Re: Not sure what to do.....can I run this by you guys?

[Tom Glod via use-livecode]

Hi everyone here is the github repo for the stack demonstrating the
solution.  May not work on all cases but can for many.  Seems to work very
well for my needs. 10 requests ...no problem. have not done more.

https://github.com/makeshyft-tom-g/lc-single-domain-asynchronous-http

On Sun, Mar 18, 2018 at 11:54 PM, Tom Glod  wrote:

> Yup, realized that when I thought about and tried mikes suggestion
>
> Yes using a ? works just the same as # but may be safer in some way..
>
> I'm just completing demo stack for this .works flawlessly.
>
> Thanks to everyone who chimed in here to help me find solution...hopefully
> i can make it easier on others to publish a demo stack.
>
>
>
>
> On Sun, Mar 18, 2018 at 10:39 AM, Ben Rubinstein via use-livecode <
> use-livecode@lists.runrev.com> wrote:
>
>> I think this depends on the base URL.  If it's with a simple URL, there's
>> no reason why an appended '#' parameter should cause a browser to request
>> it fresh; as Mike says, it indicates scrolling to an anchor in the page. So
>> it's perfectly reasonable for a smart browser - or a caching layer on the
>> server, or in a proxy somewhere between you and the server - to use a
>> cached version of that page ignoring the '#' suffix.
>>
>> With a simple URL, the better choice is to use "?" to append a parameter
>> (I usually use 'the milliseconds' to guarantee that it's unique). Neither
>> the browser nor any caching layer on the chain between the user agent and
>> the source system should recognise it as the same URL, so this should force
>> a fresh load. You can just use e.g.
>> put "?" & the milliseconds after tURL
>>
>> or to be more explicit, name the parameter something e.g.
>> put "?nocache=" & the milliseconds after tURL
>> or
>>
>> put "?uniqueid=" & the milliseconds after tURL
>>
>> The only downside of using "?" is that if the URL may already includes
>> parameters starting with "?" (which is what I mean by "not a simple URL"),
>> you need to test for that and use "&" instead of "?" to indicate that this
>> is an additional parameter; and if this is a dynamic request, which uses
>> the 'query parameters' in the URL to decide what to return, you need to be
>> sure that adding your parameter won't affect it!
>>
>>
>> On 16/03/2018 13:31, Mike Bonner via use-livecode wrote:
>>
>>> Another way around the cache problem is to use the #2 trick at the end of
>>> the url.  Send each request with a pound and different number at the end
>>> of
>>> the url and it'll be seen as a new request thus doing an end run around
>>> the
>>> cache.  Since it designates an inline anchor position on the page, it
>>> should have zero affect on the way the url functions.  (unless things
>>> have
>>> changed, the associated anchor doesn't need to exist on the page)
>>>
>>> Thanks for the neat trick Charles. :)
>>>
>>> On Fri, Mar 16, 2018 at 7:24 AM, Tom Glod via use-livecode <
>>> use-livecode@lists.runrev.com> wrote:
>>>
>>> WowI'm impressedthats quite a hack Charles..I will study all this
 see how far I get.

 Thank you gentlemenyou are Rockstars!!

 On Fri, Mar 16, 2018 at 7:13 AM, Lagi Pittas via use-livecode <
 use-livecode@lists.runrev.com> wrote:

 Maybe not 100% reliable but 
>
> https://stackoverflow.com/questions/1341089/using-meta-
> tags-to-turn-off-caching-in-all-browsers
>
> Regards Lagi
>
> On 16 March 2018 at 09:48, Charles Warwick via use-livecode
>  wrote:
>
>>   Hi Tom,
>>
>> If the site you are trying to contact has CORS enabled appropriately,
>>
> then you can do something like this...
>
>>
>> With the LiveCode browser widget, you can call JavaScript functions
>>
> from

> LC script and have the JavaScript functions call LC handlers in return.
> JavaScript has the capability to perform asynchronous HTTP requests.
>
>>
>> You can create a HTML page that you automatically load up in the
>>
> browser

> widget that has a small JavaScript function which you can call from LC
>
 with

> ‘do in widget’.   All this function needs to do is issue an
> asynchronous
> HTTP call to the URL passed to it as a parameter and when it receives
> the
> data, return it back to your LC script by calling a nominated LC
> handler
> and passing the returned data as a parameter.
>
>>
>> The HTML page would look something like this:
>>
>> 
>> 
>> Javascript Async Test
>> 
>>
>> function httpGetAsync(theUrl)
>> {
>>  var xmlHttp = new XMLHttpRequest();
>>  xmlHttp.onreadystatechange = function() {
>>  if (xmlHttp.readyState == 4 && xmlHttp.status == 200)
>>  liveCode.httpRequestComplete(theUrl,
>>
> xmlHttp.responseText);

>  }
>>  xmlHttp.open("GET", 

Re: Not sure what to do.....can I run this by you guys?

[Tom Glod via use-livecode]

Yup, realized that when I thought about and tried mikes suggestion

Yes using a ? works just the same as # but may be safer in some way..

I'm just completing demo stack for this .works flawlessly.

Thanks to everyone who chimed in here to help me find solution...hopefully
i can make it easier on others to publish a demo stack.




On Sun, Mar 18, 2018 at 10:39 AM, Ben Rubinstein via use-livecode <
use-livecode@lists.runrev.com> wrote:

> I think this depends on the base URL.  If it's with a simple URL, there's
> no reason why an appended '#' parameter should cause a browser to request
> it fresh; as Mike says, it indicates scrolling to an anchor in the page. So
> it's perfectly reasonable for a smart browser - or a caching layer on the
> server, or in a proxy somewhere between you and the server - to use a
> cached version of that page ignoring the '#' suffix.
>
> With a simple URL, the better choice is to use "?" to append a parameter
> (I usually use 'the milliseconds' to guarantee that it's unique). Neither
> the browser nor any caching layer on the chain between the user agent and
> the source system should recognise it as the same URL, so this should force
> a fresh load. You can just use e.g.
> put "?" & the milliseconds after tURL
>
> or to be more explicit, name the parameter something e.g.
> put "?nocache=" & the milliseconds after tURL
> or
>
> put "?uniqueid=" & the milliseconds after tURL
>
> The only downside of using "?" is that if the URL may already includes
> parameters starting with "?" (which is what I mean by "not a simple URL"),
> you need to test for that and use "&" instead of "?" to indicate that this
> is an additional parameter; and if this is a dynamic request, which uses
> the 'query parameters' in the URL to decide what to return, you need to be
> sure that adding your parameter won't affect it!
>
>
> On 16/03/2018 13:31, Mike Bonner via use-livecode wrote:
>
>> Another way around the cache problem is to use the #2 trick at the end of
>> the url.  Send each request with a pound and different number at the end
>> of
>> the url and it'll be seen as a new request thus doing an end run around
>> the
>> cache.  Since it designates an inline anchor position on the page, it
>> should have zero affect on the way the url functions.  (unless things have
>> changed, the associated anchor doesn't need to exist on the page)
>>
>> Thanks for the neat trick Charles. :)
>>
>> On Fri, Mar 16, 2018 at 7:24 AM, Tom Glod via use-livecode <
>> use-livecode@lists.runrev.com> wrote:
>>
>> WowI'm impressedthats quite a hack Charles..I will study all this
>>> see how far I get.
>>>
>>> Thank you gentlemenyou are Rockstars!!
>>>
>>> On Fri, Mar 16, 2018 at 7:13 AM, Lagi Pittas via use-livecode <
>>> use-livecode@lists.runrev.com> wrote:
>>>
>>> Maybe not 100% reliable but 

 https://stackoverflow.com/questions/1341089/using-meta-
 tags-to-turn-off-caching-in-all-browsers

 Regards Lagi

 On 16 March 2018 at 09:48, Charles Warwick via use-livecode
  wrote:

>   Hi Tom,
>
> If the site you are trying to contact has CORS enabled appropriately,
>
 then you can do something like this...

>
> With the LiveCode browser widget, you can call JavaScript functions
>
 from
>>>
 LC script and have the JavaScript functions call LC handlers in return.
 JavaScript has the capability to perform asynchronous HTTP requests.

>
> You can create a HTML page that you automatically load up in the
>
 browser
>>>
 widget that has a small JavaScript function which you can call from LC

>>> with
>>>
 ‘do in widget’.   All this function needs to do is issue an asynchronous
 HTTP call to the URL passed to it as a parameter and when it receives
 the
 data, return it back to your LC script by calling a nominated LC handler
 and passing the returned data as a parameter.

>
> The HTML page would look something like this:
>
> 
> 
> Javascript Async Test
> 
>
> function httpGetAsync(theUrl)
> {
>  var xmlHttp = new XMLHttpRequest();
>  xmlHttp.onreadystatechange = function() {
>  if (xmlHttp.readyState == 4 && xmlHttp.status == 200)
>  liveCode.httpRequestComplete(theUrl,
>
 xmlHttp.responseText);
>>>
  }
>  xmlHttp.open("GET", theUrl, true); // true for asynchronous
>  xmlHttp.send(null);
> }
> 
> 
> 
> 
> 
>
> You can either load that from a file into the browser widget’s URL or
>
 set its htmlText property accordingly...

>
> Then in LC, make sure you register the httpRequestComplete handler so
>
 that the widget can call it:

>
> set the javascriptHandlers of widget “browser” to “httpRequestComplete”
>
> After that, add a 

Re: Not sure what to do.....can I run this by you guys?

[Ben Rubinstein via use-livecode]

I think this depends on the base URL.  If it's with a simple URL, there's no 
reason why an appended '#' parameter should cause a browser to request it 
fresh; as Mike says, it indicates scrolling to an anchor in the page. So it's 
perfectly reasonable for a smart browser - or a caching layer on the server, 
or in a proxy somewhere between you and the server - to use a cached version 
of that page ignoring the '#' suffix.


With a simple URL, the better choice is to use "?" to append a parameter (I 
usually use 'the milliseconds' to guarantee that it's unique). Neither the 
browser nor any caching layer on the chain between the user agent and the 
source system should recognise it as the same URL, so this should force a 
fresh load. You can just use e.g.

put "?" & the milliseconds after tURL

or to be more explicit, name the parameter something e.g.
put "?nocache=" & the milliseconds after tURL
or

put "?uniqueid=" & the milliseconds after tURL

The only downside of using "?" is that if the URL may already includes 
parameters starting with "?" (which is what I mean by "not a simple URL"), you 
need to test for that and use "&" instead of "?" to indicate that this is an 
additional parameter; and if this is a dynamic request, which uses the 'query 
parameters' in the URL to decide what to return, you need to be sure that 
adding your parameter won't affect it!


On 16/03/2018 13:31, Mike Bonner via use-livecode wrote:

Another way around the cache problem is to use the #2 trick at the end of
the url.  Send each request with a pound and different number at the end of
the url and it'll be seen as a new request thus doing an end run around the
cache.  Since it designates an inline anchor position on the page, it
should have zero affect on the way the url functions.  (unless things have
changed, the associated anchor doesn't need to exist on the page)

Thanks for the neat trick Charles. :)

On Fri, Mar 16, 2018 at 7:24 AM, Tom Glod via use-livecode <
use-livecode@lists.runrev.com> wrote:


WowI'm impressedthats quite a hack Charles..I will study all this
see how far I get.

Thank you gentlemenyou are Rockstars!!

On Fri, Mar 16, 2018 at 7:13 AM, Lagi Pittas via use-livecode <
use-livecode@lists.runrev.com> wrote:


Maybe not 100% reliable but 

https://stackoverflow.com/questions/1341089/using-meta-
tags-to-turn-off-caching-in-all-browsers

Regards Lagi

On 16 March 2018 at 09:48, Charles Warwick via use-livecode
 wrote:

  Hi Tom,

If the site you are trying to contact has CORS enabled appropriately,

then you can do something like this...


With the LiveCode browser widget, you can call JavaScript functions

from

LC script and have the JavaScript functions call LC handlers in return.
JavaScript has the capability to perform asynchronous HTTP requests.


You can create a HTML page that you automatically load up in the

browser

widget that has a small JavaScript function which you can call from LC

with

‘do in widget’.   All this function needs to do is issue an asynchronous
HTTP call to the URL passed to it as a parameter and when it receives the
data, return it back to your LC script by calling a nominated LC handler
and passing the returned data as a parameter.


The HTML page would look something like this:



Javascript Async Test


function httpGetAsync(theUrl)
{
 var xmlHttp = new XMLHttpRequest();
 xmlHttp.onreadystatechange = function() {
 if (xmlHttp.readyState == 4 && xmlHttp.status == 200)
 liveCode.httpRequestComplete(theUrl,