Fix crashes when enabledBreadcrumbTypes is null

[imjoehaines]

Goal

Following on from #1466, this PR fixes crashes in various plugins that assumed enabledBreadcrumbTypes must always be an array. They now allow for the case where it may be null and handle it accordingly

Testing

Existing tests & new unit tests. Manually tested in browser, electron and react native projects

[github-actions]

@bugsnag/browser bundle size diff

	Minified	Minfied + Gzipped
Before	41.31 kB	12.72 kB
After	41.33 kB	12.72 kB
±	⚠️ +12 bytes	⚠️ +6 bytes

code coverage diff

Ok	File	Lines	Branches	Functions	Statements
✅	/home/runner/work/bugsnag-js/bugsnag-js/packages/core/client.js	99.34% (+0%)	97.62% (+0.06%)	100% (+0%)	98.8% (+0.01%)
🔴	/home/runner/work/bugsnag-js/bugsnag-js/packages/electron/src/client/renderer.js	34.38% (-1.1%)	0% (+0%)	0% (+0%)	30.56% (-0.87%)
✅	/home/runner/work/bugsnag-js/bugsnag-js/packages/plugin-electron-ipc/bugsnag-ipc-main.js	82.54% (+0%)	75% (+1.92%)	89.47% (+0%)	81.43% (+0%)
✅	/home/runner/work/bugsnag-js/bugsnag-js/packages/plugin-electron-net-breadcrumbs/net-breadcrumbs.js	100% (+0%)	90% (+2.5%)	100% (+0%)	100% (+0%)
🔴	/home/runner/work/bugsnag-js/bugsnag-js/packages/react-native/src/notifier.js	72.15% (+0.36%)	61.76% (-0.74%)	61.54% (+0%)	71.26% (+0.33%)

Total:

Lines	Branches	Functions	Statements
82.26%(-0.01%)	71.55%(+0.07%)	83.47%(+0%)	81.2%(-0.01%)

Generated by 🚫 dangerJS against 341d6c7

[kattrali]

Is it possible to either:

• Coerce enabledBreadcrumbTypes to always be an array during config validation (where null -> all)? or
• Add a helper on config to handle the null check? Most of the validation checks now are of the form if (config.breadcrumbTypes !== null && config.breadcrumbTypes.contain(type)) where maybe it could be something like if (config.isBreadcrumbTypeEnabled(type))

Trying to think of a way to prevent this kind of issue going forward since every future plugin/consumer of this config option basically needs to have this check. Stronger type validation on our part? Additional linting?

[imjoehaines]

Trying to think of a way to prevent this kind of issue going forward since every future plugin/consumer of this config option basically needs to have this check. Stronger type validation on our part? Additional linting?

I avoided doing anything extra in core to try to avoid increasing the browser bundle size, but it turns out that a helper method actually saves space because it removes duplication from plugins

I've opened #1468 for this as it affects files in this PR, #1466 and files that weren't touched by either PR