code coverage mode?

[m h <sesquile@gmail.com>]

Hey all-

I'm implementing emacs support for code coverage tools in python.

I've currently got some code [0], that will show coverage data in two ways:

* flymake mode
* compile mode

I like the flymake style, yet I still want to really use flymake for
how it was intended.  (Plus I don't want to be re-running coverage
analysis on every change, since it only really changes when you
instrument the code.  Also if you edit the code then your coverage
data is probably stale (unless you're clever about it, which I'm not)
So I guess in that way it's more like compile mode).  I'm not a lisper
by any means, I'm hacking stuff together here.  But if this were
python then I'd create a subclass of flymake and use that for
coverage, while still using the former for its normal use (syntax
checking).

My recent work has been a simple test finder.  The idea being that you
are working on a function and want to see what kind of coverage you
have for that function.  So you invoke the run-function-tests (or
something similar) and it will run only tests that invoke that
function directly and report coverage on that.  View the results,
tweak the tests to get more coverage.  Rinse, repeat....

So I figure before I go off and hack a kludgey solution, I thought I'd
ask the experts if they have any ideas or suggestions.  It'd be cooler
to implement it the "right way" if others (other languages) could take
advantage of it and use it as well.

thoughts?  Does anyone here care about coverage?  What features would they want?

(if this is OT kindly redirect me to the correct place :))

thanks much!

-matt

0 - http://github.com/mattharrison/pycoverage.el