Agree to not block teh release for the failing tests (which IMO are false negatives):
The failures seem to be related to the order in which object properties are serialized to JSON. A similar thing happened to the Abiquo provider, where tests succeeded or failed depending on the JAXB version (which kinda sucks). If you take a closer look at the failure, it is just a matter of the test, which is comparing the payloads as raw Strings, but if you take into account the JSON structure and values, the generated requests are correct. I think we could improve ExpectTests assertions in general to avoid this kind of raw String comparisons, and make the tests happy regardless of the Java/JAXB/foo versions. On 30 May 2014 09:36, Andrew Gaul <g...@apache.org> wrote: > On Tue, May 27, 2014 at 08:40:18PM +0000, Everett Toews wrote: >> This thread is for discussion of the first release candidate for Apache >> jclouds 1.7.3. Please use this thread for discussion of issues uncovered in >> the RC, questions you may have about the RC, etc. Thank you. > > I ran verify_jclouds_release.sh against jdk1.7.0_55 and all tests passed > but against jdk1.8.0_05 some cloudsigma tests failed. I attached the > results from this run. Given that this code has not changed since 1.7.2 > this should not block the release. > > I also ran integration tests against aws-s3. I encountered two types of > failures, several due to clock skew for which I submitted a pull > request[1] and several due to DNS settling issues described in > JCLOUDS-305. We may want to port the generic S3 changes to AWS-S3. > > [1] https://github.com/jclouds/jclouds/pull/386 > > -- > Andrew Gaul > http://gaul.org/
