After everything is set up for the test, I want to be very sensitive
to errors and warnings, and, you know, turn the lava lamps red and
start blasting Guns-n-Roses "Welcome to the Jungle", etc. This
particular problem occurs while loading the LFC, before I've got
everything set up, so I don't want to trigger the lava lamps.
(There aren't currently any actual lava lamps, btw.)
On Apr 24, 2006, at 3:10 PM, P T Withington wrote:
> Ok, I am doing this, but I think what you are doing is actually the
> way you want to run the tests, because what I am doing in debug
> mode is catching errors and displaying them. For your tests, you
> want the errors not to be caught, so you can fail the test, right?
>
> On 2006-04-24, at 16:31 EDT, Benjamin Shine wrote:
>
>> Smarter addText sounds good.
>>
>> I checked in a workaround, which is just to run the lztests with
>> the non-debug LFC. I make a trivial Debug object, make all the
>> (warn|info|write|warn) methods call addText, and make addText just
>> call print. So, "ant lztest" works again.
>>
>>
>>
>> On Apr 24, 2006, at 1:20 PM, P T Withington wrote:
>>
>>> What if I just make addText smarter and only use the debug window
>>> if it exists, use print if it exists, etc. Are there other
>>> choices/fallbacks?
>>>
>>> On 2006-04-24, at 16:05 EDT, Benjamin Shine wrote:
>>>
>>>> It looks to me like Debug.addText is being called while we're
>>>> still loading the LFC, before we get to my override of it. Is
>>>> there a switch I can set to turn off reportSourceWarning? I
>>>> think it's trying to report a warning when it loads the lfc, by
>>>> calling Debug.addText, which kills it.
>>>>
>>>>
>>>> On Apr 24, 2006, at 5:17 AM, P T Withington wrote:
>>>>
>>>>> Yes, but that is inside Debug.addText, which I thought you
>>>>> override in your Rhino runtime before you do anything? Oh
>>>>> wait... I have a debug call to it early on, let me try
>>>>> removing that.
>>>>>
>>>>> On 2006-04-24, at 00:42 EDT, Benjamin Shine wrote:
>>>>>
>>>>>>
>>>>>> The variable names appended to the register names help a lot.
>>>>>>
>>>>>> This makes sense why it doesn't work from rhino -- in
>>>>>> LzRuntime.lzs it says
>>>>>> 68 if ($dhtml) {
>>>>>> 69 var dw = top.frames['LaszloDebugger'];
>>>>>> 70 var dwd = dw.document;
>>>>>>
>>>>>> So, in rhino, we're in dhtml, but there's no top.frames. I can
>>>>>> get around this with the BUFakeDom, which seems like the right
>>>>>> thing to do. I want the rhino runtime to seem just like the in-
>>>>>> browser dhtml environment.
>>>>>>
>>>>>> The build will fail tonight because of this -- blame me!
>>>>>> Tomorrow I'll be working on this, and on documenting the test
>>>>>> system.
>>>>>>
>>>>>> -ben
>>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>>> On Apr 23, 2006, at 4:21 PM, P T Withington wrote:
>>>>>>
>>>>>>> Change 41722 by [EMAIL PROTECTED] on 2006/04/23 16:20:56
>>>>>>>
>>>>>>> Summary: Tweaks for Ben
>>>>>>>
>>>>>>> Technical Reviewer: ben (pending)
>>>>>>> QA Reviewer: ben (pending)
>>>>>>> Doc Reviewer: n/a (pending)
>>>>>>>
>>>>>>> Details:
>>>>>>>
>>>>>>> LzRuntime, LzDebug: Make bootstrap warning use Debug.addText
>>>>>>> so it can
>>>>>>> be captured by unit tests.
>>>>>>>
>>>>>>> JavascriptGenerator: Append variable names to register names
>>>>>>> in debug mode for increased debugability.
>>>>>>>
>>>>>>> Tests: Examine LFCdhtml-debug.js
>>>>>>>
>>>>>>> Affected files ...
>>>>>>>
>>>>>>> ... //depot/lps-legals/WEB-INF/lps/lfc/compiler/
>>>>>>> LzRuntime.lzs#3 edit
>>>>>>> ... //depot/lps-legals/WEB-INF/lps/lfc/debugger/LzDebug.lzs#3
>>>>>>> edit
>>>>>>> ... //depot/lps-legals/WEB-INF/lps/server/src/org/openlaszlo/
>>>>>>> sc/JavascriptGenerator.java#16 edit
>>>>>>>
>>>>>>
>>>>>> benjamin shine
>>>>>> software engineer
>>>>>> [EMAIL PROTECTED]
>>>>>>
>>>>>>
>>>>>>
>>>>>
>>>>
>>>> benjamin shine
>>>> software engineer
>>>> [EMAIL PROTECTED]
>>>>
>>>>
>>>>
>>>
>>
>> benjamin shine
>> software engineer
>> [EMAIL PROTECTED]
>>
>>
>>
>
benjamin shine
software engineer
[EMAIL PROTECTED]
_______________________________________________
Laszlo-dev mailing list
[email protected]
http://www.openlaszlo.org/mailman/listinfo/laszlo-dev
