-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Shlomi Fish wrote:
> Now we only have to cover the outcomes of the condition in only one
function.
> (or two different outcomes in different functions). So our test
coverage is
> not as comprehensive as the previous one.
Well, I don't think that the test is now less comprehensive. Waving
the hands a bit here, I'd say that the fact that you managed to do the
refactoring tells us something about the code structure (e.g. there
are no lexical variables used inside the "if(CONDITION)" block) which
in turn hints at the coverage previously being redundant (that is, the
code *not* in the "if(CONDITION)" block was being exercised in the
same way by both branch-coverage tests).
I feel it's OK that refactoring, by simplifying the code, would also
simplify the tests, coverage level being otherwise equal.
>
> Right now an ad-hoc solution is to make sure that one has a good test
coverage
> before one refactors the code.
Can't hurt, but the real "philosophical" issue here IMHO is whether
you should delete your tests that don't contribute towards your
coverage anymore. I am of the opinion that it depends what the tests
were created for in the first place. If they cover a bug or express a
requirement, they should be kept; if they were intended to improve
coverage they are now obviously useless. I always document my tests
with a one-liner for this reason (just to keep track of *why* they exist).
Regards,
- --
<< Tout n'y est pas parfait, mais on y honore certainement les
jardiniers >>
Dominique Quatravaux <[EMAIL PROTECTED]>
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.6 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org
iQCVAwUBRaijvvTYH7KfeIIFAQIuYQP+NLb0Ugj+ZWikzDVghrNsM+ec+rIpK2TQ
89D2yQngaqbtmwZjEDyoJjPmwP1MoQ+a7ENFxUgt4XvvIkTv0hetcCNyY5h6dZzD
qiO/+L56Y9loFM2lQO0YdGHbj4bW6694tgHlEtWwQv4mASSvlGo3Ic23w9BSKOOB
bLYp20UYB+g=
=F76W
-----END PGP SIGNATURE-----
