On Sat, Aug 15, 2009 at 1:29 AM, Brett Porter <[email protected]> wrote:

>
> On 07/08/2009, at 7:00 AM, [email protected] wrote:
>
>
>> +    public void testResolveSnapshotUniqueVersionPresent()
>> +        throws Exception
>> +    {
>> +        VersionedReference ref = new VersionedReference();
>> +        ref.setGroupId( "org.apache.archiva" );
>> +        ref.setArtifactId( "unique-version" );
>> +        ref.setVersion( "1.0-SNAPSHOT" );
>> +
>> +        try
>> +        {
>> +            ArchivaProjectModel model = resolver.resolveProjectModel( ref
>> );
>> +
>> +            assertNotNull( model );
>> +            assertEquals( "org.apache.archiva", model.getGroupId() );
>> +            assertEquals( "unique-version", model.getArtifactId() );
>> +            assertEquals( "1.0-SNAPSHOT", model.getVersion() );
>> +            assertEquals( "Unique Version Snapshot - Build 3",
>> model.getName() );
>> +        }
>> +        catch ( Exception e )
>> +        {
>> +            fail( "The latest timestamp should have been found!" );
>> +        }
>> +    }
>>
>
> Isn't leting the method throw this exception sufficient?
>

Not for all I think. I want the test to fail explicitly and not show in
error when a ProjectModelException is thrown so I added a check for the type
of exception instead. If the exception is not an instance of
ProjectModelException, I just propagated the error. (see commit -r804909)


>
>
>
>> +    public void testResolveNotFound()
>> +        throws Exception
>> +    {
>> +        VersionedReference ref = new VersionedReference();
>> +        ref.setGroupId( "org.apache.archiva" );
>> +        ref.setArtifactId( "non-existant" );
>> +        ref.setVersion( "2.0" );
>> +
>> +        try
>> +        {
>> +            resolver.resolveProjectModel( ref );
>> +            fail( "An exception should have been thrown." );
>> +        }
>> +        catch( Exception e )
>> +        {
>> +            assertTrue( true );
>> +        }
>> +    }
>>
>
> This should catch the specifically expected Exception so it doesn't pass on
> an NPE, for example.


Fixed :)

Thanks,
Deng

Reply via email to