Re: RFC: A Test For Every Patch

Bradley M Kuhn <[email protected]> writes:

> Discussion forked into "should perl define Perl" and discussion of
> formats.  I would like to comment (on this Request for Comment :) that I
> agree with Michael strategy.  I think patchers should submit a test case
> along with any patches that are submitted, within reason of course.  ;)

I agree for all code changes (at least that are testable; I can think of a
few types of code changes that aren't particularly testable and probably
aren't worth testing, but I'm sure that people will understand that rules
are just rules and there are exceptions).  I'm unconvinced for
documentation.  It's hard enough to get people to write documentation; I
think it's important to decrease the number of hoops they have to jump
through as much as possible.

-- 
Russ Allbery ([email protected])             <http://www.eyrie.org/~eagle/>