[
Date Prev][
Date Next][
Thread Prev][
Thread Next][
Date Index][
Thread Index]
[
List Home]
Re: [jakartaee-tck-dev] [External] : Re: [data-dev] New Tools and Challenges of Creating Modern TCK on Persistence Layer
|
On 02. 05. 23 14:10, Lukas Jungmann
wrote:
I may
have misunderstood this comment but wiring tests to particular
editor sounds wrong to me.
Actually I meant exactly opposite - tests not being dependent
even on Maven; and it is also a reason why JEE and all
non-jar-file-servers are now considered as obsoleted (and still
will be for a long time, but at least we started moving). You
cannot simply run and debug tests, you always have to do some
additional steps, something fails, then you have to investigate
what happened and why, destroy the environment, start again.
Instead of just using the same path which doesn't depend on
Maven/Ant/Eclipse/Idea/Netbeans/...
The maven.failsafe.debug doesn't work with "classic" servers (at
least not with GlassFish and Payara).
Most of sources are created in editors. But we all are used to
start everything from the command line - exactly because of this
issue. Not just hitting the play button or "run as" in a context
menu. We still live in 2010 or even in earlier times.
In some of previous emails I sent some trivial examples - they
can be executed in any editor just like any other standard JUnit5
test. For the TCK we just need to avoid mandatory dependency on
GlassFish.
Perhaps the TCK could just create some API jar file based on the
next "arquillian phase", so it would be possible some future
evolution, switching technologies like Arquillian, Test
Containers, or even k8s, bash scripts, etc. However I don't think
there is anyone at this point who would be able to transform the
current ant-based TCK state to such "shiny new perfect project".
--
David Matejcek | OmniFish
david.matejcek@xxxxxxxxxxx