[jira] Created: (RIVER-39) visibility semantics in transaction spec incomplete

["Jim Hurley (JIRA)" <ji...@apache.org>]

visibility semantics in transaction spec incomplete
---------------------------------------------------

                 Key: RIVER-39
                 URL: https://issues.apache.org/jira/browse/RIVER-39
             Project: River
          Issue Type: Improvement
          Components: net_jini_core
    Affects Versions: jtsk_1.1
            Reporter: Jim Hurley
            Priority: Minor


Bugtraq ID [4512970|http://bugs.sun.com/bugdatabase/view_bug.do?bug_id=4512970]

TX.3.5 doesn't make clear what happens to visibility state when a top-level transaction commits. It also doesn't make clear that there can be conflicts between locks and visibility state (deletion in particular). Parallels to the statements made about the effect of conflicts on operations that attempt to acquire locks need to made about the effect of conflicts on operations that attempt to delete.

The terminology is also somewhat confusing. There is a distinction between the created/deleted state of an object, which is inherited as subtransactions commit, and the visibility of an object to an arbitrary transaction, which is not inherited. Using the phrase "visibility state" and "creation/deletion visibility" for the former confuses it with the latter.

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.