Help Logs Database

Undernet  |  EFnet  |  Quakenet  |  Freenode  |  Ircnet  |  Dalnet
Page: 1 2 3 4 5 6 7 8

<rajesh> simple xm nodeset
<rajesh> xml
<Craig_> why would u leave out the request header?
<alexmacy> hi
<Craig_> you* sorry :P msn has got to me
<rajesh> no reason, i was setting the mimetype on the server-side
<alexmacy> i want to capture the "back" event (when the user click on the navigator back button, or when backspace is hi)
<alexmacy> anyone know how i can do this?
<Craig_> yes but the header tells the server and the browser its a form post/get
<Craig_> hence the url encoded
<rajesh> Craig_: k, i added that line in.
<Craig_> alexmacy, you can detect the backspace button being hit, but its not possible to detect the back button being clicked, sorry.
<alexmacy> rf
<alexmacy> there is absolutely no way to capture it?
<rajesh> Craig_: http://pastebin.com/397534
<Craig_> nope.
<Lebannen> I've got an xmlhttprequest question... a client with AOL (based on IE) experiences regular crashes which I've managed to replicate; it appears if a number of xmlhttprequests are triggered in quick succession, the program crashes. Is there a way to safely dispose of xmlhttprequest objects so that callbacks etc won't trigger (as I presume that's the problem)? Or does anyone know what else might cause this?
<alexmacy> raah
<alexmacy> this is a problem lol
<Craig_> return d.addCallback(test); ?
<Craig_> Lebannen, Umm callbacks?
<rajesh> Craig_: umm, i thought that d.addCallBack(test) would call test() with the xml response as a parameter? no?
<Craig_> well no, because your not actually returning the xml response, your returning the req object
<Lebannen> Craig_: Well I've no idea what's causing the crashes, but I'd guess that it's something to do with a readystatechange callback from a previous request somehow conflicting with a newer xmlhttprequest
<Craig_> rajesh, the xml document is p***ed within the object as req.respon***ML
<Craig_> Lebannen, got an example of your code I can work from?
<rajesh> Craig_: right. so, wouldn't the block inside the alert() pick out the <name> tags in the respon***ML that was p***ed as resp to test() ?
<Craig_> no.
<Craig_> use this instead
<Craig_> alert(resp.respon***ML.getElementsByTagName('name').length);
<Lebannen> Craig_: it's very standard... create an object, add a onreadystatechange function, do stuff in that function. zero errors thrown, works fine in ie and firefox, etc... but it's triggered on a onchange event on a textbox, and if five or six events trigger in rapid succession, AOL crashes.
<Craig_> Lebannen, are you making sure to check the states id?
<Lebannen> any processing happens after a (xmlReq.readyState == 4) check - should I do more?
<Craig_> for thers four different states, if you dont add a conditional brace to detect if its "4" then it wil lfire 4
<Craig_> ah, so you ahve :)
<Craig_> maybe try this too
<Lebannen> and a (xmlReq.status == 200) to boot
<rajesh> Craig_: thanks, but no go.
<Craig_> ah, beat me to it :)
<Craig_> Lebannen, no reason why it should be firing so many times, try alerting everytime it is fired, that way you will know if its being fired mroe than once
<Craig_> rajesh, hmm
<Craig_> what does addCallback to?
<rajesh> ?
<rajesh> what does it do?
<Craig_> return d.addCallback(test);
<rajesh> or what does it point to?
<Craig_> what does it do, sorry, I've never had a use for it so never known what it does.
<Lebannen> Craig_: it only fires once per change, I've checked... but say I quickly type "blabla" into the textbox, that sends six change events... I could add a delay/limit, but that would take away from the "instant response" feel normally - and it seems weird for anything to cause an application crash :(
<Craig_> application crashes are normally caused by an inifite loop or multiple functions being fired. maybe its just your browser version?
<Craig_> try getting another person with AOL to test it for you and see if it has the same effect
<rajesh> Craig_: ahh ok. well, if im not mistaken, once the response is received, it triggers the method.
<rajesh> i just wasn't sure how mochikit does it.
<Craig_> right but, is the response p***ed as a paramter, or must you define it?
<rajesh> which is what i don't know.
<Lebannen> yeah, I set up an AOL issue and replicated it, so it's not just one machine... but it may well be just that version, namely AOL 9.0. Ah well, I guess there's no easy fix then :)
<rajesh> prototype p***es it as a paramter to the callback function. mochikit, i don't know.
<rajesh> am looking
<Craig_> try alerting resp.
<Craig_> that way you know if the parameter is defined
<rajesh> alert(resp) only tells me that there is an object of type mxmlhttp...
<rajesh> so, it is defined.
<Craig_> go up in steps.
<Craig_> then alert resp.respon***ML
<rajesh> hmm, it's a null object.
<Craig_> try responseText
<rajesh> i did
<rajesh> :0
<rajesh> :)
<rajesh> interesting, i get back stuff
<rajesh> ok, something to build on now.
<rajesh> thanks a ton, Craig_
<Craig_> Is the document your querying an XML ?
<rajesh> it is.
<Craig_> because if not, respon***ML won't return anything
<Craig_> it says .html
<rajesh> oh, that's alright, i'm redirecting internally.
<Craig_> does responseText get the right document?
<rajesh> ahh, i know the problem.
<rajesh> it's in how the paramis being p***ed.
<rajesh> if i do 'url?lap', it works alright
<Craig_> ah :)
<rajesh> strange
<Craig_> simple things are the ones that get you :P
<rajesh> :)
<rajesh> and yes, in mochikit, you have to the .respon***ML.
<Craig_> ya, thats a xmlhttprequest object, so its in all scripts
<rajesh> right.
<rajesh> thought it would be shortcutted in mochi
<Craig_> what is mochi anywho?
<rajesh> mochikit.com
<Craig_> ah :)
<steve_> I have a div and when I try to do a div.innerHTML = ""; I get an exception with a failure code from firefox
<steve_> the failure code is 0x80004005
<steve_> any idea as to what that might be?
<steve_> It worked a little while ago...
<rajesh> steve_: maybe you don't want to use innerHTML
<TheXL> Can you serve me with the whole line?
<rajesh> i've read in many, many places that it shouldn't be used other than for IE
<steve_> rajesh: what should I use to replace the content of that div with other content?
<TheXL> Hmm, works in Gecko most of the time in my experience
<rajesh> if you're using innerHTML, i think it also means you're usin responseText
<frb-away> it's always worked for me
<rajesh> i just create the element, and append to it when i need it.
<steve_> alert(div.innerHTML); shows me correctly the current contents of the div
<rajesh> steve_: innerHTML failure code 0x80004005
<rajesh> google that
<rajesh> i'm sure it works for most people
<rajesh> but i wouldn't do it
<frb> then how would you set the contents of a div?
<rajesh> but then, i just started working with javascript recently
<frb> don't just say to not do something without an alternative
<rajesh> frb: document.getElementById(divid).appendChild(whatever);
<rajesh> ?
<frb> and what is the child if you want to just create text?
<frb> or do you propose that if I wanted text with 4 images, I would append each of them separately?
<rajesh> ahh, interesting. that would fail if it's text?
<rajesh> why would it?
<frb> well, what kind of child is text?
<steve_> I have a whole bunch of xhtml to _replace_ the current code. So I also need to delete the old content...
<rajesh> does innerHTML replace? or append?
<rajesh> replace.
<rajesh> ok.
<Burninator> but you shouldn't use it if you can help it, rajesh
<Burninator> use DOM!
<rajesh> Burninator: that' what i've been suggesting.
<rajesh> i was just wondering ...
<rajesh> that's all. :)
<Burninator> k
<BarnacleBob> in firefox can i trap someone hitting the back button or backspace (navigate backwards in the history). i am using an AJAX style framwork for navigation and that is just detrimental as it kicks them out entirly
<frb> Burninator: what part of the dom should be used is my question
<Burninator> frb: to do what?
<kosh[-1]> hi, I've got a page where I'm using a script block to find all the tags <textarea> and output them, it finds ONE tag, where there is clearly MORE than one, there is three
<frb> Burninator: replace the content of a node
<kosh[-1]> I've got the page source open and I can see three of them
<Burninator> frb: nodeValue
<kosh[-1]> so any idea why these textarea tags are "hidden"
<frb> Burninator: and I can put HTML in there?
<rajesh> kosh[-1]: getElementsByTagName('textarea');
<kosh[-1]> rajesh: i'm using that to find them
<Burninator> you don't insert html via dom
<Burninator> you insert nodes
<kosh[-1]> except I'm saying "textarea" and not ' '
<Burninator> you have to create each node
<frb> Burninator: so if I wanted to have text with 4 images interpspersed, I'm supposed to add text, image, text, image, text, image, text?
<Burninator> yes.
<frb> that's a pain the ***
<Burninator> it's not that bad if you do it right
<BarnacleBob> kosh[-1], objs=getElementsByTagName('textarea'); for(i=0; i<objs.length; i++){ print objs[i].value}
<frb> and to replace the whole content, I have to kill all 7 nodes, and create 7 new ones
<kosh[-1]> BarnacleBob: yup, again, I'm doing that
<kosh[-1]> i'm preparing a pastebin with the entire page source, might help


Return to javascript
or
Go to some related logs:

osdev
brainstew irc trondheim
math

Copyright © 2005 www.irclogs.ws. All rights reserved. » disclaimer » contact