fix(runtime): use bootstrap mechanism to load globals rather than setting in the alloy.js file

[ewanharris]

Fixes ALOY-1734

• alloy.bootstrap.js will be loaded at startup before the app.js, this ensures that the globals are correctly configured before the app.js is even executed.
• Check if the globals exist in the app.js and set them if they don't, this should keep us compatible with SDKs pre-7.5.0 (i.e. those without the bootstrap feature, if you're using an SDK that old ... please, please, please update ASAP)

Test steps - see jira

[build]

	Warnings
⚠️	Please ensure to add a changelog entry for your changes. Edit the CHANGELOG.md file and add your change under the Unreleased items header

	Messages
📖	✅ All tests are passing Nice one! All 3490 tests are passing.

Generated by 🚫 dangerJS against efcabf6

[jquick-axway]

LGTM

Thanks for taking care of this @ewanharris !

[brentonhouse]

@ewanharris @jquick-axway - I know you have already merged this but I've tried using bootstrap files for assigning globals in the past and although it works when running normally, it breaks everything when running in LiveView.

[jquick-axway]

@brentonhouse, I'm not sure how it could fail. When you make a file change, LiveView invokes Titanium's undocumented Ti.App._restart() method, which then reloads our "ti.main.js" script, which in turn loads all of our bootstrap scripts before requiring-in the "app.js". I've tested Android with hyperloop (which now uses a bootstrap script and binding APIs) and it restarts with LiveView fine for me.

But it doesn't hurt to re-test Alloy and LiveView again anyways. It's part of our smoke test anyways.

[brentonhouse]

What happens is there becomes multiple instances of modules. If you set a property on a module in alloy.js and go to read it in index.js, the property is not set as it is a clean instance of the module.

[brentonhouse]

What happens is there becomes multiple instances of modules. If you set a property on a module in alloy.js and go to read it in index.js, the property is not set as it is a clean instance of the module.

[brentonhouse]

We saw this before but I think it was fixed by this #955

[ewanharris]

@brentonhouse I'm unaware of any issues between alloy and the boostrap mechanism, I know we've seen issues similar to what you're describing (TIMOB-27206), but that was specifically down to liveview and it's inability to handle node_modules.

Do you have