TODO

***************************************************************************
*									  *
*   Copyright (c) 2009, Lawrence Livermore National Security, LLC. 	  *
*   Produced at the Lawrence Livermore National Laboratory 		  *
*   Written by: Gavin M. Bell (gavin@llnl.gov) 				  *
*   LLNL-CODE-420962.	BSD Open Source License				  *
*									  *
*   All rights reserved. This file is part of the:                        *
*   Earth System Grid (ESG) Data Node Software Stack                      *
*									  *
***************************************************************************

   Description: 
   A running list of items of work that need to be done

***************************************************************************

TODO:

Features:

- Shell: 
  - Mode exit function to clear mode env
  - caching DAO in admin mode

- P2P Protocol adjustments
  - don't drop messages just because you have seen them before
  - Do not resend messages with artificially touched timestamp
  - TTL logic for dropping messages.
  - Test timing code for removed messages vs stale incoming
  - filter by namespaces (regex namespaces during merge)
  - remove immediately if fails to connect
      if then recovers with ping, then reinstate
  - Send out explicit 'goodbye' message when leaving / shutting down
  - When comparing node timestamps in merge, add in skew slop function
      instead of direct comparing. Ex: isWithinSkew(ts1,ts2);
      coordinate this skew tolerance with that of ntp.
  - Adjust the regex for the PeerNetworkFilter to deal with namespace
      prefix. Ex: a candidate network foo-org.esgf is accepted when
      testing for org.esgf. (but not vise versa)


overall:
-Put in MORE JUnit testing.
-Write test cases for codebase.
-Install BDM client and simple sanity check call(s)

core:
Metrics & Monitoring
 ->Put in Query "LIMIT" clause at the end of each query to a configurable value
 ->For Monitoring put in time series results in fixed size LRU cache

-Remove faux 'gateway' from codebase.
-Put in XML Parsing of Payload message data (the catalog)
 -> Have to get or make catalog specific XSD to JAXB on.

-Find out from Bob what the deal is/was with hessian to see if I can
safely upgrade to using 3.1.5 - he alluded to hacking the hessian code
on the python side for things to work... will an upgraded hessian
screw things up?

      hessian-3.0.20.jar ->? hessian-3.1.5.jar

DONE:
xClean up InstallCert Blurting out exception stack trace simply because
 you don't have the key.  (that's the point - you're trying to get it :-)

libs: (todo's and issues)

	git rm jsr173 api use fetched stax api
	git rm jaxb1-impl use fetched jaxb-impl

xRoll back log4j from 1.2.15 to 1.2.14 to avoid ivy repo fetch failure
issue.

	git rm log4j-1.2.15.jar use fetched log4j-1.2.14.jar


xFix issue with ivy having problems with resolving "javamail" to fetch
 mailapi.jar and smtp.jar

-Keep only:
       ivy-2.1.0.jar
       [hessian-3.0.20.jar] ? Test that still wire compatible with python version.
 x     [javamail = mail.jar, smtp.jar] ? working on resoving ivy repo failure
       (everything else we fetch via ivy)

(jars that will be fetched by ivy that can be removed)

    activation.jar
    commons-dbcp-1.2.2.jar
    commons-dbutils-1.3.jar
    commons-logging-1.1.1.jar
    commons-pool-1.5.4.jar
    jaxb-api.jar
    jaxb-impl.jar
    jaxb-xjc.jar
    jaxb1-impl.jar
    jsr173_1.0_api.jar
    junit-4.8.2.jar
    log4j-1.2.15.jar
    postgresql-8.3-603.jdbc3.jar
    servlet-api.jar

xFind out how to put fetched jars in a separate lib subdirectory
 "fetched" so we can seperate the two types of libs and then have git
 be able to ignore the fetched ones and remove that directory on
 cleanup with ant.


---