Weird errors on 1: LoadError: UndefVarError: Base_Docs_4352b6d8 not defined.
They seem unrelated to this PR (same errors occur when running ] test Revise against #master).

Xref:
#505
#520
#561

The error on 1.6 & windows (of course, damn *** :/) looks related to the added test :

   in expression starting at C:\Users\RUNNER~1\AppData\Local\Temp\ML0Fd83oSC\driver.jl:3
caused by: LoadError: "invalid escape sequence"

==> Edit: fixed

Shouldn't all PkgId(mod) be mapped to PkgId(mod, string(mod) (else PkgId(mod) maps to PkgId(mod, "Main")) ?

Base uses moduleroot to get the top-level module. In this case the root really is Main, because you're not loading a package.

I think this change would break for a submodule of Base (or any package submodule):

julia> Revise.track(Base)

julia> id = Base.PkgId(Base.Filesystem)
Base [top-level]

julia> pkgdata = Revise.pkgdatas[id];   # works

julia> id = Base.PkgId(Base.Filesystem, string(Base.Filesystem))
Base.Filesystem [top-level]

julia> pkgdata = Revise.pkgdatas[id]
ERROR: KeyError: key Base.Filesystem [top-level] not found
Stacktrace:
 [1] getindex(h::Dict{Base.PkgId, Revise.PkgData}, key::Base.PkgId)
   @ Base ./dict.jl:498
 [2] top-level scope
   @ REPL[10]:1

Would it be worth first trying the version that's already in the code, and then if haskey(pkgdatas, id) fails, maybe try your version? Perhaps specifically in the case where moduleroot(mod) == Main?

Thanks very much for the test!

In this case the root really is Main, because you're not loading a package.

I see, good point.

Would it be worth first trying the version that's already in the code, and then if haskey(pkgdatas, id) fails, maybe try your version

Thanks for the hint. This is a good idea, I'll try to work something out now.

Tests on 1.6 are ✔, so I think that this PR is ready.

Thanks!