Cosmos Architecture Meeting 13-Dec-07
- Valentina Popescu
- Jack Devine
- Martin Simmonds
- John Todd
- Paul Stratton
- Hubert Leung
- David Whiteman
- Bill Muldoon
- Ali Mehregani
- Sheldon Lee-Loy
- Don Ebright
- Joel Hawkins
Notes on Error Discussion
OK... We've opened a bunch of ERs with "Error" type stuff in them. I've tried to compile the complete list below.
209227: Common Error Management System for COSMOS https://bugs.eclipse.org/bugs/show_bug.cgi?id=209227 see also: COSMOS Design 209227
209234: Error management for the contract between the UI and DataCollection https://bugs.eclipse.org/bugs/show_bug.cgi?id=209234 see also: COSMOS Design 209234
209242: Error Management for MDR interface to SML repository https://bugs.eclipse.org/bugs/show_bug.cgi?id=209242
208592: Error Handling - improve programming model https://bugs.eclipse.org/bugs/show_bug.cgi?id=208592
209223: Error Handling - client side https://bugs.eclipse.org/bugs/show_bug.cgi?id=209223
Proposed Course of Action
First, I'd like to check in the code that is attached to . This introduces the common COSMOS exception. We can, over the course of i8 & i9, refactor to have all of our exceptions extend this one.
Third, I propose that  be scoped to the following:
- The error contract between the MDR and the outside world, from a Java perspective, will be extensions of COSMOSExceptions
- The CMDBf specification v1.0 defines faults. These should be mapped into a set of COSMOS exceptions
Once this mapping is defined, this ER can be closed.
Sheldon and team have prepared a design document on . I propose that we review this and close the ER prior to the architecture call on 20-Dec-07 via e-mail on COSMOS-dev.
I'd propose that we review and close the forthcoming design of [209223 prior to the architecture call on 20-Dec-07 via e-mail on COSMOS-dev.
Errors vs. Logging....
There was some discussion/confusion that COSMOS Design 209227 was or was not going to cover our logging strategy. In prep for this discussion, I went back over some of my notes and trolled around on the wiki to try and assemble where I think we are. The net is, we wanted to keep the logging strategy related, but separate from, the error strategy. We established a page to capture our logging requirements COSMOS internal logging use cases. The intent was to use these use cases to open up the necessary ERs.
Looking at this page, we clearly have some work that needs to be done. Given that we are getting to the point where we need to start producing logs, we may want to ensure we get a good cut of this work done in i9.