I think it makes sense to error out on unknown platforms.
This doesn't conflict with skipping tests (or silently passing) for
unsupported platforms.
We just need a way to distinguish them. I'd probably go for the same
approach that Max suggested, i.e. add the list of (known) unsupported
platforms to osMap and errors out if the requested platform cannot be
found in osMap.
Valerie
On 9/27/2016 7:22 AM, Wang Weijun wrote:
Looking at the webrev, it looks we've never tested on "Linux-arm-32" and
"Linux-aarch64-64" before and we only realized it now. This is a true problem.
On the other hand, I also agree with Xuelei's concern. If a new platform is
added and it does not have NSS libs tests will fail.
How about this? If there is such a new platform called "mhos-arch-32", we can
add
osMap("mhos-arch-32", new String[]{});
and, make nssLibDirs == null a failure, but nssLibDirs.length == 0 can return
null.
Is this good?
Thanks
Max
On Sep 27, 2016, at 7:25 PM, Xuelei Fan <xuelei....@oracle.com> wrote:
I think, it is the expected behavior to ignore the test if a platform does not
support it. If showing failures, every testing on unsupported platform will
fail, and additional effort MUST be paid to evaluate the root cause of the
failure. We should try to avoid that.
Xuelei
On 9/27/2016 6:32 PM, Tim Du wrote:
Hi All:
Would you help to review the patch for sun/security/pkcs11/PKCS11Test.java?
The test keep pass on not supported platforms, it will make nobody
notice the test was skipped,which is not our expected. Update case to
show failure, when platform not supported. And add the support for
Linux-arm-32 and Linux-arm-64 platforms. Thanks.
JBS: https://bugs.openjdk.java.net/browse/JDK-8164322
Webrev: http://cr.openjdk.java.net/~tidu/8164322/webrev.00/
Regards
Tim
