2016-11-05 ~ Sample Many Iframes R1

[theo-armour]

Sample Many Iframes R1

@paulmasson

Still not a separate file.

Tested on Chrome only.

The code and the variable names are beginning to get easier to read. Performance seems OK on Chrome.

sample-many-iframes-r1.html ~ 266 iframes and texts ~ source

[paulmasson]

@theo-armour this is exactly what I wanted to see! Control over an arbitrary number of iframes and their sources in 30 lines of code. This is looking great!

Questions/ideas:

1. When this is a stand-alone file, will it need to be placed at the end of a web page or can the process be reliably initiated from document.onload? What if there is already a document.onload defined? Is there another trigger that will not conflict?

2. Jupyter notebooks process a custom.js script described here. Adding it for future reference.

3. It appears trival to add srcdoc alongside src for all iframes. If only one is ever defined, can't we set both at the same time?

[theo-armour]

@paulmasson

Control over an arbitrary number of iframes and their sources in 30 lines of code.

Can you think of variable and function names that would be more intuitive?

1. When this is a stand-alone file...

Can you mock up or outline a worst case scenario/ test file(s)? with mixed src and srddoc and onload and whatever?

No need for 266 iframes, just a few will do.

2. Jupyter notebooks process a custom.js script described here.

Noted. Thanks for the heads-up.

3. It appears trivial to add srcdoc alongside src for all iframes.

See 1,

---

Jupyter Reader

My intent is to build an easy-peasy entry-level pure client-side reader of our own. Why should they have all the fun? Yes, there can be chats with the server as and when needed.

[paulmasson]

@theo-armour variables and function names look fine to me.

It's easy for you to mock up what you need. Just copy the source for every other iframe into srcdoc after first escaping double quotes to explicit HTML " entities and remove the src reference. As for the onload I don't know what to expect right now, just trying to think ahead.

When this is completed I anticipate copying the stand-alone file over to the SageMath repository for linking to Jupyter and native Sage notebooks, so add an author reference and license at some point.

[theo-armour]

@paulmasson

My intent is to build an easy-peasy entry-level pure client-side reader of our own. Why should they have > all the fun? Yes, there can be chats with the server as and when needed.

Not worth doing. GitHub do well enough:

https://github.com/blog/1995-github-jupyter-notebooks-3
http://blog.jupyter.org/2015/05/07/rendering-notebooks-on-github/
https://github.com/ipython/ipython/wiki/A-gallery-of-interesting-IPython-Notebooks

Should probably go back to thinking about math+shaders and 100 line physics demos

[paulmasson]

Not worth doing. GitHub do well enough:

@theo-armour all those noteboks are static, not live...

[theo-armour]

People have different abilities and aspirations.

Some people are good at breathing life into historical things

Some people do crazy new shit

Which way do you think my butt faces?

On Mon, Nov 7, 2016 at 5:50 PM, Paul Masson notifications@github.com
wrote:

Not worth doing. GitHub do well enough:

@theo-armour https://github.com/theo-armour all those noteboks are
static, not live...

—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
#2 (comment),
or mute the thread
https://github.com/notifications/unsubscribe-auth/AAhbKt7xuufGp9hZiCUjkE5wVvPmepjIks5q79VZgaJpZM4Kqemh
.

[theo-armour]

@paulmasson

sample-many-iframes-r3

JavaScript file is external.

The goal is to have one set up script tags - but we are not there yest.

Go to bottom of sample file HTML to see the second set of tags and the init. ;-(

I have tried a number of ways to invoke the JavaScript just at the time the HTML but before the renderers start kicking in with only a single set script tags but to no avail so far.

I'm sure there some sort of fancy-pants way of doing this - loading and invoking immediately but I've not come across it yet. It's proably going to be something with adding parens at the end of a JavaScript function.

Feel free to make suggestions - suggest where to next with this. I picked MIF for the namespace but it could be anything?

[paulmasson]

@theo-armour the point of the separate namespace is to avoid conflicts with existing JS in the HTML, right?

[theo-armour]

@paulmasson

yes, and anything that might bubble up from the iframes

[paulmasson]

@theo-armour I was going to make some edits but I don't have access..

Have you tried adding the line

 document.addEventListener( 'load', MIF.init );

to multiple-iframes.js?

[theo-armour]

@paulmasson

i tried

 document.addEventListener( 'load', MIF.init, false );

[paulmasson]

Curious: I would have thought you could set document.onload externally, but I can't get it to fire.

Adding it to window does fire, so how about something like

window.addEventListener( 'load', function() { MIF.init() } );

This will wait for things to load of course, which might be a problem for the test file but shouldn't be for the intended application. The extra function() appears to be necessary for this to work.

[theo-armour]

@paulmasson

I also tried window.addEventListener. IIRC it loaded too late and the renderers were already being invoked.

I must have tried a dozen or more permutations. All quickly though. So sometimes the good answer is overlooked.

In any case, the lessons learned include:

The script must be loaded before the rendering starts
The script must be invoked after the HTML is loaded

[theo-armour]

@paulmasson

I will look at the init inside function tomorrow.

This might provoke immediate execution. Also could try this with document.addEve...

And there's the whole thing of :

thing = function(){}();