhat towards the
latter.  I also tend to lean towards the standardised approach, feeling
that I really should be able to understand the interactions, and program
the ones I want, rather than relying on luck to find them for me, but I
also think this says rather more about my (lack of) skill as an
exploratory tester then I would really like it to.

Paul Johnson -

Thread Previous | Thread Next Perl Programming lists via nntp and http.
Comments to Ask Bjørn Hansen at | Group listing | About