Adam Murdoch wrote:
Hi,
Just a quick update: We've been slowly chipping away at making the CI
builds more stable. Here's the current hit list, in priority order.
1. Initialisation failures in the tooling API tests. This is due to a
couple of problems with locking in the artefact cache, so that
concurrent integration tests are stamping over each other. We'll fix
these for 1.3.
2. Performance test failures. I've narrowed this down to a commit, need
to figure out a fix. We'll fix this for 1.3. There'll almost certainly
be more regressions hiding behind this first failure, and we'll need to
fix these for 1.3. too.
3. Failures on Windows and the IBM JVM due to the client thinking that
the daemon has gone missing. Not sure what's going on here yet. Looking
at the logs it seems the build result is not making it back to the
client for some reason. A fix may or may not go into 1.3. Depends if
it's a test problem or a production problem.
4. JVM crashes in JNA. These will go away as we migrate away from JNA.
5. Our old favourites ProtocolStackTest and DefaultActorFactoryTest.
The first 3 are causing the bulk of the failures at the moment. The
others pop up now and again. There are a few others that occasionally
fail, too, that are not on the list above.
As far as dev builds go, I don't know of any failures that are causing
problems. Please let me know if that's not the case.#
On my mbpro, the Test
org.gradle.process.internal.DefaultWorkerProcessTest#startThrowsExceptionOnConnectTimeout
is always failing for some time now. This seems to be caused by a weired
problem with local vmware & virtualbox network settings. Havn't yet
tracked down any further.
cheers,
René
--
Adam Murdoch
Gradle Co-founder
http://www.gradle.org
VP of Engineering, Gradleware Inc. - Gradle Training, Support, Consulting
http://www.gradleware.com
---------------------------------------------------------------------
To unsubscribe from this list, please visit:
http://xircles.codehaus.org/manage_email
