Yes, I agree but there is much work to be done before it becomes a reality
... and it won't be ready for version 1.1. I have been involved in other
stuff during the past month (mainly the creation of Octo Technology UK,
www.octo.com) and have not spent as much time as I would have liked on
Cactus. Basically I am looking for persons interested enough in Cactus to
become committers and help make it evolve and live ... Anyone willing to
help is welcome ! :)

I also like mock objects but for the moment I don't see very clearly how it
would integrate with Cactus as it is. One area of help would be to design
some way for Cactus to be able to provide both mock object testing and
in-container testing. Keeping in mind the simplicity of use and that if end
users have to write 2 suite of tests it would probably be too much a burden.
So it has to be *integrated* ... I have been thinking about a runtime
generation tool, which, depending on some flag would generate the test cases
either for mock objects or for in-container but this is still very much in
the toying around stage ... There may be other ways of doing it.

Thanks
Vincent

----- Original Message -----
From: "Alex Fernández" <[EMAIL PROTECTED]>
To: <[EMAIL PROTECTED]>
Sent: Tuesday, June 12, 2001 10:59 AM
Subject: Re: Roadmap for Cactus 1.1


> Hi Vincent!
>
> I hoped you'd include some kind of mock-object implementation in the
> next release. It would make for more flexible testing.
>
> Un saludo,
>
> Alex.
>
> Vincent Massol wrote:
> >
> > The roadmap for the next release of Cactus (i.e. 1.1) is :
> > - include all corrections/improvements already under CVS, as described
in
> > http://jakarta.apache.org/commons/cactus/changes.html
> > - correct the 2 bugs initally reported by Bob Davison and apply his
patches.
> > The bugs are :
> >   - "Result not found" bug under Tomcat. Sometimes, when the test is a
bit
> > long, under Tomcat, the client side of Cactus would read the result
*before*
> > the test result is actually committed in the application scope
> >   - "Too much data" bug under Tomcat. This an io lock that can happen
under
> > Tomcat if the returned servlet output data are big. In that case, Tomcat
> > would block and wait for the client side to read some data before
> > continuing. As the client is also waiting for the server side to store
the
> > test result in the application scope, there is a deadlock.
> >
> > Any other things that someoe might want included ?
> >
> > Thanks
> > Vincent.
>

Reply via email to