Don't export Dates.adjust

[jariji]

adjust(df::Dates.DateFunction, start, step, limit)
adjust(func::Function, start; step, limit)

Dates.adjust has been exported for a while, but has no docstring and is not in the manual in 1.10.

I've asked on Slack, GitHub, and Discourse if anybody is using it, and nobody piped up.

Since we're adding docstrings to everything for 1.11, it seems like a good time to remove the adjust export.

Resolves #52746.

[LilithHafner]

@nanosoldier runtests()

[LilithHafner]

This has never been part of the public API because the public API is defined to include only things in the manual.

[mbauman]

There is some usage in the ecosystem (Quantlib, QDates, OpenSSL, Miletus), but I think only QuantLib relies upon it being exported.

[nanosoldier]

The package evaluation job you requested has completed - possible new issues were detected.
The full report is available.

[LilithHafner]

Nanosoldier revealed two new failures:

OpenSSL extends and reexports adjust. Most of it's usage does not depend on adjust being exported. The only tests that fail are the explicit "does OpenSSL.adjust work" tests, but the tests are failfast so there could be more that fail.

Dates itself fails because it expects the symbol adjust to be public & undocumented but this unbreaks that test (this failure should come up this this PR's CI)

QuantLib did not come up in PkgEval because it is skipped. (no updates since 2020, 3 monthly downloads according to JuliaHub, 0 dependants)

[quinnj]

LGTM assuming we get Dates tests fixed as @LilithHafner mentioned

[jam-khan]

I have added docstrings and made changes to runtests.jl for adjust as part of issue #52725 in PR #52914

[DilumAluthge]

This PR broke the doctest job in CI.