Option for embedding Sizzle for older browsers

[jasononeil]

If we do this we might be able to get closer to browser support on older platforms.

I suppose you could make it optional with a compiler switch?

Code could be copied from Franco's DHX library.

[guedressel]

Is there any progress on this issue?
We would really appreciate support for IE8. Which is currently not possible because of the querySelectorAll...

[jasononeil]

No, but it should be a relatively easy fix, it just hasn't been high
priority for me at the moment.

I will look into it, again on the weekend :)

On Thu, Nov 1, 2012 at 12:22 AM, Gü notifications@github.com wrote:

Is there any progress on this issue?
We would really appreciate support for IE8. Which is currently not
possible because of the querySelectorAll...

—
Reply to this email directly or view it on GitHubhttps://github.com//issues/31#issuecomment-9951348.

[guedressel]

hi.
Guess you didn't find any time last weekend?
May I help you, or would it too much of a hazzle for a haxe greenhorn like me?

[skial]

Looking at the dhx library which Jason mentions in the opening issue, you can take the code from Sizzle.hx. You only really need the following haxe.macro.Tools.includeFile("dhx/sizzle.js");, which you can paste into your own __init__ method some place. More about init magic.

This is all just guessing, but you will of course need to change the path to point to your sizzle.js file 😄

[jasononeil]

Hi sorry,

Yep the weekend turned out to be too busy, sorry :) I'll be doing some
work on it tomorrow. If you want anything before you could try it
yourself.

I'm still undecided about the approach I want to take:

• Fall back to jQuery (which there's a high chance many people will use
  anyway, but is heavy)
• Fall back to Sizzle
  • Built in to detox?
  • As a standalone library? We could then use #if sizzle in our
    conditional compilation
  • Include the file in the haxe source, or leave it to the developer
    to include sizzle.js, or jquery.js, and we just reference it.

If you guys have any input on this, I'd appreciate it.

Whichever path we choose, Franco already has some Sizzle externs in his dhx
library:
https://github.com/fponticelli/dhx/blob/master/src/dhx/SizzleEngine.hx so
we can just use those.

I'd appreciate your feedback, and I'll try get on to this soon. Sorry
again for the delay :)

Jason

On Mon, Nov 5, 2012 at 6:53 PM, Skial Bainn notifications@github.comwrote:

Looking at the dhx libraryhttps://github.com/fponticelli/dhx/tree/master/src/dhxwhich Jason mentions in the opening issue, you can take the code from
Sizzle.hxhttps://github.com/fponticelli/dhx/blob/master/src/dhx/Sizzle.hx.
You only really need the following
haxe.macro.Tools.includeFile("dhx/sizzle.js");, which you can paste into
your own init method some place. More about init magichttp://haxe.org/doc/advanced/magic#initialization-magic
.

This is all just guessing, but you will of course need to change the path
to point to your sizzle.js file [image: 😄]

—
Reply to this email directly or view it on GitHubhttps://github.com//issues/31#issuecomment-10066501.

[skial]

I think having a haxe version of sizzle would be best, as a standalone library. But the quickest option is to just borrow/link to @fponticelli dhx externs and embed sizzle.js into the output.

Then if a haxe version of sizzle is still wanted, it can be done slowly.

[jasononeil]

@skial if we wanted to a haxe written selector engine, there is the 'selecthx' javascript library - but I'm not sure how well supported it is across browsers.

For now I've been able to set up jQuery and Sizzle to each be imported and theoretically used, but I'm still having trouble getting even a basic demo passing in the Internet Explorer browsers. I mostly run Linux (and have my wife's Mac to test on as well), but only have Win7 in a virtual machine and it's just a pain in the ass to do proper testing / debugging in it. I'll keep trying to figure it out though.

It may be that there are a bunch of problems other than the selector engine, especially for IE6 or 7. With these I might try and fall back on jQuery as often as possible. Once I get a basic demo working, I can run the unit tests and see what works and what doesn't, and get to work fixing it up.

Thanks for your help / advice on this,

Jason

[skial]

@jasononeil I forgot that selecthx is used. Before I was talking of using sizzle as a shim, but I will have to see how selecthx works, and look closer at sizzle and see if I can port bits over. Have no idea when though.

As for IE6/7 testing I can try and setup a couple of vms to test on. Getting multiple versions of IE on Windows 7 doesn't work well, its either vms or a service like browserstack. I assume it's only the munit tests that need to run and nothing else?

[skial]

I've had a go at getting a local vm setup with IE8, real pain to do. Anyway running the unit tests in IE8 gives a wonderfully pointless error of something similar to cannot access undefined field or method 😩 .

It points to dtx.single.DOMManipulation::append method, then to parent.appendChild(child).

I've tested to see if parent has appendChild which returns true. My guess is the tests are running before IE has completed loading/building the dom. IE6/7 just hang at the same spot, but doesn't report anything. Also IE8 doesn't always report an error, it just hangs like IE6/7.

I'll try and create some simple tests running which don't rely on munit, just bare haxejs/detox and see if that gets any more useful info from IE.