| Summary: | Type mismatch: cannot convert from null to boolean/int/long/float/double | ||
|---|---|---|---|
| Product: | z_Archived | Reporter: | pf yu <pfyu> |
| Component: | EDT | Assignee: | Project Inbox <edt.javagen-inbox> |
| Status: | CLOSED WORKSFORME | QA Contact: | |
| Severity: | critical | ||
| Priority: | P3 | CC: | jeffdouglas, jqian, mheitz, pfyu, svihovec |
| Version: | unspecified | ||
| Target Milestone: | --- | ||
| Hardware: | PC | ||
| OS: | Windows XP | ||
| Whiteboard: | |||
|
Description
pf yu
This is a valid defect, but the testcases should be changed to avoid it. The problem happens when null is passed as an argument to a function, and the function's parameter is not nullable. For example, fvt.librarysupport.TestMathLib2LibtestStringAsInt4() contains this line LogResult.assertBigIntEqual1(null, nullableBigint); And the first parameter of assertBigIntEqual1 is "expected bigint in". Notice that it's not nullable. In RBD you can pass null to a non-nullable parameter, and it's like passing in the variable's default value (zero, empty string, etc. depending on the parameter's type). But in EDT if you assign null to a non-nullable variable, or pass null to a non-nullable parameter, we'll throw a NullValueException. In other words, once the JavaGen error is fixed, this code won't work. a different fix must have corrected this. Verified in build 20111020 |