Move path management out of rebar_utils manual code path function
handling (which we leave there for backwards compat), and centralize
them to allow easier coordination of paths between plugins and deps.

On top of path handling, do a check of loaded modules to only purge and
reload those that actually need it done in order to prevent all kinds of
weird interaction and accidental purge kills. It also allows the
possible cohabitation of both at once, with a "in case of conflict pick
X" as a policy

Changing path handling in providers also highlighted a bunch of bugs in
some tests and appears to fix some in other providers, specifically
around plugins.

This replaces all deprecated function usage by alternative ones based on
a version switch enacted at compile time, preventing all warnings.

This will likely introduce some possible runtime errors in using a
Rebar3 compiled on OTP-20 or OTP-21 back in versions 19 and earlier, but
we can't really work around that.

A bunch of dependencies have been updated to support OTP-21 without
warnings as well.

This is done through 3 main change groups:

- replacing `~s` by `~ts` in format strings, so that strings that
  contain unicode are properly printed rather than crashing
- adding the `unicode` argument to all function of the `re` module to
  ensure transformations on strings containing unicode data are valid
  instead of crashing (see issue #1302)
- replacing `ec_cnv:to_binary/1` and `ec_cnv:to_list/1` with matching
  functions in `rebar_utils`.

The last point has been done, rather than modifying and updating erlware
commons, because binary and list conversions can be a contentious
subject. For example, if what is being handled is actually bytes from a
given binary stream, then forcing a byte-oriented interpretation of the
data can corrupt it. As such, it does not appear safe to modify erlware
commons' conversion functions since it may not be safe for all its
users.

Instead, rebar3 reimplements a subset of them (only converting
atoms and chardata, ignoring numbers) with the explicit purpose of
handling unicode string data.

Tests were left as unchanged as possible. This may impact the ability to
run rebar3's own suites in a unicode path, but respects a principle of
least change for such a large patch.

Fixes issue #880

May break backwards compat with projects that manually called the
dialyzer formatter, but we never documented or expected this to be
exposed.

* exclude_apps - never use applications for PLT/analysis
* base_plt_mods - add modules to base PLT (overrules exclude_apps)
* plt_extra_mods - add modules to PLT (overrules exclude_apps)
* exclude_mods - never use modules for PLT/analysis (overrules all)

* All configuration is moved to `{dialyzer, PropList}`
* `dialyzer_warnings` renamed `warnings`
* `dialyzer_plt_warnings` renamed `get_warnings`
* `dialyzer_plt_apps` renamed `plt_extra_apps`
* `dialyzer_plt_dir` renamed `plt_location`
* `{plt_location, local}` uses profile base directory
* `dialyzer_plt` removed
* `dialyzer_plt_prefix` sets prefix of PLT name (default "rebar3")
* `dialyzer_base_plt_apps` renamed `base_plt_apps`
* `dialyzer_base_plt_dir` renamed `base_plt_location`
* `{base_plt_location, global}` uses rebar3's global cache directory
* `dialyzer_base_plt` removed
* `base_plt_prefix` sets prefix of PLT name (default  "rebar3")

PLT files are always named `<prefix>_<otp_version>_plt` to match rebar.
`plt_prefix` and `base_plt_prefix` are not available in rebar, where the
prefix is always set to the project's top level application.

* Only include direct dependencies + base plt
* Error on missing direct dependency
* Always use beams in ebin instead of reading .app file (copying rebar2)
* Remove fake unknown warnings (will clash with unknown warning in 18.0)
* Remove redundant search for .app file (code path includes deps)

- Crashes in providers lib when no providers in a namespace are bare
- Making sure bareness matches semantics; i.e. a bare provider is
  visible, a non-bare provider is hidden.

Also expose it in rebar_api

Previously dialyzer stored the project PLT in ./build/, which meant the
same PLT was used for all profiles. This could mean partial rebuilding
of a PLT when switching between profiles.

It is not convenient for everyone to handle unmatched_returns and
underspecs warnings. Extra default warnings are removed to match
dialyzer's behaviour.

Previously dialyzer would crash when not doing success typing analysis.

When dialyzer_plt_warnings is false no warnings should be returned when
updating a PLT. Previously any callgraph warnings would be returned when
updating a PLT regardless of the get_warnings option.