----- Original Message -----
From: <[EMAIL PROTECTED]>
To: "Ant Developers List" <[EMAIL PROTECTED]>
Sent: Thursday, May 30, 2002 6:33 AM
Subject: Re: Ant test framework


>
> Ok, where to from here....do you want someone to come up with some tasks
> and go from there? Or should we discuss the implementation details here
> first?

well, I have only very intermittent time on this right now; but I will do
what I can. help is needed.

lets start with use cases

1. Ant's own build
 -run tests against many build files, many corner cases of tasks themselves
checked, like exception triggering events. Minimal toplevel build file
config, coalesced reporting.

2. third party task
 -same thing as ant, we just want to make it easier to do

3. project with a build file per test.
-the example here is xml-axis. This has build files for each soap endpoint
they are testing against; fetching the WSDL, creating the client stubs,
running them. The aim here is for zero-build-file-rework addition of new
tests, good reporting, and execution of build files that either succeed or
fail.

Someone working on Axis is using a <forEach> variant to set this up, I would
push them towards an <antOn> implementation with failonerror handling
(somehow). I know <anton> was always disparaged because a master build
should know about its sub projects, when you have many test cases, each to
their own build file, this no longer holds.


--
To unsubscribe, e-mail:   <mailto:[EMAIL PROTECTED]>
For additional commands, e-mail: <mailto:[EMAIL PROTECTED]>

Reply via email to