Consider removing character snapshot tests?

[matthewmayer]

These were originally added in #3276

Although this was my idea originally, I would say in the time its been active it has not been very effective - it has caught zero actual bugs (people accidentally adding incorrect characters) and it has some negative side-effects:

• it trips up a lot of newbies as it means even trivial locale changes often need a full re-run of preflight
• there can be conflicts around this file

[Shinigami92]

IMO these are helpful for us maintainers/reviewers so we can see directly which characters are used in a language.
This is more important for languages in asian countries.
Then you can see if e.g. accidentally some non asian characters slipped in. Or the other way around, that asian characters slipped into non asian locales.