[Laszlo-dev] For Review: Change 20080827-ptw-k Summary:

[André Bargull]

'Debug.write("")' does not print a new prompt (regression for dhtml, swf 
was also broken before the change).

On 8/31/2008 9:52 PM, André Bargull wrote:
> And "canvas.sprite" prints a "SyntaxError" (also dhtml-only):
>> «Error «object| SyntaxError:…» computing __typeof#44| {owner: This is 
>> th [...]
>> <javascript:window.parent.$modules.lz.Debug.displayObj(44)>
>
>
> On 8/31/2008 9:46 PM, André Bargull wrote:
>> Functions are displayed differently in dhtml, for instance type 
>> "Class" into the debugger and inspect the object:
>>
>> new:
>>> «Object#8| Class» {
>>> _dbg_name: 'Class'
>>> addProperty: «Class static function(2)» 
>>> <javascript:window.parent.$modules.lz.Debug.displayObj(5)>
>>> addStaticProperty: addStaticProperty($1_name, $2_value) {
>>> this[$1_name] = $2_value;
>>> if <javascript:window.parent.$modules.lz.Debug.displayObj(6)>
>>> allClasses: {…} 
>>> <javascript:window.parent.$modules.lz.Debug.displayObj(15)>
>>> make: make($1_classname, $2_traitsAndSuperclass, 
>>> $3_instanceProperties, $4_staticProperties) {
>>> var nc = function constructor() {this.constructor = 
>>> arguments.callee;this.validateClassStructure();if 
>>> <javascript:window.parent.$modules.lz.Debug.displayObj(16)>
>>> prototype: {…} 
>>> <javascript:window.parent.$modules.lz.Debug.displayObj(18)>
>>> }
>>
>> old:
>>> «Object#87| Class» {
>>> _dbg_name: 'Class'
>>> addProperty: addProperty 
>>> <javascript:window.parent.$modules.lz.Debug.displayObj(83)>
>>> addStaticProperty: addStaticProperty 
>>> <javascript:window.parent.$modules.lz.Debug.displayObj(84)>
>>> allClasses: {...} 
>>> <javascript:window.parent.$modules.lz.Debug.displayObj(95)>
>>> make: make <javascript:window.parent.$modules.lz.Debug.displayObj(96)>
>>> prototype: {...} 
>>> <javascript:window.parent.$modules.lz.Debug.displayObj(97)>
>>> }
>>
>>
>>
>> On 8/30/2008 8:11 PM, André Bargull wrote:
>>> Great stuff, that will make live easier for everyone who needs to 
>>> touch the debugger.
>>> Right now I'm going through the files, until now just (mini-)source 
>>> changes:
>>>
>>> LzRuntime.lzs, encodeURIComponent: why not decodeURIComponent?,
>>> LzBootstrapDebugConsole#addText(..): string coercion looks odd, 
>>> belongs into the catch-clause?
>>> LzBootstrapDebugLogger#log(..): loadVariables() should be 
>>> loadVariablesNum(), right?
>>> LzBootstrapDebugService: _dbg_log_all_writes, why a global?
>>> LzBootstrapDebugService#xmlEscape(): we really ought to consolidate 
>>> all our different xml-escape routines...
>>> LzBootstrapDebugService, "var Debug": you've left out the 
>>> "var"-declaration for the generic and runtime-specific debuggers, 
>>> why not here? (as the first "var Debug" is declared in LzRuntime.lzs)
>>> LzBootstrapDebugService, $reportNotFunction and 
>>> $reportUndefinedMethod: message concatenation looks odd
>>> LzCompiler, $reportSourceWarning: "if (LzWarning)" is (now) always 
>>> true, so the else-clause can be removed
>>> LzBootstrapMessage#indexOf(..args) vs LzDebugMessage#indexOf(key): 
>>> not a compatible override
>>> LzBootstrapMessage#toString(), #toHTML(): timestamp is off by three 
>>> years ;-)
>>> LzBootstrapMessage: String.prototype.toHTML() uses 
>>> "Debug.xmlEscape", which is only available in debug-mode
>>> LzSourceMessage: missing declaration of "message" in class
>>> LzSourceMessage: constructor typing wrong, "message" is either 
>>> String or LzMessage
>>> LzDebugService.internalPropertyPrefixes: use an Object instead of an 
>>> Array (it's used as an Object in LzDebugService#internalProperty())
>>> debugger-component: update the comment at file-beginning and mark 
>>> the methods which implement the LzBootstrapDebugConsole-interface
>>> LzDHTMLDebugConsole declares "doEval" as overriden, but there is no 
>>> "doEval" in LzBootstrapDebugConsole, missing?
>>>
>>> Does it make sense to replace these lines (occur several times in 
>>> the debugger-files):
>>>> var color = '#0000ff';
>>>> if (attrs && attrs.color) { color = attrs.color };
>>> with a shorter replacement (for LPP-6892): var color = (attrs && 
>>> attrs.color) || '#0000ff';
>>>
>>> As you see, that's merely me being nit-picking. And I haven't yet 
>>> tried the new debugger, I'll do that tomorrow.
>>>
>>> On 8/28/2008 10:14 PM, P T Withington wrote:
>>>> Change 20080827-ptw-k by ptw at dueling-banjos.home on 2008-08-27 
>>>> 14:46:24 EDT
>>>> in /Users/ptw/OpenLaszlo/krispy-kreme
>>>> for http://svn.openlaszlo.org/openlaszlo/trunk
>>>>
>>>> Summary: Classify the debugger
>>>>
>>>> Bugs Fixed:
>>>> LPP-6226 'make the compiler debug flag work with swf9' (step 0)
>>>>
>>>> Technical Reviewer: hminsky (pending)
>>>> QA Reviewer: a.bargull at intensis.de (pending)
>>>>
>>>> Summary:
>>>> This is a pretty massive change, not easy to review from the
>>>> diffs. I propose checking it in so we can just get some usage on
>>>> it.
>>>>
>>>> I re-factored the debugger so that it is written as a service
>>>> class with a singleton instance. There is a bootstrap class
>>>> (which works nearly immediately), a generic sub-class (which
>>>> contains the debugger proper) and per-platform sub-sub-class
>>>> (which adds platform-dependent overrides and extensions). I also
>>>> broke out the interface portion of the debugger so that different
>>>> interfaces can be mixed and matched with the same debugger.
>>>> Similarly, the logging portion of the debugger is broken out;
>>>> although for now there is just one logger which understands how to
>>>> log back to the server for swf or to the browser console (e.g.,
>>>> firebug) for dhtml.
>>>>
>>>> When the system starts up, the bootstrap debugger is created and a
>>>> bootstrap console attached. The bootstrap debugger supports the
>>>> various debug message levels and the compiler warnings and errors
>>>> in a very simplified fashion that does not depend on much else
>>>> working. The bootstrap console simply records any output (which
>>>> will be replayed once a real console is attached).
>>>>
>>>> As soon as the necessary substrate is available, the platform-specific
>>>> debugger is instantiated to replace the bootstrap debugger, the
>>>> appropriate console is attached (based on runtime and query
>>>> options), and the bootstrap console's output is replayed to the
>>>> actual console.
>>>>
>>>> Details:
>>>> Now that try/catch is working in swf8, I was able to re-merge a
>>>> lot of the platform-specific parts of the debugger that were only
>>>> broken out because of that missing functionality.
>>>>
>>>> LzInit is now generic.
>>>>
>>>> LzMessage (and its subclasses) is now classified (and a sub-class
>>>> of the bootstrap version). The one bummer about this is that it
>>>> is no longer a subclass of String, so I had to create a new
>>>> predicate isStringLike, that replaces (typeof == 'string' || is
>>>> String).
>>>>
>>>> __typeof, __String and inspectInternal now are generic instead of
>>>> being platform-specific. Platform sub-classes can override to add
>>>> platform-specific types, e.g., the DHTML debugger knows how to
>>>> print DOM nodes.
>>>>
>>>> The debugger now uses LzFormatter as a mixin, rather than
>>>> trampolining out to it.
>>>>
>>>> Remodularized the debugger component and remote console to obey
>>>> the new console protocol. The component debugger no longer plays
>>>> any surgury games to insert itself into the prototype chain of the
>>>> debugger, now that there is a clean(er) console protocol.
>>>>
>>>> ResponderEVAL: Fixed the logger to translate debug message levels
>>>> in to the corrsponding log4j levels, so logged warnings and errors
>>>> will show up (now that the default log level is warn).
>>>>
>>>> Updated console debugger to use displayObj rather than eval, so
>>>> inspecting looks pretty.
>>>>
>>>> Tests:
>>>> smokecheck, ant lztest, verified that console debugger still works
>>>>
>>>> Files:
>>>> M WEB-INF/lps/lfc/debugger/swf9stubs.lzs
>>>> A + WEB-INF/lps/lfc/debugger/LzInit.lzs
>>>> M WEB-INF/lps/lfc/debugger/LzMessage.lzs
>>>> M WEB-INF/lps/lfc/debugger/LzDebug.lzs
>>>> M WEB-INF/lps/lfc/debugger/platform/swf/LzFlashRemote.as
>>>> D WEB-INF/lps/lfc/debugger/platform/swf/kernel.as
>>>> D WEB-INF/lps/lfc/debugger/platform/swf/LzInit.as
>>>> M WEB-INF/lps/lfc/debugger/platform/swf/LzDebug.as
>>>> D WEB-INF/lps/lfc/debugger/platform/dhtml/LzInit.js
>>>> M WEB-INF/lps/lfc/debugger/platform/dhtml/LzDebug.js
>>>> D WEB-INF/lps/lfc/debugger/platform/dhtml/kernel.js
>>>> M WEB-INF/lps/lfc/debugger/LzMonitor.lzs
>>>> D WEB-INF/lps/lfc/debugger/LzFormat.lzs
>>>> M WEB-INF/lps/lfc/debugger/LzMemory.lzs
>>>> M WEB-INF/lps/lfc/debugger/Library.lzs
>>>> M WEB-INF/lps/lfc/debugger/LzTrace.lzs
>>>> M WEB-INF/lps/lfc/debugger/LzBacktrace.lzs
>>>> M WEB-INF/lps/lfc/LaszloLibrary.lzs
>>>> M WEB-INF/lps/lfc/events/LaszloEvents.lzs
>>>> M WEB-INF/lps/lfc/compiler/LzFormatter.lzs
>>>> M WEB-INF/lps/lfc/compiler/LzRuntime.lzs
>>>> M WEB-INF/lps/lfc/compiler/LzMessage.lzs
>>>> A WEB-INF/lps/lfc/compiler/LzBootstrapDebugService.lzs
>>>> M WEB-INF/lps/lfc/compiler/Library.lzs
>>>> M 
>>>> WEB-INF/lps/server/src/org/openlaszlo/servlets/responders/ResponderEVAL.java 
>>>>
>>>> M lps/components/debugger/debugger.lzx
>>>> M lps/admin/dev-console.lzx.swf
>>>> M lps/admin/dev-console.lzx
>>>> M lps/admin/dev-console.lzx.js
>>>>
>>>> Changeset: 
>>>> http://svn.openlaszlo.org/openlaszlo/patches/20080827-ptw-k.tar
>>>>
>>>
>>
>