• Robert Clipsham (20/20) Jun 08 2011 I've just read:

Robert Clipsham <robert octarineparrot.com> writes:

I've just read:

http://google-engtools.blogspot.com/2011/06/testing-at-speed-and-
cale-of-google.html 
( http://goo.gl/OKqBk in case that link gets broken due to its length).

Which introduces the interesting notion of partial testing. Perhaps 
something like this could be integrated into unittesting? Something like:

$ dmd -partial-unittest list.d of.d files.d that.d have.d change.d

Could be used to cause only the unit tests of those files to run, and 
their dependencies. This could be further improved with named unittests, 
allowing only relevant blocks of tests in the file to run.

Perhaps an even more interesting way of doing it would be:

$ git diff | dmd -partial-unittest -
- or -
$ git diff > my.diff
$ dmd -partial-unittest my.diff

Where dmd takes a diff, knows exactly what has changed, and runs the 
relevant unittests for the blocks of code that have changed, if they 
fail, their dependencies, and so on.

-- 
Robert
http://octarineparrot.com/