chore: Rework SDK initialisation

[benjackwhite]

Changes

TAKE 2! Had a PR a while ago for this but it got stalled by tests. Since then the test suite improved so now they work!

We get a lot of issues with people initialising PostHog incorrectly, sometimes due to think an env var is set when it isn't.
I started trying to add this and the spaghetti code of initialisation caused such a headache that I felt it was time we improved this.

Big change so we should validate this heavily but this:

• Removes the multiple states that create a lib, instead using PostHog class as the basis and moving all initialisation into the two methods
• Simplifies the global state side of things
• Still keeps the snippet specific stuff but now it is relatively straightforward acting on the initialised PostHog instance rather than a hacked global override

TODO

• Check the multiple paths
• Check we have tests for these paths and they all pass
• Ensure pre-loaded calls still work

Checklist

• Tests for new code (see advice on the tests we use)
• Accounted for the impact of any changes across different browsers

[posthog-bot]

Hey @benjackwhite! 👋
This pull request seems to contain no description. Please add useful context, rationale, and/or any other information that will help make sense of this change now and in the distant Mars-based future.

[github-actions]

Size Change: -3.85 kB (0%)

Total Size: 858 kB

Filename	Size	Change
dist/array.full.js	184 kB	-1 kB (-1%)
dist/array.js	126 kB	-991 B (-1%)
dist/es.js	126 kB	-926 B (-1%)
dist/module.js	126 kB	-926 B (-1%)

ℹ️ View Unchanged

Filename	Size
dist/exception-autocapture.js	12 kB
dist/recorder-v2.js	106 kB
dist/recorder.js	58.6 kB
dist/surveys-module-previews.js	62 kB
dist/surveys.js	58.3 kB

_{compressed-size-action}

[mariusandra]

Makes sense to me!

I hope nobody was accidentally relying on the undocumented old initialization path. Seems low risk.

[posthog-bot]

This PR hasn't seen activity in a week! Should it be merged, closed, or further worked on? If you want to keep it open, post a comment or remove the stale label – otherwise this will be closed in another week.

[pauldambra]

LGTM

runs locally, is smaller when bundled 👨‍🍳 👌

[posthog-bot]

This PR hasn't seen activity in a week! Should it be merged, closed, or further worked on? If you want to keep it open, post a comment or remove the stale label – otherwise this will be closed in another week.

[robbie-c]

I was debugging https://posthoghelp.zendesk.com/agent/tickets/11815 and I believe this PR is the most likely root cause. The global variable posthog isn't fully initialized when the loaded callback is called. The fix was pretty small.

Might be worth delaying the loaded callback so this fix isn't necessary.