[py-dev] the seeming need of task-based build-tool properties in py.test

[Ronny Pfannschmidt]

On Mon, 2010-02-15 at 14:50 +0100, Adam wrote:
> Ronny Pfannschmidt wrote:
> 
> > as needs like dependend tests, [...]
> > it seems like we are going tinto the direction of a task-based build
> > tool where each test item is a 'task' that may depend on other tasks and
> > has possible checked preconditions for being executed.
> 
> A test dependency is IMO very different to a build dependency. With a
> build dependency the task can't be run until it's dependency have
> produced their product. If they are run in a wrong order, it produces a
> broken build.

ordering issues indeed get important as soon as one creates test-items
that depend on the order and execution of other test items

this is already happening for glashammers acceptance-testsuite
another part of the issue is tests that take long,
but are not required to run every time - like for example code
validators

-- Ronny

> 
> OTOH the test dependencies are only relevant for interpreting why a test
> failed. If (at least) one dependency failed, I don't care if the test
> passt or not. In which order the two were run is irrelevant. Sure, if I
> ran the dependency first, I can save time, but as (with the exception of
> looponfailing) '#failed test << #tests' it shouldn't make much of a
> difference).
> 
> --Adam