[
https://issues.apache.org/jira/browse/DAFFODIL-1298?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16868713#comment-16868713
]
Michael Beckerle commented on DAFFODIL-1298:
--------------------------------------------
Here's what we need to do to resolve this:
The tests that have the same name should be scrutinized in the TDML file.
If identical, the 2nd should be deleted.
If not-identical, the 2nd should be renamed, and the scala driver updated to
the new name. Then if they pass, all good. However, the tests might not be
identical, but still might be redundantly checking something that it is obvious
some other test is checking. In that case this 2nd test should also be deleted.
But when in doubt, keep it.
If this new "2nd" test fails, however, then the test must be scrutinized to see
if it is a proper/correct test, and if it is determined that the failure
actually represents a bug a Jira ticket needs to be created, the scala driver
call commented out listing the Jira ticket.
> Duplicate test cases in Functions.tdml
> --------------------------------------
>
> Key: DAFFODIL-1298
> URL: https://issues.apache.org/jira/browse/DAFFODIL-1298
> Project: Daffodil
> Issue Type: Bug
> Components: Clean Ups, QA
> Affects Versions: 1.0.0
> Reporter: Michael Beckerle
> Priority: Minor
> Labels: beg
>
> I tweeked the TDML runner to reject duplicate test names in the same TDML
> file, and I found that the file Functions.tdml has quite a few duplicates in
> it. These tests are passing because the first one found is used. The later
> test with duplicate name is not used.
> However, really we should not be depending on this. All tests in a TDML file
> should have unique names. The schema for TDML files is supposed to have a
> uniqueness constraint that assures this, but it seems either that uniqueness
> constraint is not properly expressed, or we're invoking the validation of
> TDML files without the options that tell it to validate the key/uniqueness
> constraints (which can be disabled because they are so expensive to execute
> by the Xerces XML validator). I believe the TDML runner should turn on "full
> schema checking" or something like that to get the uniqueness constraints
> checked. (TDML runner defect is DFDL-1299)
> Here's the list of duplicates from Functions.tdml
> More than one test case for name 'hexBinary_constructor_03'.
> More than one test case for name 'nonNeg_constructor_03'.
> More than one test case for name 'byte_constructor_03'.
> More than one test case for name 'uint_constructor_01'.
> More than one test case for name 'hexBinary_constructor_04'.
> More than one test case for name 'uint_constructor_04'.
> More than one test case for name 'byte_constructor_04'.
> More than one test case for name 'ubyte_constructor_01'.
> More than one test case for name 'ubyte_constructor_02'.
> More than one test case for name 'ubyte_constructor_03'.
> More than one test case for name 'hexBinary_constructor_01'.
> More than one test case for name 'nonNeg_constructor_01'.
> More than one test case for name 'byte_constructor_01'.
> More than one test case for name 'uint_constructor_03'.
> More than one test case for name 'date_constructor_01'.
> More than one test case for name 'hexBinary_constructor_02'.
> More than one test case for name 'uint_constructor_02'.
> More than one test case for name 'nonNeg_constructor_04'.
> More than one test case for name 'byte_constructor_02'.
> More than one test case for name 'ubyte_constructor_04'.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
