Improve scope of validation warning errors

[jerseyrobot]

If a resource class has 2 GET's with same media types but different parameter
lists, the deployment is OK and the generated WADL would have both method listed.

I think if the usage is wrong deployment should fail in the first place, also
any validations should catch it during design time. Specs/documentation should
also make it clear, if not already.

Environment

Operating System: All
Platform: All

[jerseyrobot]

@glassfishrobot Commented
Reported by ntn@java.net

[jerseyrobot]

@glassfishrobot Commented
Was assigned to jersey-issues

[jerseyrobot]

@glassfishrobot Commented
ntn@java.net said:
correct issue summary

[jerseyrobot]

@glassfishrobot Commented
sandoz@java.net said:
Yes, this is a validation issue, that should result in a deployment error or a
warning.

There are a whole bunch of these we need to enumerate in the spec and check for
in the implementation.

[jerseyrobot]

@glassfishrobot Commented
sandoz@java.net said:
More generally the same validation error/warning should be produced for Java
methods with different method names and signatures but the same consuming media
(and URI paths, if present).

[jerseyrobot]

@glassfishrobot Commented
japod@java.net said:
I think, we need to add some more validation tests to
BasicValidator#visitAbstractResource method
in the first place.

Later on, if performance tests will prove it is needed,
we can update the validation schema (it uses visitor pattern),
so that it will keep a context information when validating
individual methods, so that we could save one iteration loop.

[jerseyrobot]

@glassfishrobot Commented
sandoz@java.net said:
Another case two sub-resource resource methods or locators:

class R {
@path("x") @get String getX()

{ ... }
@path("x") @get String getX() { ... }

}

In addition the following would need to be checked as well:

class R {
@path("x") @get String getX()

{ ... }
@path("x/") @get String getX() { ... }

}

as matching of terminated and non-terminated '/' paths as above is equivalent.

[jerseyrobot]

@glassfishrobot Commented
japod@java.net said:
All mentioned cases are now being checked during resource model validation.

[jerseyrobot]

@glassfishrobot Commented
Marked as fixed on Sunday, September 7th 2008, 6:22:47 pm

[jerseyrobot]

@glassfishrobot Commented
This issue was imported from java.net JIRA JERSEY-33

[jerseyrobot]

• Issue Imported From: https://github.com/jersey/jersey/issues/305
• Original Issue Raised By:@glassfishrobot
• Original Issue Assigned To: @mpotociar
• Original Issue Closed By:@glassfishrobot