Hello, Does anyone else have a problem with Ajax.Updater where you have some links "x, y and z" which call Updater to load content into your target container on the page somewhere.. Click "x".. that''s fine... then maybe click "y".. etc. But, if the user''s connection or the server''s a bit slow and they click "y" before the "x" has finished loading the page kinda flickers.. and the user experience is a little spoiled because the page starts to not keep up with the user''s interaction... I''ve made a little change to our version of prototype ( half a dozen lines or so ) which, if you give the parameter "override:true" to Ajax.Updater() it will record the stack of requests going to each target. That way, if there are two or more concurrent requests who''s content will eventually end up in the same place only the last one gets executed. What this _doesn''t_ do is actually cancel the requests.. ( I don''t know how best to do that! ) it simply makes sure that only the last request in the stack get''s inserted into the target ( and at the moment only the last request''s javascript gets eval()''ed ) . Anyhow - just wondering if this sort of thing is worthwhile enough to comit back into the cvs?? Would anyone find this useful, is it the right sort of thing to commit? Thanks, Jimbo. PS. Since this is my first post to this list: Thanks Sam, et al! Great library! --~--~---------~--~----~------------~-------~--~----~ You received this message because you are subscribed to the Google Groups "Ruby on Rails: Spinoffs" group. To post to this group, send email to rubyonrails-spinoffs-/JYPxA39Uh5TLH3MbocFFw@public.gmane.org To unsubscribe from this group, send email to rubyonrails-spinoffs-unsubscribe-/JYPxA39Uh5TLH3MbocFFw@public.gmane.org For more options, visit this group at http://groups.google.com/group/rubyonrails-spinoffs?hl=en -~----------~----~----~----~------~----~------~--~---
Marius Feraru
2007-Feb-01 12:22 UTC
Re: Prototype, Ajax.Updater - stacking requests (latency patch?)
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Jimbo wrote:> I''ve made a little change to our version of prototype ( half a dozen > lines or so ) which, if you give the parameter "override:true" to > Ajax.Updater() it will record the stack of requests going to each > target. > That way, if there are two or more concurrent requests who''s content > will eventually end up in the same place only the last one gets > executed.Seeing some code would have been nice. :( IMHO, there''s no need for such an "override" parameter. Better yet, simply twist Ajax.Updater to behave like this. Anyway, you''ll have to be pretty careful when deciding if an "older" request has been already triggered, and I specifically mean to mind about the "insertion" parameter. A simple rule would be: "there is no such thing as <<older>> request if <<insertion>> is <<true>>", but that again is wrong if you think about it, because order matters when doing insertions.> What this _doesn''t_ do is actually cancel the requests.. ( I don''t > know how best to do that! )Agreed, we currently don''t have a fancy "abort()" method. IIRC there was some discussion about this right on this list (or maybe on our tracker?)> Anyhow - just wondering if this sort of thing is worthwhile enough to > comit back into the cvs?? Would anyone find this useful, is it the > right sort of thing to commit?Useful? Yes, could be. But let''s discuss a little bit about it, to find the best/safest way of implementing it, in order not to get things even worse as they are right now ;-) cheers - -- Marius Feraru -----BEGIN PGP SIGNATURE----- iD8DBQFFwdudtZHp/AYZiNkRAuZHAKDunS645uJi2y9Hs9zd1LudehhGQgCeIPtl 6fsVk1uw/TWSvicLWafmXJ0=aOm3 -----END PGP SIGNATURE----- --~--~---------~--~----~------------~-------~--~----~ You received this message because you are subscribed to the Google Groups "Ruby on Rails: Spinoffs" group. To post to this group, send email to rubyonrails-spinoffs-/JYPxA39Uh5TLH3MbocFFw@public.gmane.org To unsubscribe from this group, send email to rubyonrails-spinoffs-unsubscribe-/JYPxA39Uh5TLH3MbocFFw@public.gmane.org For more options, visit this group at http://groups.google.com/group/rubyonrails-spinoffs?hl=en -~----------~----~----~----~------~----~------~--~---
Colin Mollenhour
2007-Feb-01 14:56 UTC
Re: Prototype, Ajax.Updater - stacking requests (latency patch?)
<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<html>
<head>
<meta content="text/html;charset=ISO-8859-1"
http-equiv="Content-Type">
</head>
<body bgcolor="#ffffff" text="#000000">
Sounds like a good idea, although I think this could be handled by the
programmer quite easily.<br>
<br>
For example, override the insertion function with one that checks a
counter before calling Element.update manually:<br>
function doUpdate(what){<br>
if(!element.updateCounter){ element.updateCounter
= 0; }<br>
element.updateCounter++;<br>
new
Ajax.Updater(element,''page'',{<br>
parameters:
''getData=''+what,<br>
insertion:
function(element,update){<br>
element.updateCounter--;<br>
if(element.updateCounter == 0){
Element.update(element,update); }<br>
}<br>
});<br>
}<br>
<br>
Note, this wouldn''t prevent evalScripts from happening if it was
enabled, or onComplete, etc.. from being called, so that would be a
benefit to the patch if it handles these cases correctly.<br>
<br>
<blockquote
cite="mid:45C1DB9D.5020809-9gptZ63fvgw@public.gmane.org"
type="cite">
<pre wrap=""><!---->Seeing some code would have been
nice. :(
</pre>
</blockquote>
Agreed, we gotta see the code!<br>
<blockquote
cite="mid:45C1DB9D.5020809-9gptZ63fvgw@public.gmane.org"
type="cite">
<pre wrap="">
IMHO, there''s no need for such an "override" parameter.
Better yet, simply
twist Ajax.Updater to behave like this.</pre>
</blockquote>
I''m not so sure.. Say your insertion is something besides
Element.update. If it is Insertion.Bottom then you may not want any
insertions overridden. Rather than having Prototype test which
insertion you''re using, the override parameter may be a simpler and
cleaner solution. Let the programmer decide!<br>
<br>
See <a class="moz-txt-link-freetext"
href="http://prototypejs.org/contribute">http://prototypejs.org/contribute</a>
for info on submitting a patch.<br>
<br>
Colin<br>
<br>
--~--~---------~--~----~------------~-------~--~----~<br>
You received this message because you are subscribed to the Google Groups
"Ruby on Rails: Spinoffs" group. <br> To post to this
group, send email to
rubyonrails-spinoffs-/JYPxA39Uh5TLH3MbocFFw@public.gmane.org <br> To
unsubscribe from this group, send email to
rubyonrails-spinoffs-unsubscribe-/JYPxA39Uh5TLH3MbocFFw@public.gmane.org
<br> For more options, visit this group at
http://groups.google.com/group/rubyonrails-spinoffs?hl=en <br>
-~----------~----~----~----~------~----~------~--~---<br>
</body>
</html>
<br>
Marius Feraru
2007-Feb-01 16:33 UTC
Re: Prototype, Ajax.Updater - stacking requests (latency patch?)
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Colin Mollenhour wrote:> For example, override the insertion function with one that checks a > counter before calling Element.update manually:Nifty. Continuing the idea, let''s do even more: switch "counter" to "updaters", an array (or maybe hash?) containing pending updaters for that element. This way they could be manipulated easily, both internally (by Ajax.Updater methods) and externally (by any picky user) :)>> IMHO, there''s no need for such an "override" parameter. Better yet, >> simply twist Ajax.Updater to behave like this. > I''m not so sure.. Say your insertion is something besides Element.update. > If it is Insertion.Bottom then you may not want any insertions > overridden.That''s exactly what I was proposing (it there''s an "insertion" provided, then don''t bother about replacing old requests). But I can see where you''re going...> Rather than having Prototype test which insertion you''re using, the > override parameter may be a simpler and cleaner solution. Let the > programmer decide!Good enough, so an "override" with "false" as default would be better. :)) - -- Marius Feraru -----BEGIN PGP SIGNATURE----- iD8DBQFFwhZttZHp/AYZiNkRArLvAKDehUV0o0Or4ifimG58PRfKwT7MIwCfYXiy nVEMkZq3UdZlK6XSWuC2KuU=gpah -----END PGP SIGNATURE----- --~--~---------~--~----~------------~-------~--~----~ You received this message because you are subscribed to the Google Groups "Ruby on Rails: Spinoffs" group. To post to this group, send email to rubyonrails-spinoffs-/JYPxA39Uh5TLH3MbocFFw@public.gmane.org To unsubscribe from this group, send email to rubyonrails-spinoffs-unsubscribe-/JYPxA39Uh5TLH3MbocFFw@public.gmane.org For more options, visit this group at http://groups.google.com/group/rubyonrails-spinoffs?hl=en -~----------~----~----~----~------~----~------~--~---
Jimbo
2007-Feb-01 19:16 UTC
Re: Prototype, Ajax.Updater - stacking requests (latency patch?)
Sorry guys! Was just dipping my toe in the water, hence no code...
plus had just come from the Google Maps API newsgroup which has says
"don''t post code" in big red letters! ;-)
Here''s the code then:
First, added our stack object to "Ajax" itself... just next to
activeRequestCount
=== var Ajax ==
targetStack : {}
=== end =====
. . . then added the flag to "options". It defaults to false for
backward compatibility..
=== Ajax.Base.prototype ===
this.options = {
method: ''post'',
asynchronous: true,
contentType: ''application/x-www-form-urlencoded'',
encoding: ''UTF-8'',
parameters: '''',
override: false
}
=== end ====
Ok... this is the bit of code that records the stacking .. it goes in
the Ajax.Updater''s initialize just below
"this.setOptions(options)" ...
=== [ Ajax.Updater : initialize ] ==
if ( this.options.override ){
// the target where I''ll be inserted.
var targetId = this.container.success.id
if ( !Ajax.targetStack[targetId] ){
// ** the first to this target.
// stack me incase another comes along.
Ajax.targetStack[targetId] = [ this ];
} else {
// ** the n''th item in this stack.
// mark the previous item as overridden ...
Ajax.targetStack[targetId][ Ajax.targetStack[targetId].length -
1 ].overridden = 1;
// ... and add me to the stack.
Ajax.targetStack[targetId].push ( this );
}
}
=== [ end ] =====
.. and then last but not least, in the updateContent function just
wrap some of it with an if..
=== [ updateContent ] ====
if ( !this.overridden ){
/**
* I''ve included the "if (!this.options.evalScripts)"
* and "if (receiver = $(receiver)) {" statements in here
* but then closed out before the "if (this.success()) {"
* which would seem to need to run, just cos you''re masking the
content insert.
*/
}
Ok.. so I don''t know how neatly that''s going to present so
I''ve posted
a patch to ajax.js here:
http://yetanotheruser.blogspot.com/2007/02/prototype-update-stack.html
... if that''s easier..
.. also - it''s not been exhaustively tested so I''m not sure if
it''s
behaving in firefox... :-/
I agree with you guys, being able to control it a bit more would be
fun... or set ''priorities'' even... :-)
Re. override = true/false... I''ve got it false by default at the
moment, as that would seem more compatible?
Re. hashing it : yep, should behave like you want. In my example for
instance we have a main content block in the centre of the page which
displays listings ( this is a recruitment website ) of jobs let''s
say.. sometimes they take a while if it''s a big list and a slow
connection.
We''ve also got a little e-poll on the LHS which uses Ajax.Updater() to
give you the stats when you hit an answer.
Set up like this though, you can click a few listings links and it''s
all fine - the last one to get clicked is the one the users sees in
the main content block. While they''re waiting though, if they answer
the questionnaire that request ( for the answers/graph ) doesn''t
interfere with the stack for the main page body...
Anyhow... hope this helps :-) ...
Cheerio,
Jim
On Feb 1, 4:33 pm, Marius Feraru <altb...-9gptZ63fvgw@public.gmane.org>
wrote:> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
> Colin Mollenhour wrote:
> > For example, override the insertion function with one that checks a
> > counter before calling Element.update manually:
>
> Nifty. Continuing the idea, let''s do even more: switch
"counter" to
> "updaters", an array (or maybe hash?) containing pending updaters
for that
> element. This way they could be manipulated easily, both internally (by
> Ajax.Updater methods) and externally (by any picky user) :)
>
> >> IMHO, there''s no need for such an "override"
parameter. Better yet,
> >> simply twist Ajax.Updater to behave like this.
> > I''m not so sure.. Say your insertion is something besides
Element.update.
> > If it is Insertion.Bottom then you may not want any insertions
> > overridden.
>
> That''s exactly what I was proposing (it there''s an
"insertion" provided,
> then don''t bother about replacing old requests). But I can see
where you''re
> going...
>
> > Rather than having Prototype test which insertion you''re
using, the
> > override parameter may be a simpler and cleaner solution. Let the
> > programmer decide!
>
> Good enough, so an "override" with "false" as default
would be better. :))
>
> - --
> Marius Feraru
> -----BEGIN PGP SIGNATURE-----
>
> iD8DBQFFwhZttZHp/AYZiNkRArLvAKDehUV0o0Or4ifimG58PRfKwT7MIwCfYXiy
> nVEMkZq3UdZlK6XSWuC2KuU> =gpah
> -----END PGP SIGNATURE-----
--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups
"Ruby on Rails: Spinoffs" group.
To post to this group, send email to
rubyonrails-spinoffs-/JYPxA39Uh5TLH3MbocFFw@public.gmane.org
To unsubscribe from this group, send email to
rubyonrails-spinoffs-unsubscribe-/JYPxA39Uh5TLH3MbocFFw@public.gmane.org
For more options, visit this group at
http://groups.google.com/group/rubyonrails-spinoffs?hl=en
-~----------~----~----~----~------~----~------~--~---