P T Withington | 2 Nov 02:09 2009
Picon

Re: [Laszlo-reviews] For Review: Change 20091031-ptw-3 Summary: Work around Adobe bug ASC-3852

Thanks for the test cases.  Am looking into it.

Looks like labelled statements are just plain broken!

On 2009-10-31, at 18:11, André Bargull wrote:

> Not approved, these methods didn't compile:
>
> <method name="returnVoid" returns="void">
> for (var s in {}) {
>   if (1) return;
> }
> </method>
>
> <method name="returnFunction" returns="Function">
> for (var s in {}) {
>   if (1) return (function () {});
> }
> return null;
> </method>
>
> And labels don't work, but I think that's a different issue:
>
> <!-- only works in non-debug mode -->
> <method name="labeledFor" >
> foo: for (var s in {}) {
>   if (1) return 1; else break foo;
> }
> </method>
>
(Continue reading)

Raju Bitter | 2 Nov 12:50 2009

Splash screen / 1 >= 0

Hi all,

I tried to compile an SWF8 splash screen into an OL 4.0.18 app, and  
got the following warning: Can't import /branding/shared/splash/ 
splash_progress.swf 1 >= 0
While 1 might be >= 0, what is the system trying to tell me here?
Thanks,
Raju

Raju Bitter | 2 Nov 12:54 2009

[BULK] Re: Splash screen / 1 >= 0

Looking into some source I'm wondering: do we still need this method?

     private String[] getBlogList() {
         return new String[] {
             "http://wetmachine.com/",
             "http://osteele.com/",
             "http://www.davidtemkin.com/",
             "http://www.ultrasaurus.com/",
             "http://pt.withy.org/ptalk/"
         };
     }

Found in http://svn.openlaszlo.org/openlaszlo/trunk/WEB-INF/lps/server/src/org/openlaszlo/compiler/SplashCompiler.java

;-)

On Nov 2, 2009, at 12:50 PM, Raju Bitter wrote:

> Hi all,
>
> I tried to compile an SWF8 splash screen into an OL 4.0.18 app, and  
> got the following warning: Can't import /branding/shared/splash/ 
> splash_progress.swf 1 >= 0
> While 1 might be >= 0, what is the system trying to tell me here?
> Thanks,
> Raju
>

P T Withington | 2 Nov 13:37 2009
Picon

Do you know how in the swf9 compiler...

It diverts the top-level statements from processing and inserts them  
into the main class?  I can't seem to follow the logic.

P T Withington | 2 Nov 14:14 2009
Picon

Re: [BULK] Re: Splash screen / 1 >= 0

We're keeping that for backwards compatibility with 2.0.  :)

As to your question about 1 >= 0, I don't think I can help.  I grepped  
for `rror.*>=` and found nothing.  As a wild guess, this is a multi- 
frame resource and it is missing frame 0?

On 2009-11-02, at 06:54, Raju Bitter wrote:

> Looking into some source I'm wondering: do we still need this method?
>
>    private String[] getBlogList() {
>        return new String[] {
>            "http://wetmachine.com/",
>            "http://osteele.com/",
>            "http://www.davidtemkin.com/",
>            "http://www.ultrasaurus.com/",
>            "http://pt.withy.org/ptalk/"
>        };
>    }
>
> Found in http://svn.openlaszlo.org/openlaszlo/trunk/WEB-INF/lps/server/src/org/openlaszlo/compiler/SplashCompiler.java
>
> ;-)
>
> On Nov 2, 2009, at 12:50 PM, Raju Bitter wrote:
>
>> Hi all,
>>
>> I tried to compile an SWF8 splash screen into an OL 4.0.18 app, and  
>> got the following warning: Can't import /branding/shared/splash/ 
(Continue reading)

Raju Bitter | 2 Nov 14:27 2009

Re: [BULK] Re: Splash screen / 1 >= 0

Thanks, Tucker. Here's some info I got from both SWF (the one which  
works and the new one) through SWF Decompile:

Default splash_progress.swf for Webtop:
   SWF Version 7
   Width 312px
   Height 312px
   Frame Rate 12
   FrameNumber 100
There's no frame 0 in that SWF file as it looks.

The new SWF file:
   SWF Version 8 (uses some Flash 8 effects)
   Width 804px
   Height 343px
   Frame Rate 12
   Frame Number 100

The new SWF was built based upon an example FLA file Antun gave me.  
Maybe SWF8 is not supported for preloaders?

- Raju

On Nov 2, 2009, at 2:14 PM, P T Withington wrote:

> We're keeping that for backwards compatibility with 2.0.  :)
>
> As to your question about 1 >= 0, I don't think I can help.  I  
> grepped for `rror.*>=` and found nothing.  As a wild guess, this is  
> a multi-frame resource and it is missing frame 0?
(Continue reading)

Henry Minsky | 2 Nov 16:34 2009

Re: Do you know how in the swf9 compiler...

There was some hack for that, I remember setting up these constants to pass to the
script compiler

        props.put(org.openlaszlo.sc.Compiler.SWF9_APPLICATION_PREAMBLE, makeApplicationPreamble());
        props.put(org.openlaszlo.sc.Compiler.SWF9_APP_CLASSNAME, MAIN_APP_CLASSNAME);
        props.put(org.openlaszlo.sc.Compiler.SWF9_WRAPPER_CLASSNAME, EXEC_APP_CLASSNAME);

There's some code in the script compiler that sets and refers to config.mainClassName, which somehow the parsetreeprinter decides to put things into, but I'm not sure how.

  public List makeTranslationUnits(SimpleNode translatedNode, boolean compress, boolean obfuscate)
  {
  ...
...

    config.mainClassName = (String) options.get(Compiler.SWF9_APP_CLASSNAME);
    //System.err.println("mainClassName = "+config.mainClassName);





On Mon, Nov 2, 2009 at 7:37 AM, P T Withington <ptw <at> pobox.com> wrote:
It diverts the top-level statements from processing and inserts them into the main class?  I can't seem to follow the logic.



--
Henry Minsky
Software Architect
hminsky <at> laszlosystems.com


P T Withington | 2 Nov 16:46 2009
Picon

Re: Do you know how in the swf9 compiler...

Yeah, I am finding that.  Digging into it.

André discovered that labeled statements don't work, and that has led  
me down a rat hole about how the unparser unparses statements and  
blocks.  The parser is inserting all these extra levels of statements  
that hold statements, and I am trying to simplify that, but our tree- 
walkers are pretty fragile...

<bleah/>

On 2009-11-02, at 10:34, Henry Minsky wrote:

> There was some hack for that, I remember setting up these constants  
> to pass
> to the
> script compiler
>
>        props.put(org.openlaszlo.sc.Compiler.SWF9_APPLICATION_PREAMBLE,
> makeApplicationPreamble());
>        props.put(org.openlaszlo.sc.Compiler.SWF9_APP_CLASSNAME,
> MAIN_APP_CLASSNAME);
>        props.put(org.openlaszlo.sc.Compiler.SWF9_WRAPPER_CLASSNAME,
> EXEC_APP_CLASSNAME);
>
> There's some code in the script compiler that sets and refers to
> config.mainClassName, which somehow the parsetreeprinter decides to  
> put
> things into, but I'm not sure how.
>
>  public List makeTranslationUnits(SimpleNode translatedNode, boolean
> compress, boolean obfuscate)
>  {
>  ...
> ...
>
>    config.mainClassName = (String)
> options.get(Compiler.SWF9_APP_CLASSNAME);
>    //System.err.println("mainClassName = "+config.mainClassName);
>
>
>
>
>
> On Mon, Nov 2, 2009 at 7:37 AM, P T Withington <ptw <at> pobox.com> wrote:
>
>> It diverts the top-level statements from processing and inserts  
>> them into
>> the main class?  I can't seem to follow the logic.
>>
>
>
>
> -- 
> Henry Minsky
> Software Architect
> hminsky <at> laszlosystems.com

Henry Minsky | 3 Nov 01:54 2009
Picon

[Laszlo-reviews] For Review: Change 20091102-hqm-F Summary: fix for computing maxscroll on DHTML text view

Change 20091102-hqm-F by hqm <at> badtzmaru.home on 2009-11-02 19:50:14 EST
     in /Users/hqm/openlaszlo/trunk-clean
     for http://svn.openlaszlo.org/openlaszlo/trunk

Summary: fix for computing maxscroll on DHTML text view

New Features:

Bugs Fixed: LPP-8583

Technical Reviewer: max
QA Reviewer: ptw
Doc Reviewer: (pending)

Documentation:

Release Notes:

Details:

+ having setscrollevents=true causes the  
LzTextSprite.setScrolling(true) to be called, but
__updatefieldsize needs to be called to get the new scrollHeight and  
send it back to the LFC.
But you need to actually wait for some time period to let the browser  
re-layout the scrolldiv.
So I used a 200msec timer. This seems like a non-robust solution, but  
I am not sure how
to fix it otherwise.

Tests:

testcase.lzx from bug report, and then check that tfield.maxscroll is  
greater than 1

Files:
M      WEB-INF/lps/lfc/kernel/dhtml/LzTextSprite.js

Changeset: http://svn.openlaszlo.org/openlaszlo/patches/20091102-hqm-F.tar
Henry Minsky | 3 Nov 04:23 2009
Picon

Re: [Laszlo-reviews] For Review: Change 20091030-hqm-t Summary: fix implementation of xpath last()

if anyone gets a chance to review this..

On Fri, Oct 30, 2009 at 5:10 PM, Henry Minsky <henry.minsky <at> gmail.com> wrote:
Change 20091030-hqm-t by hqm <at> badtzmaru.home on 2009-10-30 16:51:30 EDT
   in /Users/hqm/openlaszlo/trunk6
   for http://svn.openlaszlo.org/openlaszlo/trunk

Summary: fix implementation of xpath last()

New Features:

Bugs Fixed: LPP-298

Technical Reviewer: andre
QA Reviewer: max
Doc Reviewer: (pending)

Documentation:

Release Notes:

Details:

Andre's suggested implemenation of last() operator on LzDataAttrBind.__LZHandleMultiNodes


Tests:

test case from bug report

<canvas debug="true" >
 <dataset name="dsBooks" >
 <bookshelf>
  <book /><book />   <book /><book />
 </bookshelf>
 </dataset>

 <text>
 <attribute name="books" value="$path{'dsBooks:/bookshelf/book/last()'}" type="number"/>
 <handler name="onbooks" >
   this.format("found %d book(s)", this.books);
 </handler>
 </text>
</canvas>


Files:
M      WEB-INF/lps/lfc/data/LzDataAttrBind.lzs

Changeset: http://svn.openlaszlo.org/openlaszlo/patches/20091030-hqm-t.tar



--
Henry Minsky
Software Architect
hminsky <at> laszlosystems.com


_______________________________________________
Laszlo-reviews mailing list
Laszlo-reviews <at> www.openlaszlo.org
http://www.openlaszlo.org/mailman/listinfo/laszlo-reviews

Gmane