On Thu, 2007-09-27 at 14:54 +0100, Rupert Smith wrote: > Make no mistake. The integration test are very well documented... I > always write my Javadoc ;)
As I am :) I was more referring to making a list of the tests we run with a description of what they effectively test. That way we would make a kind of table and identify what is and is not tested. I was not blaming anybody for not documenting the java doc as I even sometime don't write it (shame on me). Arnaud > On 27/09/2007, Arnaud Simon <[EMAIL PROTECTED]> wrote: > > > > /trunk/qpidtests ? Are these language independent tests > then? Do we > > have /trunk/qpidtests/java /trunk/qpidtests/cpp ? That > would denote > > the language the test was written in... but what about which > broker it > > is ok to test against? > > > > More details before we make such a decision pls. > > > Here is what I am doing > 1) Update the client module tests for getting them running > against the > 0_10 code > 2) Add 0_10 specific tests (that run only against a 0_10 > broker > > Once I am done with that I would suggest we/I do the > following > 1) Document the tests that are in the client module ( yes I am > volunteering) > 2) List the tests that are pure JMS > 3) list the sys tests > 4) Document sys and integration tests > 5) make sure that the sys tests are run with the build > 6) move the client module tests in the identified module > 7) Define a more global testing strategy > > I am not saying this is funny doing all of that but I really > think it > will help. We should stat speaking about our global testing > strategy > knowing exactly what we have. So, I would only answer this > question once > we all know (at least me) better about the big picture. > > Arnaud > > > >
