[Laszlo-dev] [Laszlo-user] setting timeout to longer than 30 seconds

[Sarah Allen]

  FYI


-------- Begin forwarded message --------
Subject: Re: FW: Re: [Laszlo-user] setting timeout to longer than 30 
seconds
Date: 12/27/2007  5:37:46 PM
 From: Adam Wolff <adam at elasticprocess.com>
To: Sarah Allen <sallen at laszlosystems.com>
CC: Max Carlson <max at laszlosystems.com>

well the issue is that you can't tell the difference (in Flash
6 at least) between a 404 and a timeout, so a really long
timeout setting leads to long waits if you have requests that
don't complete. it was a learn-ability thing; users don't
expect everything to get gummed up if you make a few bad
requests.

A

On Dec 27, Sarah Allen wrote:

> any memory of this?
>
>
> -------- Begin forwarded message --------
> Subject: Re: [Laszlo-user] setting timeout to longer than 30 seconds
> Date: 12/27/2007  2:53:52 PM
> From: Henry Minsky <henry.minsky at gmail.com>
> To: Sarah Allen <sallen at laszlosystems.com>, OpenLaszlo development 
> list <laszlo-dev at openlaszlo.org>
>
> I can't remember what the original thinking was on having a minute or
> less timeout  by default, need to ask david/adam/max.
>
> I think longer timeouts ought to be fine and should probably be the
> default, as Tucker has suggested from time to time.
>
>
>
>
> On Dec 27, 2007 5:48 PM, Sarah Allen <sallen at laszlosystems.com> wrote:
>> huh.  I always thought the 30 second limit was a Flash setting, not 
>> an
>> OL setting.  Is there any underlying limit that we should be 
>> concerned
>> about?  We're seeing legitimate requests from some users that take
>> longer than 30 seconds on slow connections and I was thinking of just
>> increasing this to two or three minutes.  Other than the delay in
>> finding out your server is down, are there any other drawbacks to a 
>> long
>> timeout?
>>
>> Thanks,
>> Sarah
>>
>>
>>
>> On Thu, Dec 27, 2007 at 12:54 PM, Henry Minsky wrote:
>>
>>> Here is the flow of control for SWF data request in LPS 4
>>>
>>> + The constructor for LzDataset sets its timeout in this code:
>>>
>>>     if ('timeout' in args && args.timeout) {
>>>         this.timeout = args.timeout;
>>>     } else {
>>>         this.timeout = canvas.dataloadtimeout;
>>>     }
>>>
>>>
>>> So it default to the value of the canvas.dataloadtimeout, if no
>>> explicit timeout init arg
>>> is passed in.
>>>
>>> + In the LPS 4 world where we have the new LzHTTPDataProvider, the
>>> timeout
>>> is passed along by LzHTTPDataProvider .doRequest() to the  SWF 
>>> kernel
>>> LzHTTPLoader object
>>>
>>>         tloader.setTimeout(dreq.timeout);
>>>
>>>
>>>
>>> + This timeout gets passed in to the good old LzLoader kernel 
>>> object,
>>> which gets
>>> then gets passed  to the LzLoadQueue service.
>>>
>>> LzHTTPLoader.prototype.setTimeout = function (timeout) {
>>>     this.timeout = timeout;
>>>     this.lzloader.timeout = timeout;
>>> }
>>>
>>> + The LzLoader creates a "load request object", which happens to be 
>>> a
>>> Flash native XML object, and it sets the "timeout" property to the
>>> LzLoader's timeout value.
>>>
>>>
>>> + Then the "load object" is passed to LzLoadQueue, which is
>>> responsible for actually sending the data requests. The timeout
>>> property is used to checked
>>> by a delegate which is triggered by an LzTimer event.
>>>
>>>
>>> On Dec 27, 2007 3:42 PM, Sarah Allen <sallen at laszlosystems.com> 
>>> wrote:
>>>> what does this  do "under the hood" and will it apply to a Flash
>>>> Upload
>>>> request which is not currently supprted by OpenLaszlo but which we 
>>>> do
>>>> with a custom compoent which embeds ActionScript?
>>>>
>>>> Thanks,
>>>> Sarah
>>>>
>>
>
>
>
>