| |
| |
| |
|
Page: 1 2 3 4 5 6 7
Comments:
<0> :)
<1> whoa
<1> i just made windows crash inside of vmware :D
<0> heh
<0> that's always fun
<1> indeed
<2> i'm using a onchange for a drop down meny with a window.location.reload if the value in the drop-down is changed. what I want to do is to p*** the value of the drop down to the reload of the page, is that possible?
<2> or do I have to use location, and specify the new param
<2> er, replace() even
<0> eydaimon: Yes, just use location, and p*** the param via URL
<2> document.location.whatgoeshere?
<0> document.location.href
<0> So I think you can do: document.location = document.ocation.href + "?myNewVar=" + blah.selectedIndex;
<3> document.locate.****.my.***.mr.garrison
<0> Of course you need to check to see if it already has a ? or not, and that's not exactly the right code either
<2> thanks Sembiance
<2> Sembiance: I'll use "&mynewvar" :)
<3> that's not right either
<2> it is if I already have a ?oldvar=
<2> no?
<3> if your href ends with #someplace, you need to add query strings before it
<2> oic
<3> protocol://host/path?query&query2#anchor
<2> i know what you mean. it doesn't, so I will ignore that for now to keep it simple
<2> doesn't seem to be reloading the window with the new values with just document.location though
<4> I have a stupid question, maybe as much of an http question as a javascript question (or general programming)
<4> can one embed an HTTP server in a browser window?
<4> it would be cool if one javascript window, when it's corresponding remote object has changed, for the remote object to propagate it's changes back to the browser by doing an http request of the browser
<5> [2]Andrew you would have to channel the socket through the web server and use flash or java
<4> so like a "java servlet" type of thingy but as an applet?
<5> something like that, i saw an irc client made with js <-> flash <-> server
<4> what do you mean "channel" the socket?
<6> anyoen here particularly knowledgeable about using onbeforeunload with both IE and Firefox?
<2> i see what I did wrong :)
<5> [2]Andrew i thought you meant that other clients should be able to connect to the server
<4> this is what I'm thinking
<4> you have javascipt "proxy" objects which live in the browser and represent "presentations" of database objects on the server
<7> nonix4: what do need to confirm?
<4> lets say you drag and drop one presentation into another
<4> you drag "V-8 Engine" into "mustang" and kickout "V-6 Engine"
<8> can $F("form_element_id") get the current value of an option from a select dropdown?
<4> on the javascript side
<4> the "mustang" then HTTP Requests the server to tell it that it should change engines in the database
<5> [2]Andrew i would use a protocol that allows for a stay-alive connection, such as IRC
<4> but sitll use IE?
<7> [2]Andrew: the browser sending a request to the webserver?
<9> if I do if (...) { var abc="hi" } will the value of abc only be accessible within the scope of the if statement?
<7> airwave: no
<9> why not?
<4> ok...the "presentation mustang" sends signal "change engine v-8" to the server
<7> airwave: its scope will be the whole function
<9> whats the point of var anyways?
<4> but this would make it so that other objects would propagate changes through the OO dbase, since it's actually an active knowledge-base
<4> so that means other "presentations" in the browser would be "stale"
<4> those presentations wouldn't know that they are supposed to update
<7> airwave: it declares the scope of a variable
<4> that's my basic philosophical flaw
<7> airwave: so that it's separate from a variable with the same name outside
<6> okay, lemme rephrase my question :) Using onbeforeunload, is it possible to return something that prevents the browser window from closing?
<6> (something that works in both Firefox and IE)
<7> [2]Andrew: in the same browser?
<10> hi, does someone know if it is possible to create css cl***es on the fly with js ?
<4> yeah, for simplicity, lets say just the same browser window the user always has up
<4> I thought maybe at the same time the "change engine" event fires in javascript, javascript could do a force-reupdate
<4> of all it's objects
<7> yeah, why not update the other views locally?
<4> but that's brute force
<7> how so?
<7> defbyte: CSS doesn't have cl***es
<4> because you only want objects talking to the database if they are changed, not everytime a change happens to any object on the page
<4> what if you have 2 dozen windows open, but only one is actually dependent on "mustang"
<7> then only update that one
<4> how do you know that one?
<4> only the server knows
<7> how did those windows get created?
<4> some kind of javascript event fires (such as onclick) and Javascript pops up a window and populates with the xml data from an http request
<4> there is an object id stored by javascript that is a backpointer to the actual database object
<7> the javascript that opens the window gets a reference to it
<7> it can then set variables and call functions in it
<4> my thinking is that the only way to make javascript know what windows are dependent on what is to implement the knowledge base in javascript
<4> and that is a big waste of time, since everybody needs to get data from a central point
<7> can the dependency information be sent to the script?
<4> maybe
<4> it sounds error prone though
<7> or the script could ask the server
<7> "I have windows showing X, Y and Z; which ones need to be updated?"
<4> yeah, I would have to cut into the source code of the knowledge base to do that
<4> but i could
<4> it would have to happen simultaneously with the actual dependency updating
<10> deltab: when i define in css: .testcl*** { margin:0; padding:0; } ... how to creat this with js without, for example, obj.style.margin ....
<4> so the knowledge base says: attribute a1 of object A depends on attribute b1 of object B, serving back to javascript as it traverses the dependency tree
<7> would it be easier for the server to notify each window separately?
<4> that is where my first question was
<10> deltab: so i create one cl*** and apply that to my 50 divs =)
<4> could I tell my server to do an http-request to a java applet in the browser with a list of updates to the windows there
<7> I suspect it'd be easiest and most efficient to work out the consequences of a change at one time, and send them all at once, but I don't know the architecture of your system
<4> or it would be even cooler if each javascript presentation were its own server as well as client
<7> defbyte: ah, that's not defining the cl***, just giving a rule about what styles should be applied to elements of that cl***
<4> on the server, you have tree-shaped hierarchy of information (but cl***/instance OO, not like flat XML or something)
<4> and then I have written a little window thingy for IE, but I am thinking of going with "prototype windows" which someone here suggested
<7> defbyte: support for doing that in script is poor: http://www.webdevout.net/browser_support_dom.php#dom2style
<4> I could do a server-push from javascript every 2 seconds or so, but that would be clunky
<7> [2]Andrew: I don't think running a server in the browser is going to be too easy, useful though it would be
<7> [2]Andrew: why every two seconds? why not only when there's something to say?
<4> now if I got rid of the knowledge-base entirely and went with a non-dependency tracked OO on the server side i could probably eliminate my problem
<7> a user-side server is going to be limited by sandboxing and firewalls
<4> then, all "fresh" changes live in javascript and javascript just tells the server what to do
<4> KISS theory
<4> isn't that essentially AJAX?
<7> sounds like it
<4> AJAX would be totally screwed if the server changed the state of other objects without javascript's prior knowledge, i think we have determined that
<7> so there can be changes on the server that the client does not cause but needs to know about?
<4> so I guess Javascript always has to have knowledge of the "freshest" state of it's objects, and then when a change happens it HTTP calls the server to tell it what to do with it's own objects
<0> is this some sort of theoretical discussion?
<4> yeah, that is the fundamental problem with using a knowledge base on the server side
<4> everything has to be a page-wide "pull"
<7> [2]Andrew: if your server supports it, you can keep a connection open and the server can send updates through it
<4> <0>: I'm wanting to build a presentation system for a knowledge base which exists on a server
<0> it's code writing time for you!
<4> I know, but that could be error-prone too, a lot of exception handling logic would have to be created
<10> deltab: thank you
<7> defbyte: maybe there's a better way to do what you need
<7> [2]Andrew: what would be?
<4> what would be what?
<7> 9[2]Andrew: I know, but that could be error-prone too, a lot of exception handling logic would have to be created
<7> what is that referring to?
<4> keep-alive connections using XMLHTTPRequest
<7> how would it differ from the server connecting to the clients?
<4> the client initiates
<7> that's just a difference in how the connection is established though
<4> personally, i don't know how to write a server in javascript, though I could probably adapt somebody miniature java server into an applet
<7> I don't think you can create a listening socket in an applet
<7> I know you can't in javascript in a web page
<4> you are probably right :(
<11> writing a server-applet wouldn't be all that useful anyway
<11> firewall/nat problems
<4> right
<7> all you need is a connection, though it doesn't matter for this which end opens it
<7> once a connection is open, the server can send messages through it
<4> i wonder if XMLHTTPRequest supports persistent connections
<4> I would probably have to write a special client
<4> a client that "listens"
<7> http://ajaxpatterns.org/HTTP_Streaming
<12> anyone know how (if it's possible) to manipulate a css pseudo cl*** from js?
<4> looks interesting <7>
<13> mattypython: no its not
<7> mattypython: the DOM Inspector in Firefox apparently does it, but I don't know how
<12> i was afraid of that... is there anyway to manipulate a cl*** without it blowing away it's pseudo cl***es?
<7> "manipulate a cl***"?
<12> change the 'display' for example
Return to
#javascript
or
Go to some related
logs:
RunAsUser perl
you appear to lack the pcap(3) library
#css
#debian
ubuntu lirc_serial
#linux
#osdev
extended away +gaim
install wxpython on ubuntu livecd
ventrilo_srv: permission denied
|
|