fix: cache subdomain discovery

[pauldambra]

We use setting cookies to discover the first non-public domain so that we can identify what domain to set a cookie on.

E.g. ww.example.com.au and foo.example.com.au are both on example.com.au. This is because the length of the public part is variable and the list of public domains is very long.

However Firefox prints a warning when it can't set a cookie because of an invalid domain (maybe it only started recently 🤷). This doesn't help anyone. But it does prompt users to ask us why there are so many warnings logged.

You can't change (sub-)domain without a full page refresh. So, it's safe to store that value in memory.

This should mean that someone only sees at most four logs per page refresh

[github-actions]

Size Change: +144 B (0%)

Total Size: 745 kB

Filename	Size	Change
dist/array.full.js	176 kB	+36 B (0%)
dist/array.js	118 kB	+36 B (0%)
dist/es.js	118 kB	+36 B (0%)
dist/module.js	118 kB	+36 B (0%)

ℹ️ View Unchanged

Filename	Size
dist/exception-autocapture.js	12 kB
dist/recorder-v2.js	104 kB
dist/recorder.js	58.4 kB
dist/surveys.js	41.5 kB

_{compressed-size-action}