| --- | --- |
| Bugzilla Link | 302675 |
| Status | NEW |
| Importance | P3 normal |
| Reported | Feb 12, 2010 01:33 EDT |
| Modified | Nov 08, 2011 04:56 EDT |
| Version | 3.0.0 |
| Blocks | 318368 |
| Reporter | Ed Willink |
Description
Specific problem:
Deep messages such as those now generated within the OCLValidationDelegates cannot propagate an exception through EMF and so the ParserException is converted to a plugin log, which is never tested and which under standalone testing is not even visible in the log for visual verification.
As a minimum DEBUG tracing should be enabled during all tests.
A mechanism, such as a LogRecorder, should be provided to diagnose that the accumulation of such messages during a test is as expected.
General problem:
There is significant cut and paste duplication between OCL{,Ecore,UML}Plugin and three copies of OCLDebugOptions. Surely ecore and uml should share the core debug tracing selection, not require duplicate settings/lost messages?
| --- | --- | | Bugzilla Link | 302675 | | Status | NEW | | Importance | P3 normal | | Reported | Feb 12, 2010 01:33 EDT | | Modified | Nov 08, 2011 04:56 EDT | | Version | 3.0.0 | | Blocks | 318368 | | Reporter | Ed Willink |
Description
Specific problem:
Deep messages such as those now generated within the OCLValidationDelegates cannot propagate an exception through EMF and so the ParserException is converted to a plugin log, which is never tested and which under standalone testing is not even visible in the log for visual verification.
As a minimum DEBUG tracing should be enabled during all tests.
A mechanism, such as a LogRecorder, should be provided to diagnose that the accumulation of such messages during a test is as expected.
General problem:
There is significant cut and paste duplication between OCL{,Ecore,UML}Plugin and three copies of OCLDebugOptions. Surely ecore and uml should share the core debug tracing selection, not require duplicate settings/lost messages?