Conflicts:
	THANKS

Moves ct_extra_params to the end of the generated ct_run command.
This allows users to pass commands to the underlying emulator
using -erl_args. The included rt test demonstrates that it is
possible to pass an addtional option to ct_run and -erl_args at
the same time. Finally, the test executes in regular and verbose
modes because rebar constructs the ct_run command differently in
verbose mode.

This reverts commit 07e2232847baa634b7d09bdb87dafda6c6bcc4a4, reversing
changes made to 37cf470ae9.

Quote include/lib paths

Fix #249 (erlc regression)

Add gen_event template

Switch template instructions

Add code coverage analysis functionality to `qc'

Partially revert naming changes introduced in 93689703c1:
CoverageModules -> FilteredModules
get_coverage_modules -> get_matching_modules

Having the name "coverage" meaning "filtered/selected modules" can be
confused with code coverage.

- Use `cover' with QuickCheck testing
- Reuse the `cover_*' rebar.config options
- Refactor cover-related code to separate module (`qc_cover_utils')
  for use with both `eunit' and `qc'

This reverts commit 3297ffec2c89c6b7135165b4bfaeaeb07167c33c, reversing
changes made to 0401debb3c.

Move include/rebar.hrl to src/rebar.hrl

ct fixes

Add REBAR to environment before executing hooks

Some optimisation/refactoring using stdlib's functions

Add qualified name tests specification (see #118)

Fix #176

Add support for the Perforce VCS client via the "p4" tool

* Corrected regexp: `[x|y]` -> `[xy]`.
* Used an re:replace/4 option instead of multiple IOlist functions.

Namespaced types: fix build for 17.0

We use namespaced_types option to choose between dict() and dict:dict() types.

REBAR will be set to the rebar binary which was executed and runs the
builds. Enables the use of the same binary for rebar invocations as
part of a pre or post hook like so:

    ${REBAR} escriptize

erlc: replace if expression with case of

Motivated by the bug fix in 2c4d7d1.

erlc: fix typo in update_erlcinfo/3 clause

The combination of changes to rebar_erlc_compiler, and the fact
that erl_first_files is inherited, caused a regression. To fix
that, ensure every project uses its own .rebar/erlcinfo. While at
it, fix the issue that erl_first_files entries were not included
when initializing the dep digraph.

Reported-by: Louis-Philippe Gauthier
Reported-by: Roland Karlsson

Thanks: Tuncer Ayaz

suggested by @tuncer

noted by @tuncer

When trying to skip spec files under deps/ directory,
ignore "deps" component which is also included in Cwd.
For example, "/home/deps/src/myapp/test/cover.spec"
contains "deps" component but should not be skipped if
Cwd is "/home/deps/src/myapp/".

Extend CONTRIBUTING.md with regards to tests and credit.

Augment 'tests' option of 'rebar eunit' command with ability to specify
tests to run using module-qualified names. This change also forced me
to change the way modules for coverage and for testing itself are
selected - module-qualified tests specifications are now taken into
consideration. Extend tests to cover new functionality. Update
dialyzer_reference accordingly.

Error 'Command not found' when sname is used

improve output when using `rebar shell`

When a node is configured with -sname the REMSH_NAME_PART argument should be defined with using command 'hostname' but error occurs due extra space

attempt to emulate the behavior of
`erl -pa ebin -pa deps/*/ebin`

fix error messages and formatting issues of `rebar shell` by
shutting down and restarting the user subsystem in a mode more
hospitable to the shell than the simple user started when run
as an escript. emulate `error_logger` behaviour when the shell
is run via `erl`

add documentation of the shell command

limitations:

the erlang interrupt handler is not enabled when running as an
escript and there is no interface to re-enable it via erlang code.
this means `ctrl-c` will immediately exit the running process
unlike when running the shell via `erl`. `ctrl-g` is, however,
unaffected

the user subsystem is killed and restarted but not supervised. if
your code somehow relies on the user subsystem crashing and
restarting `rebar shell` may interfere with it's operation

rebar.hrl is only meant to be used by src/*.

Slim release support