Em Sexta 21 Abril 2006 12:21, Sylvain Hellegouarch escreveu:
> Well, I assume most unit tests that TG has are not testing the behavior of
> TG as a whoile but layer by layer right? If so, it may be a good idea to
> setup scripts which checks if TH as a framework behaves as expected... a
> mega-unit-test I suppose ;)
A "united-test" instead of "unit-test" ;-) hehehehe... Sorry for the bad
joke... ;-)
> Forgive ne though if this is already the case as I don't keep a very close
> eyes on TG (I stick to CP only for now ;)
I believe unit tests should test individual features but their behavior
shouldn't differ when integrated together otherwise the test is invalid in a
real situation.
So, we have a requirement for ">" that is there because of features or fixes
we need that are only available in those versions. I believe that
restricting the upper version is much more complicated. It should be a
problem if we follow "alpha" or "unreleased" versions -- and it happens --
but ASAP we should add a requirement for the stable version.
To solve this I only see one solution:
- documentation AND
- releasing a new TG version when we see something breaking and adding
a "<"
version requirement to avoid using the broken version.
Docs should tell that to users and make it clear that the latest release is
recommended, is stable and will contain such information for
incompatibilities of new versions of its dependencies.
This and unit testing (I don't believe that the results should differ with the
mega-unit-test and if they do then this is a bug in our tests or our code)
should be enough.
--
Jorge Godoy <[EMAIL PROTECTED]>
--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups
"TurboGears" group.
To post to this group, send email to [email protected]
To unsubscribe from this group, send email to [EMAIL PROTECTED]
For more options, visit this group at http://groups.google.com/group/turbogears
-~----------~----~----~----~------~----~------~--~---
