[Laszlo-dev] lzunit improvment [Was: testing]

[P T Withington]

I see the problem.  Some bitrot because lzunit was poking at debugger  
internals.  I'll make a fix.

On 2008-12-02, at 16:18EST, Henry Minsky wrote:

> lztest was printing the name of the test method that failed, but not
> the class (subclass of TestSuite) that it belonged to.
>
>
>
> On Tue, Dec 2, 2008 at 3:58 PM, P T Withington <ptw at pobox.com> wrote:
>> On 2008-12-02, at 15:31EST, André Bargull wrote:
>>
>>>
>>> On 12/2/2008 8:17 PM, P T Withington wrote:
>>>>
>>>> On 2008-12-02, at 11:30EST, André Bargull wrote:
>>>>>
>>>>> I've changed this line in lzunit.lzx (in "Test#fail(..)"):
>>>>>>
>>>>>> this.result.addFailure(message.toString());
>>>>>
>>>>> to:
>>>>>>
>>>>>> this.result.addFailure(this.constructor.classname + "#" +
>>>>>> message.toString());
>>>>>
>>>>> to get the test name.
>>>>
>>>> If you turn on backtrace, failures should be echoed to the debug  
>>>> console
>>>> and should include source file/line (and a backtrace).  Is that  
>>>> not working?
>>>
>>> Don't know, most times I don't work with backtrace turned on...
>>
>> Well, for future reference:  If you are trying to figure out where  
>> a unit
>> test is failing, turn on backtraces -- then errors get file/line  
>> annotation.
>>
>> But, I'm confused by your (and Henry's comments):  Are we talking  
>> about
>> lps/components/lzunit?  Because that already captures the test name  
>> when
>> there is a failure and reports it in the display.  If I'm missing  
>> something,
>> please file an improvement and give me an example of where lzunit  
>> is failing
>> to tell you what test failed.  (I know a lot of people have poked  
>> at lzunit
>> since I first built it and our error strategy has evolved a lot, so  
>> maybe I
>> am just confused about what the problem is.)
>
>
>
> -- 
> Henry Minsky
> Software Architect
> hminsky at laszlosystems.com