Wed, 22 Aug 2007
A Testing WTF
I'm in the process of writing up a "when and how to test" screed, and I discovered this:
Karl Fogel's book, Producing Open Source Software, has precisely two keyword hits to testing in the ToC.
WTF?
--titus
posted at: 12:11 | path: /aug-07 | 4 comments
Comments:
Posted by Ricardo Niederberger Cabral at Wed Aug 22 14:00:20 2007:
I counted three. But really, is there any real technical difference between testing OSS projects versus commercial ones (which the author probably assumed as something already extensively covered elsewhere)?
Posted by Rams at Wed Aug 22 21:17:39 2007:
All non-trivial SVN patches require a test to be submitted; the whole test suite is in python. The "HACKING" doc has the details.
But you are right though, the book doesn't emphasize testing - many hackers like the author do not have testing (as a topic worthy of attention by itself) on their radar. I am far more concerned about "Committers" are gods attitude in the book. It doesn't talk about how to detect and eliminate cliques and cabals. Why does SVN have a private committers only mailing list ? What are the exact criteria for for becoming a committer - Why is it so difficult to set it down in black and white ?
Posted by Titus Brown at Thu Aug 23 07:53:06 2007:
Rams -- what project are you talking about?
Ricardo -- I think there are differences in testing OSS projects, just like there are differences in communication structure, responsibilities, code writing, etc... Testing is actually a huge force multiplier when working on a collaborative project, especially one where most people are, in theory, equal contributors.
--titus
Posted by Rams at Fri Aug 24 12:55:54 2007:
Titus Brown,
I meant the SVN project itself, since that project is used throughout the book as the primary example.
Here is the HACKING doc:http://subversion.tigris.org/hacking.html