Some Eclipse Foundation services are deprecated, or will be soon. Please ensure you've read this important communication.

Bug 368906

Summary: EUnit Java test regression
Product: z_Archived Reporter: Huang Ji Yong <hjiyong>
Component: EDTAssignee: Project Inbox <edt.javagen-inbox>
Status: CLOSED FIXED QA Contact:
Severity: normal    
Priority: P3 CC: jeffdouglas
Version: unspecified   
Target Milestone: ---   
Hardware: PC   
OS: Windows 7   
Whiteboard:
Attachments:
Description Flags
Compare using beyondCompare none

Description Huang Ji Yong CLA 2012-01-18 00:18:11 EST
Created attachment 209654 [details]
Compare using beyondCompare

Compare to the result of 0.7, the some of the variation in the test cases failed when they are passed in 0.7

api.eglx_lang.EAny001
fvt.datatypes.TestAnyTypeLib
fvt.datatypes.TestCollectionsLib
fvt.json.JSONTestsLib
fvt.operators.TestIsaOperatorLib
fvt.primitives.TestDictionaryLib
fvt.xml.CDATALib
fvt.xml.FlexRecordLib
fvt.xml.NullHandlingLib
fvt.xml.weatherLi
lang.expression.parameter.AnyParameters007

Please refer to the attached file for the comparison of the result of 0.7 and 0.8
Comment 1 Jeff Douglas CLA 2012-01-18 07:43:27 EST
I'll take a look at the differences, however, there was a change made to the eglar's that affect certain nullable conditions (PLUS change, etc), which might make the results that we were getting in 0.7 invalid, and might be correct now. That would mean the testcase needs fixing. We'll see when I get a look at it.
Comment 2 Jeff Douglas CLA 2012-01-19 15:38:48 EST
There was a change in the IR that caused about 90 of these failures. I have since added the 1 line of code that is needed in the EAny class that additionally checks for a comparison of 2 Number types. Previously, this was not needed, as this comparison was sent to the ENumber class for comparison. Which this IR change, only some got sent over.

There are a few differences now left and I am investigating.
Comment 3 Jeff Douglas CLA 2012-01-19 22:13:14 EST
I found a bug in the function invocation logic where it was trying to determine if boxing expressions needed to be altered to wraps. The logic was not handling embedded expressions. This change has been released to cvs. It fixes several more of the problems. There are now only 9 differences left. Will continue working on it.
Comment 4 Jeff Douglas CLA 2012-01-25 10:26:38 EST
There is 1 outstanding issue having to do with Isa on an anyboxedobject of an array. I am opening a new defect for this, and marking this as resolved.
Comment 5 Huang Ji Yong CLA 2012-02-19 22:17:22 EST
Verified in build  0.8.0.v201202190901-1Co-FjuJ6QMNRcrynsZh8h8C