RFR(s): 8212828: (process) Change the Process launch mechanism default on Linux to be posix_spawn

[Thomas Stüfe]

I think there are two kinds of skipping:
- Skipping where it just does not make sense to execute the test, e.g. on
the wrong OS. That is unexciting.
- Skipping where prerequisites are not met and I skip to reduce the test
analysis load but actually I would like to execute the test. Those tests
one may want to force-execute from time to time. Martins Zombies is a good
example - you may never run it without ever realizing if you have not perl
installed.

But maybe this is all sloppy thinking - one could say the test is
either important enough to be executed always, or it is not. In the former
case all prerequisites should be installed and skipping is not the right
thing to do.

Thank you for the TestNG hint. I will check it out.

Cheers, Thomas


On Tue, Feb 12, 2019 at 3:53 PM Roger Riggs <Roger.Riggs at oracle.com> wrote:

> Hi Thomas,
>
> For tests that are not applicable but should be noted,
> a recent addition is throwing jtreg.SkippedException. [1]
>
> Adding @library /test/lib to the test header
> and import jtreg.SkippedException.
> Throw it when appropriate.
>
> See What if a test does not apply in a given situation?
> <http://openjdk.java.net/jtreg/faq.html#what-if-a-test-does-not-apply-in-a-given-situation>
>
> http://openjdk.java.net/jtreg/faq.html#what-if-a-test-does-not-apply-in-a-given-situation
>
> Testng is generally preferred for new test over main tests and it has its
> own Skipped test mechanism.
>
> Regards, Roger
>
> On 02/12/2019 01:41 AM, Thomas Stüfe wrote:
>
>
> (As a side note, I wonder whether we could have a mechanism to signal
> requirements not met, eg. with a TestRequirementsNotMetException, and then
> let the test executor decide what to do: warn, ignore, error...)
>
>
>
>