"To avoid such compatability issues in future, is it a good idea to
come up with a grammar/spec. for V3?"
Precisely. I am working on that spec (Kedar and a few others have
seen it), but it won't be for Prelude.
Lloyd
On Jul 18, 2008, at 10:09 AM, Sreenivas Munnangi wrote:
> Lloyd Chambers wrote:
>>
>> Comments on comments—
>>
>> With the structure of domain.xml changing substantially in some
>> area, combined with the lack of a grammar and aliasing, backward
>> compatibilty for dotted names should be done only as a
>> compatibility layer, and even then will have its challenges.
>>
> To avoid such compatability issues in future, is it a good idea to
> come up with a grammar/spec. for V3 ?
>> Lloyd
>>
>> On Jul 17, 2008, at 11:37 PM, Nazrul Islam wrote:
>>
>>> Prashanth Abbagani wrote:
>>>>
>>>> Initial draft of Monitoring one pager for V3 Prelude is
>>>> available. Please take a took and provide feedback, comments,
>>>> concerns, if any.
>>>>
>>>> http://wiki.glassfish.java.net/Wiki.jsp?page=GFV3MonitoringOnePager
>>> Thanks for starting this thread. Here are some notes from today's
>>> review...
>>> 1) Section 4.1.2.3.3: Please ensure that the object view
>>> hierarchy is aligned with GFv2 and the dotted names are backward
>>> compatible.
>>> 2) Section 4.1.2.5: Please ensure that monitor_type is correctly
>>> mapped for GFv2 types also.
>>> 3) Section 4.1.2.5: asadmin set command does not work without
>>> domain.xml persistence.
>>> 4) Section 4.12: Please specify how GFv2 monitoring levels (OFF,
>>> LOW, HIGH) will work.
>>> 5) Please ensure that this works for both JDK 5 and JDK6.
>>> 6) Please expose the monitoring information with JMX MBeans. REST
>>> support may come post Prelude.
>>> 7) Please check all the APIs for naming consistency and align.
>>> 8) Secion 4.1.1.1: Please specify that appName can be null.
>>> 9) Section 4.3: Please specify what is in scope for GFv3 Prelude.
>>> 10) Section 4.4: Please clearly specify what is out of scope so
>>> that future design goals are understood.
>>> 11) Section 4.5.1.2: Do we need to expose the probe annotations
>>> (@ProbeName, @ProbeParams, @MethodEntry) and ProbeProviderInfo?
>>> 12) Section 4.5.1.2: Is there any risk with exposing remove method
>>> on TreeNode to 3rd parties? Should we have exposed only the
>>> integration point (example parentId)? See example at: http://wiki.glassfish.java.net/Wiki.jsp?page=GFV3PluggabilityOnePager#section-GFV3PluggabilityOnePager-4.1.1AdminConsole
>>> 13) Section 4.1.2.5: [Question] Example shows "gf:tx:*". Should
>>> the first tupple be "gf" in the example? How does this relate to
>>> DTrace?
>>> 14) Please clarify support for JVM, JDBC, JPA, jRubby monitoring
>>> and CallFlow support.
>>>
>>>
>>> --
>>> Nazrul Islam - (408) 276-6468 - Sun Microsystems, Inc.
>>>
>>
>