可以将文章内容翻译成中文,广告屏蔽插件可能会导致该功能失效(如失效,请关闭广告屏蔽插件后再试):
问题:
I'm trying to use the jQuery alerts dialog library from http://abeautifulsite.net/notebook/87 instead of the default alerts (which look pretty awful in my opinion). This seems to be a great library, but there is not an example of how to use the jConfirm library.
I need to do something like this:
function confirm() {
var result = false;
var response = false;
jConfirm('are you sure?', 'Confirmation Dialog',
function(r) {
result = r;
response = true;
return r;
});
if (response == true) {
alert(result);
return result;
}
else {
//wait for response
alert('hi');
}
}
and my call from my .net button:
I've posted a comment on the plugin's website (just this morning) and did Google searches for javascript and waiting for a callback to complete with no results.
Any ideas on how to use the callback correctly to get the result, before the rest of the javascript executes?
Thanks.
回答1:
jConfirm('are you sure?', 'Confirmation Dialog',
function(r) {
result = r;
response = true;
return r;
}
);
if (response == true) {
This betrays a misunderstanding of the sequence of events that occurs using asynchronous code. Just because you've written it inline doesn't mean it's going to execute strictly top-to-bottom.
- jConfirm is called, receiving a function as one of its parameters, which it remembers.
- jConfirm displays its UI on the page and returns immediately.
- The 'if (response==true)' line executes. Really this should just read 'if (response)', the boolean comparison is superfluous. But in any case response is of course false. Your function gives up and exits, giving control back to the browser.
- The user clicks jConfirm's UI.
- jConfirm only now jumps into action and calls back the function you gave it and it remembered earlier.
- Your nested function sets response true, far too late for the 'if (response==true)' condition to do anything with it.
You have written "//wait for response" as an alternative, but there is no JavaScript code you can write that will actually do that. Your function must return to give control back to the browser, before the browser can fire the click events on the jConfirm UI that make processing proceed.
Ways to make asynchronous code work in a synchronous context (and vice versa) exist - in particular threads and coroutines (and their limited relation generators). But JavaScript has none of these features, so you must write your code to fit the synchronous-or-asynchronous model your library is using.
回答2:
You've just hit a big limitation in JavaScript. Once your code enters the asynchronous world, there is no way to get back to a classic procedural execution flow.
In your example, the solution would be to make a loop waiting for the response to be filled. The problem is that JavaScript does not provide any instruction that will allow you to loop indefinitely without taking 100% of the processing power. So you will end up blocking the browser, sometimes to the point where your user won't be able to answer the actual question.
The only solution here is to stick to the asynchronous model and keep it. My advice is that you should add a callback to any function that must do some asynchronous work, so that the caller can execute something at the end of your function.
function confirm(fnCallback)
{
jConfirm('are you sure?', 'Confirmation Dialog', function(r)
{
// Do something with r
fnCallback && fnCallback(r); // call the callback if provided
});
}
// in the caller
alert('begin');
confirm(function(r)
{
alert(r);
alert('end');
})
回答3:
Since the callback is asynchronous (at least, in the sense that it's waiting on the user to do something), it might be easier to handle what you need to inside the callback:
function confirm() {
jConfirm('are you sure?', 'Confirmation Dialog', function(r) {
if (r) doSomething();
});
}
@klogan [comments]
I assume you got these from here?
The page gives you your answer: (look under Usage)
These methods do not return the same values as confirm() and prompt(). You must access the resulting values using a callback function. (See the demo for more details.)
@klogan
The point I'm trying to make is that there isn't really an easy way to accomplish what you want. You're trying to correlate procedural and event-driven programming -- something JavaScript doesn't help you do.
The simplest (though, risky) solution is to use a pseudo-infinite-loop. But, if callback
never gets called, you now have an actual infinite loop. And, depending on the JavaScript engine, you might kill the browser waiting.
Point: Your best bet is to avoid this trying to force event-driven into procedural.
function confirm() {
var result = false;
var response = false;
jConfirm('are you sure?', 'Confirmation Dialog',
function(r) {
result = r;
response = true;
});
while(!response) continue; // wait
return result;
}
回答4:
Technically, yes, but I wouldn't do that on a website.
Take a look at Narrative JavaScript, which is based off Narcissus.
Narrative JavaScript is a small extension to the JavaScript language that enables blocking capabilities for asynchronous event callbacks. This makes asynchronous code refreshingly readable and comprehensible.
Selenium uses this technology.
Update
Check out JavaScript Strands:
JavaScript Strands adds coroutine and
cooperative threading support to the
JavaScript language to enable blocking
capabilities for asynchronous event
callbacks. This makes code that
utilizes asynchronous operation much
more linear, readable, and manageable.
Strands is built upon Narrative
JavaScript written by Neil Mix, and
much of Narrative JavaScript has
remained in Strands including much of
this documentation.
In JavaScript your code can't simply
wait until an event has fired -- the
event must always be handled by a
separate, asynchronous event handler.
Sometimes this is fine, but it often
forces what ought to be a simple
sequence of statements into gnarly
contortions. It also breaks the
ability to encapsulate functionality
because calling functions must know to
provide a callback handler. Strands
provides the ability to suspend and
resume threads of execution. Execution
can suspend resume when the event is
finished. This allows you to write
hard-to-read asynchronous event
handling in simple, linear, readable
code that encapsulates implementation.
回答5:
I would assume you'd have to grab the response like this. I don't think you need a callback.
function confirm() {
var response = jConfirm('are you sure?', 'Confirmation Dialog');
if (response) {
alert(result);
}
else {
//wait for response
alert('hi');
}
}
回答6:
I think I have come up with a possible solution to this problem. I was reading this article:
http://treasure4developer.wordpress.com/2008/06/23/calling-postback-event-from-javascript/
Basically the idea is that you force the postback from javascript, at first I found that the postback would work but wouldn't call my button event, but after reading the article I found that I could detect if it was javascript postback and just call a method to do the processing
Regards DotnetShadow
回答7:
How the dudes said, there is no way! but... how we use to say in my country (Brazil), le gambi:
we can do something like it:
<h:commandLink styleClass="linkValor xExcluir" value="x"
onclick="if(confirmar('Confirmar excluir.','Deseja realmente excluir este registro?', this)) return true; else return false;"
action="#{mBeanPesquisarLinhas.excluir}"/>
and the javascript:
function confirmar(titulo, msg, elemento) {
if ($(elemento).attr('sim')) {
$(elemento).removeAttr('sim');
return true;
} else if ($(elemento).attr('nao')) {
$(elemento).removeAttr('nao');
return false;
} else {
$("#dialog-confirm").html('<p>' + msg + '</p>').dialog({
resizable : false,
height : 200,
modal : true,
title : titulo,
buttons : {
"Sim" : function() {
$(this).dialog("close");
$(elemento).attr('sim', 'sim');
$(elemento).click();
},
"Não" : function() {
$(this).dialog("close");
$(elemento).attr('nao', 'nao');
$(elemento).click();
}
}
});
}
return false;
}
it's the same problem but using jquery-ui.
Bye and I hope this can help somebody.
回答8:
Think about callbacks as sending messages and it will result in better structure in your code.
回答9:
This thing works. Needs jQuery.
function myconfirm(kyssa, elm, e){ // neG
if(jQuery('#confirmquestion').data('result')){
var V = jQuery('#confirmquestion').data('result');
jQuery('#confirmquestion').remove();
return V == 'Y' ? true : false;
}else if(!jQuery('#confirmquestion').length){
jQuery('body').append('<div id="confirmquestion">'+
'<h4>Kinnitus</h4>'+
'<div id="kyssa">'+kyssa+'</div>'+
'<center>'+
'<button onclick="jQuery(\'#confirmquestion\').data(\'result\', \'Y\');">Jah</button> '+
'<button onclick="jQuery(\'#confirmquestion\').data(\'result\', \'N\');">Ei</button></div>'+
'</center>');
jQuery('#confirmquestion button').click(function(){
jQuery(elm).trigger(e.type);
})
}
return false;
}
onChange="if(myconfirm(\'Saada kiri: \'+jQuery(this).find(\'option:selected\').html()+\' ?\', this, event)) { ... }"
CSS
#confirmquestion{
border:1px solid #999;
background:white;
padding-bottom: 30px;
position:fixed;
width:300px;
font-size:12px;
top:45%;
left:50%;
margin-left:-150px;
}
#confirmquestion h4 {
background:blue;
color:white;
margin:0;
padding: 2px 5px;
border-bottom:#777;
text-align:center;
}
#confirmquestion #kyssa {
padding: 30px 25px;
}