Allow using LongPackageName as LPN and change export conventions.

[LilithHafner]

Before:

julia> using Statistics as Stat
ERROR: ParseError:
# Error @ REPL[1]:1:6
using Statistics as Stat
#    └─────────────────┘ ── `using` with `as` renaming requires a `:` and context module
Stacktrace:
 [1] top-level scope
   @ none:1

After:

julia> using Statistics as Stat

julia> Statistics
ERROR: UndefVarError: `Statistics` not defined in `Main`
Suggestion: check for spelling errors or missing imports.
Hint: a global variable of this name also exists in Statistics.

julia> Stat
Statistics

julia> mean
mean (generic function with 6 methods)

This PR is implemented by

• No longer automatically export module names
• Explicitly loading module names on using (i.e. creates an explicit imported const binding)
• On using Mod as M, add Mod to the list of "used" modules and create an explicit imported binding M => Mod.

(non)breaking change 1

• Base.isexported(Statistics, :Statistics) will now return false (in Julia <=1.10 Base.isexported was not part of the public API)
• Base.ispublic(Statistics, :Statistics) will now return false (in Julia <=1.10 Base.public was not defined)

Breaking change 2

names(Mod) currently no longer includes :Mod, I'd like to stop implicitly marking the module itself as public because I don't think Statistics.Statistics is something that should be public by default. Specifically, I think that in the package M1,

module M1
    public API
    const API = M2
    module M2
        public f
        f() = 0
    end
end

The code M1.API.M2.f() depends on internals (the name of the module M2, which was never marked public)

Breaking change 3

The code

julia> module M1
           module M2
           end
       end
Main.M1

julia> module M2
           using ..M1.M2
       end
Main.M2

julia> M2.M2 === M1.M2
false

julia> M2.M2 === M2
true

now throws on

julia> module M1
           module M2
           end
       end
Main.M1

julia> module M2
           using ..M1.M2
       end
ERROR: invalid redefinition of constant M2.M2

because the using statement tries and fails to create an explicit binding. This is a code pattern that is currently in use in Pkg.jl, but I do think it's a bit of an antipattern, and this PR creates a great workaround. However, maybe we should downgrade that into a warning.

Implements #52784 (resolves #52784)
Fixes #52817
Is a superset of and therefore closes #52810 (NFC)

Depends on JuliaLang/JuliaSyntax.jl#406
Depends on JuliaLang/Pkg.jl#3750

[LilithHafner]

I'm open to more test cases, this is surprisingly finicky.

[LilithHafner]

Dropping this link here as a TODO once we update JuliaSyntax and unblock this PR: review this discord thread and consider if/how the changes here impact that situation