This is the mail archive of the mauve-discuss@sourceware.cygnus.com mailing list for the Mauve project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]

Re: Mauve compatibility intent


>>>>> "Dave" == Dave Roberts <dave@droberts.com> writes:

Dave> Regarding tags in the test code, yes, I think it's fine to
Dave> parameterize things. I think we should avoid using that as an
Dave> excuse to not have a clearly stated goal, however. If everybody
Dave> goes off and writes tests without a clearly stated goal, even if
Dave> parameterized, then no one set of tests gains enough critical
Dave> mass to be useful.

I agree we need a clearly stated goal.

It seems clear that either approach (tracking the "spec" or tracking
Sun's implementation) is losing.  Tracking the spec sucks because the
spec is clearly wrong, or underspecified, or self-contradictory.
Tracking Sun sucks because their implementations have bugs, and change
from release to release.

I think in the long term the best approach for Java would be to have a
real specification which is as unambiguous as possible.  So, I think
basing our tests on the spec is the best approach, because it is in
line with this goal.  That is, tracking Sun's implementation does not
further the goal of getting a real specification for Java, whereas
tracking the spec at least doesn't hurt.

Comments?

Tom

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]