test infrastructure ideas appreciated ...

Michael Meeks michael.meeks at collabora.com
Wed Jun 10 07:14:35 PDT 2015


On Wed, 2015-06-10 at 09:35 -0400, Kohei Yoshida wrote:
> I am happy to see that you and I are on the same term on the majority of the
> points. But let me nitpick on a few points below.

	:-)

> > 	My yard-stick would be that if writing the unit-test takes longer than
> > finding & fixing the bug - then we have a problem, 
> 
> While I think I agree with the general point you are trying to make here, I
> would have to slightly disagree with how you chose to word it.

	=)

> By this wording, if a bug fix takes a mere 5 minute, or even 30 minute

	So - of course, its well worth spending that time.

> I personally would still spend a few hours writing a test for a bug I fixed in
> 10 to 30 minutes, because, even if the fix took less than an hour today, the
> same fix at some arbitrary point in the future may take days or weeks just
> because the code may look totally different by then.

	Yep; makes sense; so if people have the dedication to do that - that's
great. I like to try to persist at writing tests to the bitter end - it
often gives a far better understanding of the real fix. Then again -
there are some areas where it is just too expensive currently. eg.
layout - though I hope (with some investment) we can fix that so they
become almost easy =) My hope is that if we approach this from several
sides: improved automation, improved ease of testing, more & better test
infrastructure, and also your idea of highlighting unit-testing heros
(who perhaps fix fewer bugs per unit time, but they stay fixed) - then
we can make a real difference.

> Anyway, I just wanted to make these points clear. I hope you didn't
> mind my nitpicking.

	I love your precision =) ( and passion for testing ), we need more of
that.

	ATB,

		Michael.

-- 
 michael.meeks at collabora.com  <><, Pseudo Engineer, itinerant idiot



More information about the LibreOffice mailing list