Massive Konqueror Regression

[David van Hoose]

Then I'm sorry to say it, but they need to learn how to write regression
tests. You create the tests from throwing together an example for EVERY
object of concern. It is almost as much work writing the tests initially as
it is writing the code of which you are testing. However, it pays off in the
long run since you know everything works. How do you think Microsoft does
it? Do you honestly think that they blindly release something? No. They test
it with homemade examples. If KDE wants to be the best, the KDE team needs
to start thinking like a commercial entity when it comes to testing. Coding
something, testing it with a few basic examples is not going to make a
project great. You need proper regression tests.

-David van Hoose

-----Original Message-----
From: Kevin Krammer [mailto:kevin.krammer at gmx.at]
Sent: Sunday, August 14, 2005 3:49 PM
To: kde-quality at kde.org
Subject: Re: Massive Konqueror Regression


On Sunday 14 August 2005 21:37, David van Hoose wrote:
> Hi.. Actually all I want to see is some regression tests that the KDE team
> runs on a production box to confirm their code works for a release. Too
> often the release seems like a buggy snapshot that sometimes doesn't even
> compile.

The KHTML developers acutally run regression tests
http://ktown.kde.org/~coolo/regression/

But of course they can't test all possible websites, they have to rely on
contributions of people who are donating some time in testing the alpha and
beta releases and release candidates.

Cheers,
Kevin

--
Kevin Krammer <kevin.krammer at gmx.at>
Qt/KDE Developer, Debian User
Moderator: www.mrunix.de (German), www.qtforum.org