[Laszlo-dev] Query args width and height processed by LaszloCanvas.lzs#construct()

[P T Withington]

Maybe what you are getting at is that we finally need to address this  
improvement:

   http://jira.openlaszlo.org/jira/browse/LPP-3479

so that private LZX parameters do not pollute the query param space  
that a user program might like to use?

On 2009-07-20, at 16:59EDT, Raju Bitter wrote:

> Yes, I know that the LaszloCanvas code is looking for those args.  
> Isn't it like there are 4 sources for the width/height value?
>
> 1) canvas attributes
> 2) embedding code
> 3) query args for SWF file
> 4) default values 400/400, in case the query args contain an invalid  
> value
>
> But what you are saying means: if a developer wants to use full  
> screen feature, he must take caution not pass any values for width  
> and height into the app at runtime, or full screen mode might not  
> auto-adjust the canvas size.
>
>
> On Jul 20, 2009, at 5:18 PM, P T Withington wrote:
>
>> I don't know the history, but notice that the canvas is _looking_  
>> for those params.  It's not that query params override an  
>> application var or setting (which could be a security issue).  The  
>> developer of the canvas specifically asks for these params.
>>
>> The history of canvas size is long and tortured, and there are a  
>> lot of crazy paths in the compiler to handle absolute vs.  
>> percentage sizing.  I am a big proponent for letting the size be  
>> dynamic, taken from the browser or query args, rather than having  
>> to recompile your app to resize it.
>>
>> On 2009-07-19, at 11:21EDT, Max Carlson wrote:
>>
>>> Perhaps we want a different behavior when fullscreen is on.  It  
>>> should act much more like the default of 100% by 100%. Maybe the  
>>> part of canvas.construct() that registers for screen size  
>>> notifications should be refactored to a separate method so it can  
>>> be called when fullscreen is turned on...
>>>
>>> -----Original Message-----
>>> From: Raju Bitter <rajubitter at me.com>
>>> Sent: Sunday, July 19, 2009 8:47 AM
>>> To: Raju Bitter <rajubitter at me.com>
>>> Cc: Laszlo-dev bug reporting <laszlo-dev at openlaszlo.org>
>>> Subject: Re: [Laszlo-dev] Query args width and height processed by  
>>> LaszloCanvas.lzs#construct()
>>>
>>> And if you don't want to read the forum post: he attached a width  
>>> and
>>> height parameter to the SWF for whatever reason (customer
>>> requirement), and never expected that to influence the behavior full
>>> screen.
>>>
>>> So my question actually is:
>>> 1) is it a documented behavior, that you can override width and  
>>> height
>>> with query vars
>>> 2) if it's a documented behavior or a "feature", I need to mention  
>>> in
>>> the fullscreen documentation what effect it will have
>>>
>>> On Jul 19, 2009, at 4:41 PM, Raju Bitter wrote:
>>>
>>>> That means: if you have the app running at 600x600 without full
>>>> screen, and you go into full screen on a 1440x900 screen, the  
>>>> canvas
>>>> will remain at 600x600 in the upper left corner of the screen.
>>>>
>>>> Of course you can easily avoid that by NOT passing any value  
>>>> through
>>>> query vars into the Laszlo app, but if you look into the forum post
>>>> that caused this discussion you'll see that a user got confused by
>>>> that behavior.
>>>>
>>>> I was just investigating what's going on to help this guy.
>>>>
>>>> On Jul 19, 2009, at 3:46 PM, Sarah Allen wrote:
>>>>
>>>>> um... so what is the conflict?
>>>>>
>>>>> On Jul 19, 2009, at 6:30 AM, Raju Bitter wrote:
>>>>>
>>>>>> Actually it happens on canvas construction, in the construct()
>>>>>> method. Which means, it overrides the values in your <canvas />
>>>>>> declaration within LZX. As a result canvas will - even in full
>>>>>> screen mode - still report the static width & height passed into
>>>>>> the app.
>>>>>>
>>>>>> On Jul 19, 2009, at 3:09 AM, Sarah Allen wrote:
>>>>>>
>>>>>>> It does seem like a useful feature in case you want to use a swf
>>>>>>> embedded in something that is not a browser (although I suppose
>>>>>>> you could always write it yourself in code, so maybe it is not
>>>>>>> worth maintaining)
>>>>>>>
>>>>>>> Just curious... why does it interfere with full screen mode?  I
>>>>>>> would think that would happen long after canvas construct...
>>>>>>>
>>>>>>> On Jul 18, 2009, at 5:47 PM, Raju Bitter wrote:
>>>>>>>
>>>>>>>> A user in the forums ran into a problem with f
>>>
>>> [The entire original message is not included]
>>>
>>
>