refactor: don't set readonly by default

[abonander]

Summary

• Don't set readonly by default to avoid conflicts with existing user settings.
  • It can still be set/overridden using Client::with_option() or Query::with_option().

closes #343
fixes #373
closes #361 (superceded PR)

Checklist

Delete items not relevant to your PR:

• Unit and integration tests covering the common scenarios were added
  • Examples
  • Integration test (waiting on bikeshedding)
• A human-readable description of the changes was provided so that we can include it in CHANGELOG later
• Questions:
  • I chose the method name with_read_only to match the ReadOnly enum even though the option name is readonly. If it was with_readonly() then for consistency the enum should be named Readonly but that doesn't look right. Does this make sense?
  • Is with_read_only / the ReadOnly enum overkill? I imagine most of the time the reason to set readonly is to override what we set by default, which I'm deleting anyway as it's causing a regression (v0.14.1 is incompatible with settings profiles setting readonly=2 #373). Otherwise it's always possible to set through with_option/set_option.
  • Does it make sense that with_read_only(false) unsets the option rather than sending readonly="0"? Practically speaking, there isn't really a use for sending readonly="0" because it's either redundant (the user/session already has read-write permissions) or not allowed (the user/session is already read-only).

[abonander]

cc @svix-jbrown

[codecov]

Codecov Report

❌ Patch coverage is 25.00000% with 3 lines in your changes missing coverage. Please review.

Files with missing lines	Patch %	Lines
src/query.rs	25.00%	0 Missing and 3 partials ⚠️

📢 Thoughts on this report? Let us know!

[loyd]

Is with_read_only / the ReadOnly enum overkill? I imagine most of the time the reason to set readonly is to override what we set by default, which I'm deleting anyway as it's causing a regression (#373). Otherwise it's always possible to set through with_option/set_option.

Probably it's overkill, creates other questions about the API around this method and requires more code to satisfy code coverage. While there are cases when someone wants to lower their rights, it's uncommon.

[abonander]

Deleted the with_read_only setters and just kept the change to the default.

[abonander]

I want to say we should ignore the codecov failure this time. There's no way deleting code made the diff coverage go down.