The previous iteration of the patch worked somewhat by accident. After
digging in and figuring out why the two dep sources are the way they
are, the patch is now properly working with a well-documented
explanatiion inline.

When the operation for an unlock takes place in the config of a umbrella
application, the `unlock' provider does not see the dependency in the
`deps' value of the config (since it only includes the deps at the root
of the project) and ignores these.

Don't die when source deps with [raw] are used

Debug log update

This adds support for their format both on the first read and when
looking for modifications to dependencies.

adjust `data_dir_correct` test in ct suite

don't bother checking the profile in the path to the data dir

filter available package versions individually by buildtools

deploy escript to travis releases instead of s3

Fix path command to consider {deps_dir, "deps"} in rebar.config

#1185 fix windows git working directory

fix issue #1185 git working directory issues due to command line options in Windows

don't needlessly update the code path twice if no hooks to run

don't generated coverage for `extra_src_dirs`

closes #1057 and #1179

fix otp-19 bootstrap compiling

upgrade erlware_commons for OTP-19 fix

Fix shell node startup

Add support for xref_ignores

Allows overwrite default cache dir using REBAR_CACHE_DIR

Fixes logging from remote shells in rebar3 shell

Add a provider hook for edoc

Change to new hex.pm repo URL

upgrade relx to 3.19.0

Allows `pre_hooks` and `post_hooks` to fire for the `edoc` provider.

Somehow swapping the tty handler once more fixes everything. I guess we
were missing a step somehow.

upgrade relx and erlware_commons

supported dist_node in ct and eunit

Recursively load configurations in shell

The sys.config loading in the shell command did not implement recursion
and thus processed the file different from a `erl -config cfg/path`.

Instead of reading every time that rebar_dir:global_cache_dir/1 is
called