The failure_handler is a jtreg extension that we build and put in the
test image. It's then used when running our product tests. As a stand
alone product, it has its own set of tests. The old makefile had a
target for running those so I felt I needed to keep that functionality.
I agree the make target can be confusing, but I couldn't find a better one.
/Erik
On 2016-04-13 12:48, Magnus Ihse Bursie wrote:
Is the test-failure-handle used to test that the failure handler is correct? It
seems a bit odd, since most other tests has their main purpose to be run, and
the build part is just a necessary thing, but here I presume that the main
thing is to build the handler so other tests can use it, and it is not clear
what the test itself mean. Otoh, I can't see any other way to express this.
/Magnus
12 apr. 2016 kl. 16:00 skrev Erik Joelsson <erik.joels...@oracle.com>:
Please review this change which adds a proper makefile and build sequence for
the failure-handler jtreg plugin. It also adds the failure handler to the test
image and makes the hotspot and jdk test/Makefile's pick it up when available.
Bug: https://bugs.openjdk.java.net/browse/JDK-8149777
Webrev: http://cr.openjdk.java.net/~erikj/8149777/webrev.03/
/Erik
