Comments on: The dojo parser.

By: dante

dante — Sun, 19 Apr 2009 13:15:20 +0000

@ricardochois – You might want to read this newer article: http://dojocampus.org/content/2009/04/15/declarative-vs-programatic/ … most of the work being done in your paste above is unnecessary, especially with the base Dojo API’s already available to you.

By: ricardochois

ricardochois — Sat, 18 Apr 2009 17:48:02 +0000

Hi I’m new with dojo.

I want to dinamically create a widget, so i create an HTML element (TextBox) and then try to parse it, but it just looks like an basic html textbox and not like a dojoType textBox.

Here is the function i’m using, what is wrong???

function agregarFilaAsignatura(argumentos){
totalAgregadas++;
tabla = document.getElementById(”tablaDatos”);
var elmTR;
var elmTD;
var elmText;
elmTR = tabla.insertRow(fila+1);

actualizarIdsFila();
elmTR.id = fila+1;

for (var i=0; i<5; i++) {

elmTD = elmTR.insertCell(i);

elmText = document.createTextNode(argumentos.rbAsignatura.split(”-”)[1]);

elmInput = document.createElement(’input’);
elmInput.type = ‘hidden’;
elmInput.value = argumentos.rbAsignatura.split(”-”)[0];
elmInput.id = “asignaturas” + idDocente;
elmInput.name = “asignaturas” + idDocente;

elmInputHoras = document.createElement(’input’);
elmInputHoras.type = ‘text’;
elmInputHoras.dojoType = ‘dijit.form.TextBox’;
elmInputHoras.style.width = ‘40px’;
elmInputHoras.value = argumentos.rbAsignatura.split(”-”)[2];
elmInputHoras.id = “horasAsignaturas” + idDocente;
elmInputHoras.name = “horasAsignaturas” + idDocente;

dojo.parser.parse(dojo.byId(”horasAsignaturas” + idDocente));

if(i==1) elmTD.appendChild(elmInput);
if(i==3) elmTD.appendChild(elmText);
if(i==4) elmTD.appendChild(elmInputHoras);

}

By: ribob2

ribob2 — Mon, 13 Apr 2009 22:10:36 +0000

Dojo object and element creation timing? When do things get parsed and created with parseOnLoad? I did this:

lpThumb = new dijit.layout.ContentPane();
lpThumb.domNode.id = “leftThumbtray”;
lpThumb.attr(”title”, “Thumb”);
lp.addChild(lpThumb);

(where lp is an existing TabContainer) There are two problems: 1) at the time postCreate is called on lbThumb, it seems that the widget domNode has not be inserted into the DOM tree (ie createElement was done, but it wasn’t appendChild()ed into the tree) 2) it seems that immediately after the lp.addChild(lpThumb) executes, the lpThumb.addChild is not defined in that object

I’d like to understand what is going on here, or more generally, what is the flow of operations and timing when things are created (so you can know when you can start using things that are being created.

By: ribob2

ribob2 — Mon, 13 Apr 2009 21:56:57 +0000

Passing atttributes resolved? I think I figured it out: Dojo may replace the original you create with its own, losing all attributes (except those it creates) It only looks for attributes matching the names of the ‘instance variables’ defined in widget declaration. So you can pass things as long as they are named as instance variables.

By: ribob2

ribob2 — Mon, 13 Apr 2009 20:16:31 +0000

Passing information from the original HTML to the widget methods ?? I want to allow the original HTML that specifies a widget to be able to pass information to widget methods or be saved in instance variables. I tried adding some attributes of my own to the initial div, but by the time is is parsed and the widget is loaded, those attributes have disappeared. How do I pass info from the HTML to the widget object? Also, is there more documentation on the details of the parsing process?

var catWidget = dijit.byId(”cat”);
var catDiv = catWidget.domNode;
var species = catDiv.getAttribute(”species”); // Doesn’t work, no species attr in the div

By: gtducati

gtducati — Fri, 18 Apr 2008 17:59:32 +0000

Ah … good, so I wasn’t going crazy Thanks!

By: dante

dante — Fri, 18 Apr 2008 17:57:38 +0000

Actually, it looks like there is no node id=”container” in the themeTester anymore … which means dojo.byId(”container”) is returning null || undefined, which means dojo.parser.parse() is being called without any arguments, thus parsing the entire body. Sorry for the confusion. I swear it did have and id=”container” at one point

By: gtducati

gtducati — Fri, 18 Apr 2008 17:48:55 +0000

Hey dante, thanks for the quick response. To clarify, when I said everything, I meant everything in the document body.

In any case, I was really just having problems with the statement “It scans the dom (or a portion of it) turning nodes with dojoType=”" attributes into declared classes”, I was confused about the use of dojoType=”" in said quote. I was assuming that the statement implied the parser was looking for ‘empty’ dojoType attributes, since the phrase dojoType=”" is used. In fact, all it’s really saying is that it scans the body for all nodes with a dojoType attribute with a qualified.class.name for rendering. So, yes, you did answer my question.

I do have another question; this one is in regards to the themeTester and its use of dojo.parser.parse(’container’) in the current 1.1.0 release. What is this exactly looking for? How is dojo.parser.parse(’container’) rendering the entire themeTester? Does ‘container’ imply the document body too? Am I missing something here? Thanks in advance.

By: dante

dante — Fri, 18 Apr 2008 15:15:16 +0000

dojo.body() is everything? You can’t have markup outside of body, so when you call dojo.parser.parse() or dojo.parser.parse(dojo.body()) it’s identical. or perhaps I’m missing your question.

By: gtducati

gtducati — Fri, 18 Apr 2008 14:58:32 +0000

I really like this cookie; however, there are two conflicting statements made concerning the call: dojo.parser.parse();

You first say that the document is scanned for all types of dojoType=”". However, in the Dojo-Preloader example their is a comment that says “// run the parser. defaults to dojo.body()”, which indicates that the call of “dojo.parser.parse();” is essentially scanning everything for rendering. Which is it?

I’ve tested this myself, and it seems that the latter is true.