Permalink
Commits on Jun 12, 2012
Commits on Jun 11, 2012
  1. Remove noisy debug

    lance committed Jun 11, 2012
  2. The cache service name can't be an alias, so make it conditional base…

    …d on isClustered()
    lance committed Jun 11, 2012
  3. [TORQUE-806] Add org.torquebox.cache.as.CacheService.

    Add a CacheService that contains an infinispan CacheContainer and some
    metadata - specifically, a boolean flag that will tell us if we're
    running the AS in clustered mode or not.  Change the cache container
    name to "polyglot" in anticipation of moving this out to jboss-polyglot.
    lance committed Jun 11, 2012
  4. Checkpoint commit

    lance committed Jun 11, 2012
Commits on Jun 8, 2012
  1. Add clustered integration specs for cache.

    This integration tests only works because I've made a change to
    Cache#cache to _not_ reconfigure the cache at runtime. If I reconfigure,
    we can't determine that we're running in clustered mode, so the cache
    thinks it's in LOCAL mode and reconfigures itself as such.
    lance committed Jun 8, 2012
Commits on Jun 7, 2012
  1. [TORQUE-806] Refactor TorqueBox::Infinispan::Cache to work in container.

    These changes ensure a few things.
    1) All ruby runtimes are using the same cache, so that data can be
    shared between them regardless of the persistence behavior.
    
    2) All Infinispan cache containers are non-clustered when run in
    standalone mode, and clustered when run from standalone-ha.
    
    3) When not run in a torquebox context, the cache shouldn't barf.
    
    There are still some failing specs. In gems/cache, attempting to test
    clustered mode when not in-container is proving problematic. In the
    integration tests, cache rollback behaves badly in failed distributed
    transactions.
    lance committed Jun 7, 2012
  2. [TORQUE-806] Refactor TorqueBox::Infinispan::Cache to work in container.

    These changes ensure a few things.
    1) All ruby runtimes are using the same cache, so that data can be
    shared between them regardless of the persistence behavior.
    
    2) All Infinispan cache containers are non-clustered when run in
    standalone mode, and clustered when run from standalone-ha.
    
    3) When not run in a torquebox context, the cache shouldn't barf.
    
    There are still some failing specs. In gems/cache, attempting to test
    clustered mode when not in-container is proving problematic. In the
    integration tests, cache rollback behaves badly in failed distributed
    transactions.
    lance committed Jun 7, 2012
  3. Support windoze betterer

    jcrossley3 committed Jun 7, 2012
Commits on Jun 6, 2012
  1. typo

    bbrowning committed Jun 6, 2012
Commits on Jun 3, 2012
Commits on Jun 1, 2012
  1. Hard to be humble

    jcrossley3 committed Jun 1, 2012
  2. [TORQUE-806] Add integration tests for shared caches in standalone mode.

    Currently, shared caches do not work unless they are persisted with the
    :persist => [true|path] option. In-memory cache sharing is pending.
    lance committed Jun 1, 2012
Commits on May 31, 2012
  1. Pass jmx4r a URL instead of process regex in integs (TORQUE-842)

    To connect via URL we're using the remoting-jmx connector in AS7,
    which is why all the extra test dependencies are needed in
    integration-tests/pom.xml.
    bbrowning committed May 31, 2012
  2. Merge pull request #98 from jkutner/server_spec

    cleared the Gem paths before loading torquebox-server to make sure it's ...
    jkutner committed May 31, 2012
  3. Merge pull request #97 from jkutner/2x-dev

    explicitly rescue NameError when loading logger to fix [TORQUE-527]
    jkutner committed May 31, 2012
  4. Support -Xa.b style JRUBY_OPTS properties (TORQUE-633)

    These will be converted to -Djruby.a.b when TorqueBox is started via
    torquebox run or rake torquebox:run.
    bbrowning committed May 31, 2012
Commits on May 30, 2012