You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@wicket.apache.org by Gerolf Seitz <ge...@gmail.com> on 2007/12/03 00:24:47 UTC
serious change for ajax related unit tests
hi all,
there has been an issue, that ajax related unit tests didn't test with an
ajax request,
but rather with a normal request[0].
as a consequence, the generated markup inside the ajax response was actually
the markup for normal requests
(eg. included wicket tags and wicket attributes, which is not the case for
"real" ajax responses).
the effect of this change is, that your ajax related unit tests will
probably fail with the current trunk.
i expect that most failed tests can be fixed by correcting the expected ajax
result.
it's rather unfortunate that this happens so close to the 1.3.0 final
release, but then again it's fortunate
that we caught it before the release at all ;)
as usual, if there are any questions left unanswered, don't hesitate to
ask/complain/investigate/...
Regards,
Gerolf
[0] https://issues.apache.org/jira/browse/WICKET-1199
Re: serious change for ajax related unit tests
Posted by Gerolf Seitz <ge...@gmail.com>.
i ran a coverage analysis with EclEmma [0] (recommended by Al Maw for the
occasional check [1]),
and posted the results to my p.a.o site [2].
i don't know about the reliability of the results, but if this is anywhere
near correct, everybody knows in what areas wicket
lacks unit tests -> patches welcome ;)
Cheers,
Gerolf
[0] http://www.eclemma.org
[1] http://herebebeasties.com/2007-06-20/eclemma-code-coverage-made-easy/
[2] http://people.apache.org/~gseitz/wicket-coverage/
On Dec 3, 2007 2:11 AM, Ryan Sonnek <ry...@gmail.com> wrote:
> Just curious if anyone has run clover or cobertura maven reports
> against the codebase to see what kind of unit test coverage their
> currently is. I have *no* doubt that wicket's unit test coverage is
> *extremely* high compared to other web frameworks, but it might be
> good to see if there are any important gaps (like this) that need
> coverage.
>
> From a community perspective, I know I would be comfortable
> contributing unit tests to help out.
>
> On Dec 2, 2007 6:24 PM, Gerolf Seitz <ge...@gmail.com> wrote:
> > hi all,
> >
> > there has been an issue, that ajax related unit tests didn't test with
> an
> > ajax request,
> > but rather with a normal request[0].
> > as a consequence, the generated markup inside the ajax response was
> actually
> > the markup for normal requests
> > (eg. included wicket tags and wicket attributes, which is not the case
> for
> > "real" ajax responses).
> >
> > the effect of this change is, that your ajax related unit tests will
> > probably fail with the current trunk.
> > i expect that most failed tests can be fixed by correcting the expected
> ajax
> > result.
> >
> > it's rather unfortunate that this happens so close to the 1.3.0 final
> > release, but then again it's fortunate
> > that we caught it before the release at all ;)
> >
> > as usual, if there are any questions left unanswered, don't hesitate to
> > ask/complain/investigate/...
> >
> > Regards,
> > Gerolf
> >
> > [0] https://issues.apache.org/jira/browse/WICKET-1199
> >
>
Re: serious change for ajax related unit tests
Posted by Gerolf Seitz <ge...@gmail.com>.
i ran a coverage analysis with EclEmma [0] (recommended by Al Maw for the
occasional check [1]),
and posted the results to my p.a.o site [2].
i don't know about the reliability of the results, but if this is anywhere
near correct, everybody knows in what areas wicket
lacks unit tests -> patches welcome ;)
Cheers,
Gerolf
[0] http://www.eclemma.org
[1] http://herebebeasties.com/2007-06-20/eclemma-code-coverage-made-easy/
[2] http://people.apache.org/~gseitz/wicket-coverage/
On Dec 3, 2007 2:11 AM, Ryan Sonnek <ry...@gmail.com> wrote:
> Just curious if anyone has run clover or cobertura maven reports
> against the codebase to see what kind of unit test coverage their
> currently is. I have *no* doubt that wicket's unit test coverage is
> *extremely* high compared to other web frameworks, but it might be
> good to see if there are any important gaps (like this) that need
> coverage.
>
> From a community perspective, I know I would be comfortable
> contributing unit tests to help out.
>
> On Dec 2, 2007 6:24 PM, Gerolf Seitz <ge...@gmail.com> wrote:
> > hi all,
> >
> > there has been an issue, that ajax related unit tests didn't test with
> an
> > ajax request,
> > but rather with a normal request[0].
> > as a consequence, the generated markup inside the ajax response was
> actually
> > the markup for normal requests
> > (eg. included wicket tags and wicket attributes, which is not the case
> for
> > "real" ajax responses).
> >
> > the effect of this change is, that your ajax related unit tests will
> > probably fail with the current trunk.
> > i expect that most failed tests can be fixed by correcting the expected
> ajax
> > result.
> >
> > it's rather unfortunate that this happens so close to the 1.3.0 final
> > release, but then again it's fortunate
> > that we caught it before the release at all ;)
> >
> > as usual, if there are any questions left unanswered, don't hesitate to
> > ask/complain/investigate/...
> >
> > Regards,
> > Gerolf
> >
> > [0] https://issues.apache.org/jira/browse/WICKET-1199
> >
>
Re: serious change for ajax related unit tests
Posted by Ryan Sonnek <ry...@gmail.com>.
Just curious if anyone has run clover or cobertura maven reports
against the codebase to see what kind of unit test coverage their
currently is. I have *no* doubt that wicket's unit test coverage is
*extremely* high compared to other web frameworks, but it might be
good to see if there are any important gaps (like this) that need
coverage.
>From a community perspective, I know I would be comfortable
contributing unit tests to help out.
On Dec 2, 2007 6:24 PM, Gerolf Seitz <ge...@gmail.com> wrote:
> hi all,
>
> there has been an issue, that ajax related unit tests didn't test with an
> ajax request,
> but rather with a normal request[0].
> as a consequence, the generated markup inside the ajax response was actually
> the markup for normal requests
> (eg. included wicket tags and wicket attributes, which is not the case for
> "real" ajax responses).
>
> the effect of this change is, that your ajax related unit tests will
> probably fail with the current trunk.
> i expect that most failed tests can be fixed by correcting the expected ajax
> result.
>
> it's rather unfortunate that this happens so close to the 1.3.0 final
> release, but then again it's fortunate
> that we caught it before the release at all ;)
>
> as usual, if there are any questions left unanswered, don't hesitate to
> ask/complain/investigate/...
>
> Regards,
> Gerolf
>
> [0] https://issues.apache.org/jira/browse/WICKET-1199
>