Re: anyone familiar with sb-cover?

["Kyle R. Burton" <kyle.burton@gmail.com>]

>  Perhaps the other macro is is residence of the utility, otherwise it should not have come out uncover.  If you are getting 100% from the primary code you should be alright.  The pointer will do no good.

Yeah, I've got 4 packages, A and B (A's tests), C and D (C's tests).

I've got macros in A that are showing 100% coverage based on tests is
in B - though I think I'm using them directly in B. I've got a macro
in C that is used by 2 fn's in C which are covered by the tests in D,
but the macro itself isn't showing coverage.  It must be b/c its not
used directly by the unit test...actually now that I've said that...

Macros are expanded at 'compile' time - when the fasl is written.  So
now I'm thinking maybe it's not covered by virtue of how I'm executing
the runner:

  1 - turn on coverage
  2 - require / recompile the base libraries (A and C)
  3 - use 'load to run/exec the unit test code (B and D)
  4 - produce coverage report

It's probably at 2 where the macro from C is actually 'executed' and
not at 3.  So maybe it's that distinction...it looks like I should
create a simpler example and have a few things to work out for myself.
 Exporting the macro isn't the end of the world, it's just feels dirty
- exposing too many of the innards of the module itself.  I wonder if
just a macroexpand-1 in the unit test is enough, lisp-unit does in
fact have an assert-expansion...


Thanks,

Kyle