At 12:39 AM 5/18/2007, Jim Grandy wrote:
OpenLaszlo 4.0.2 was released to the web today. This is a minor
release resolving almost 50 issues reported against OpenLaszlo
4.0.0, and adding a few new features.
Thanks - glad to here it.
So far, I'm wondering if anyone else has seen the following
issues going from 4.0.0 to 4.0.2:
1. <roundrectbutton> seems to have broken under DHTML. On first
running 4.0.2 I got 7 compile errors not found when using
4.0.0. The message complained about <boxview> and worked
if I commented the <boxview> out. (at least under SWF)
But then the exact same code running with the <boxview>
removed in SWF compiled but failed to initialize in DHTML.
The debugger complains "this.context has no properties"
and I get no screen what so ever. Firebug narrowed down
the exact code being run as:
this.context.clearRect(0,0,this.width,this.height)
Which has a stack trace including:
initialize(roundrectbutton name: selectButton id: selectButton ,
Object $refs=Object borderColor=10066329 borderRadius=15, [Object
attrs=Object name=view, Object name=__LZUserClassPlacementObject
attrs=content], true)
Given that, I switch the two <roundrectbutton>'s to plain
<button>'s and it all worked again. I then put the
<boxview> back in and that now worked and compiled
error free too!
It seems <roundrectbutton> broke in 4.0.2, both in terms
of the runtime code, and compiling in DHTML seems to
have some kind of conflict.
2. Compiling under 4.0.2 seems to be about the same speed,
but the drawing of the screens on initialization has just
slowed down dramatically using DHTML. I'm getting serious
hour glass behavior in DHML but quick behavior in SWF.
3. I usually include in my own app the same Style picker that is
seen in the Component Sampler. When I switch styles,
my <menubar> works for any menus one or two levels deep,
but not three levels. It displays the menu but immediately
hides it after one second. Even if you're quick enough
to navigate to menu three levels deep, it does not respond
to clicking or even highlight before the click. This
occurs only in DHTML, and it also appears only to
happen when I have previously clicked on the menu.
If I have not used the menu, and change styles, there
is no problem. It's only the sequence of Menu Click ->
change style -> Access a menu three levels deep.
4. I have a data grid with a check box in a grid column
and either "Yes" or "No" text to the side of it. Code
was copied from some samples and looks like:
<checkbox datapath="CLEARED/text()" text="Yes" value="true">
<method event="onclick">
if ( this.value == true ){
this.setAttribute("text","Yes");
} else {
this.setAttribute("text","No");
}
</method>
</checkbox>
This seemed to work in 4.0.0 but is giving very strange behavior
in both SWF and DHTML. When ever I click the check box the
text does correctly change from either Yes to No or No to Yes.
However, scrolling through the grid, a second page of items
shows data items that were checked with No beside it.
The wrong box seems to continuously travel down the page
by one item. In other words, if I changed row 2 on page 1
of the items, from Checked/Yes to Unchecked/No and scroll
down one page, row 3 on page 2 shows Checked, but No.
The correct data value in the dataset is Yes for that item.
If I scroll to page 3 of items, then row 4 is Checked but
says No while the item is Yes in the dataset.
That is basically about the text part being wrong -
when I uncheck an item scroll down, but back up
again, the same item is magically back to being
checked again.
I should say I do not believe this happened in
4.0.0, but then again perhaps it did and I
just didn't notice. In any case, with the wrong
text seemingly traveling down the screen one
row at a time, I'm almost thinking somewhere in
the grid code is an off by one error. (but,
well, maybe that's just me) ;-)
5. This is not an issue, but something I noticed
got cleared up. (and may not have been reported)
DHTML used to have some issues where resizing a
screen worked the first time, but resizing a
second time didn't. The second time the window
began to very annoyingly jump to a smaller size
the minute you clicked on the resize image.
And then when you moved the mouse it would
resize, but not where your mouse was located -
perhaps offset 50 pixels less in each of the
X and Y directions. This behavior is now
gone. Whew, that's a good thing.
=======================================================================
Geoff Crawford Phone: (973) 361 - 4224
Innov8 Computer Solutions, LLC FAX: (973) 537 - 6946
711 Route 10 East, Suite 204 Email: [EMAIL PROTECTED]
Randolph NJ 07869 Web: http://www.processware.biz
ProcessWare ERP for the Chemical Industry
