Below list is a collection of relatively recent [as of early june, 2015] test failures on iojs jenkins slaves.
These tests fails because the test runs out of time (defaults to 60, higher on arm):
- test-cluster-worker-wait-server-close (osx1010, centos5-32, centos5-64, freebsd101-64)
- test-net-server-max-connections (smartos32, smartos64)
test-child-process-double-pipe (win2008r2, win2012r2)test-child-process-exit-code (win2008r2, win2012r2)test-child-process-spawnsync-input (win2008r2, win2012r2)
- test-vm-new-script-new-context (most slaves, example): next branch, related to v8 upgrade
- undefined-reference-in-new-context (most slaves - example): same as above
- test-debug-signal-cluster.js (smartos32): possible output buffering issue
- test-sync-io-option.js (smartos32, smartos64, armv7-ubuntu1404) - looks limited to branch/commit, but later found at pi1-rasbian-wheezy
- test-timers-unrefd-interval-still-fires (pi1-raspbian-wheezy): issue 1781
- test-cluster-disconnect (armv7-wheezy)
- test-process-env (armv7-wheezy)
- test-child-process-stdio (armv7-wheezy)
- test-fs-non-number-arguments-throw (win2012r2)
- test-child-process-emfile (freebsd101-32)
test-child-process-spawnsync (win2008r2, win2012r2): most likely a timeout issue since it seems coupled with other tests that behaves similarly
These tests are run on each vm outside of jenkins invoked by make test or .\vcbuild.bat release nosign x64 test. It calls tools/test.py with the -J flag which checks the number of cores on the host and run that amount of tests in parallel. Each vm runs the suite twice. Tests haven't been run on arm-* because of access.
- test-child-process-spawnsync (smartos32, smartos64)
- test-cluster-master-kill (smartos32, smartos64)
- test-cluster-master-error (smartos32)
- test-child-process-fork-net2 (smartos64)
- test-debug-signal-cluster (smartos64)
- test-fs-access (ubuntu1404-64)
test-child-process-stdout-flush-exit (centos7-64, freebsd*)(fixed here)
These are on the
nextbranch but I don't know if they are new or not