bash-fun/test/shunit2

#! /bin/sh
# vim:et:ft=sh:sts=2:sw=2
#
# Copyright 2008-2018 Kate Ward. All Rights Reserved.
# Released under the Apache 2.0 license.
#
# shUnit2 -- Unit testing framework for Unix shell scripts.
# https://github.com/kward/shunit2
#
# Author: kate.ward@forestent.com (Kate Ward)
#
# shUnit2 is a xUnit based unit test framework for Bourne shell scripts. It is
# based on the popular JUnit unit testing framework for Java.
#
# $() are not fully portable (POSIX != portable).
#   shellcheck disable=SC2006
# expr may be antiquated, but it is the only solution in some cases.
#   shellcheck disable=SC2003
# Commands are purposely escaped so they can be mocked outside shUnit2.
#   shellcheck disable=SC1001,SC1012

# Return if shunit2 already loaded.
\[ -n "${SHUNIT_VERSION:-}" ] && exit 0
SHUNIT_VERSION='2.1.7'

# Return values that scripts can use.
SHUNIT_TRUE=0
SHUNIT_FALSE=1
SHUNIT_ERROR=2

# Logging functions.
_shunit_warn() {
  echo "${__shunit_ansi_yellow}shunit2:WARN${__shunit_ansi_none} $*" >&2
}
_shunit_error() {
  echo "${__shunit_ansi_red}shunit2:ERROR${__shunit_ansi_none} $*" >&2
}
_shunit_fatal() {
  echo "${__shunit_ansi_red}shunit2:FATAL${__shunit_ansi_none} $*" >&2
  exit ${SHUNIT_ERROR}
}

# Determine some reasonable command defaults.
__SHUNIT_UNAME_S=`uname -s`
case "${__SHUNIT_UNAME_S}" in
  BSD) __SHUNIT_CMD_EXPR='gexpr' ;;
  *) __SHUNIT_CMD_EXPR='expr' ;;
esac

__SHUNIT_CMD_ECHO_ESC='echo -e'
# shellcheck disable=SC2039
\[ "`echo -e test`" = '-e test' ] && __SHUNIT_CMD_ECHO_ESC='echo'

# Commands a user can override if needed.
SHUNIT_CMD_EXPR=${SHUNIT_CMD_EXPR:-${__SHUNIT_CMD_EXPR}}

# Enable color output. Options are 'never', 'always', or 'auto'.
SHUNIT_COLOR=${SHUNIT_COLOR:-auto}

# Specific shell checks.
if \[ -n "${ZSH_VERSION:-}" ]; then
  setopt |grep "^shwordsplit$" >/dev/null
  if \[ $? -ne ${SHUNIT_TRUE} ]; then
    _shunit_fatal 'zsh shwordsplit option is required for proper operation'
  fi
  if \[ -z "${SHUNIT_PARENT:-}" ]; then
    _shunit_fatal "zsh does not pass \$0 through properly. please declare \
\"SHUNIT_PARENT=\$0\" before calling shUnit2"
  fi
fi

#
# Constants
#

__SHUNIT_MODE_SOURCED='sourced'
__SHUNIT_MODE_STANDALONE='standalone'
__SHUNIT_PARENT=${SHUNIT_PARENT:-$0}

# ANSI colors.
__SHUNIT_ANSI_NONE='\033[0m'
__SHUNIT_ANSI_RED='\033[1;31m'
__SHUNIT_ANSI_GREEN='\033[1;32m'
__SHUNIT_ANSI_YELLOW='\033[1;33m'
__SHUNIT_ANSI_CYAN='\033[1;36m'

# Set the constants readonly.
__shunit_constants=`set |grep '^__SHUNIT_' |cut -d= -f1`
echo "${__shunit_constants}" |grep '^Binary file' >/dev/null && \
    __shunit_constants=`set |grep -a '^__SHUNIT_' |cut -d= -f1`
for __shunit_const in ${__shunit_constants}; do
  if \[ -z "${ZSH_VERSION:-}" ]; then
    readonly "${__shunit_const}"
  else
    case ${ZSH_VERSION} in
      [123].*) readonly "${__shunit_const}" ;;
      *) readonly -g "${__shunit_const}"  # Declare readonly constants globally.
    esac
  fi
done
unset __shunit_const __shunit_constants

#
# Internal variables.
#

# Variables.
__shunit_lineno=''  # Line number of executed test.
__shunit_mode=${__SHUNIT_MODE_SOURCED}  # Operating mode.
__shunit_reportGenerated=${SHUNIT_FALSE}  # Is report generated.
__shunit_script=''  # Filename of unittest script (standalone mode).
__shunit_skip=${SHUNIT_FALSE}  # Is skipping enabled.
__shunit_suite=''  # Suite of tests to execute.

# ANSI colors (populated by _shunit_configureColor()).
__shunit_ansi_none=''
__shunit_ansi_red=''
__shunit_ansi_green=''
__shunit_ansi_yellow=''
__shunit_ansi_cyan=''

# Counts of tests.
__shunit_testSuccess=${SHUNIT_TRUE}
__shunit_testsTotal=0
__shunit_testsPassed=0
__shunit_testsFailed=0

# Counts of asserts.
__shunit_assertsTotal=0
__shunit_assertsPassed=0
__shunit_assertsFailed=0
__shunit_assertsSkipped=0

#
# Macros.
#

# shellcheck disable=SC2016,SC2089
_SHUNIT_LINENO_='eval __shunit_lineno=""; if \[ "${1:-}" = "--lineno" ]; then \[ -n "$2" ] && __shunit_lineno="[$2] "; shift 2; fi'

#-----------------------------------------------------------------------------
# Assertion functions.
#

# Assert that two values are equal to one another.
#
# Args:
#   message: string: failure message [optional]
#   expected: string: expected value
#   actual: string: actual value
# Returns:
#   integer: success (TRUE/FALSE/ERROR constant)
assertEquals() {
  # shellcheck disable=SC2090
  ${_SHUNIT_LINENO_}
  if \[ $# -lt 2 -o $# -gt 3 ]; then
    _shunit_error "assertEquals() requires two or three arguments; $# given"
    _shunit_assertFail
    return ${SHUNIT_ERROR}
  fi
  _shunit_shouldSkip && return ${SHUNIT_TRUE}

  shunit_message_=${__shunit_lineno}
  if \[ $# -eq 3 ]; then
    shunit_message_="${shunit_message_}$1"
    shift
  fi
  shunit_expected_=$1
  shunit_actual_=$2

  shunit_return=${SHUNIT_TRUE}
  if \[ "${shunit_expected_}" = "${shunit_actual_}" ]; then
    _shunit_assertPass
  else
    failNotEquals "${shunit_message_}" "${shunit_expected_}" "${shunit_actual_}"
    shunit_return=${SHUNIT_FALSE}
  fi

  unset shunit_message_ shunit_expected_ shunit_actual_
  return ${shunit_return}
}
# shellcheck disable=SC2016,SC2034
_ASSERT_EQUALS_='eval assertEquals --lineno "${LINENO:-}"'

# Assert that two values are not equal to one another.
#
# Args:
#   message: string: failure message [optional]
#   expected: string: expected value
#   actual: string: actual value
# Returns:
#   integer: success (TRUE/FALSE/ERROR constant)
assertNotEquals() {
  # shellcheck disable=SC2090
  ${_SHUNIT_LINENO_}
  if \[ $# -lt 2 -o $# -gt 3 ]; then
    _shunit_error "assertNotEquals() requires two or three arguments; $# given"
    _shunit_assertFail
    return ${SHUNIT_ERROR}
  fi
  _shunit_shouldSkip && return ${SHUNIT_TRUE}

  shunit_message_=${__shunit_lineno}
  if \[ $# -eq 3 ]; then
    shunit_message_="${shunit_message_}$1"
    shift
  fi
  shunit_expected_=$1
  shunit_actual_=$2

  shunit_return=${SHUNIT_TRUE}
  if \[ "${shunit_expected_}" != "${shunit_actual_}" ]; then
    _shunit_assertPass
  else
    failSame "${shunit_message_}" "$@"
    shunit_return=${SHUNIT_FALSE}
  fi

  unset shunit_message_ shunit_expected_ shunit_actual_
  return ${shunit_return}
}
# shellcheck disable=SC2016,SC2034
_ASSERT_NOT_EQUALS_='eval assertNotEquals --lineno "${LINENO:-}"'

# Assert that a value is null (i.e. an empty string)
#
# Args:
#   message: string: failure message [optional]
#   actual: string: actual value
# Returns:
#   integer: success (TRUE/FALSE/ERROR constant)
assertNull() {
  # shellcheck disable=SC2090
  ${_SHUNIT_LINENO_}
  if \[ $# -lt 1 -o $# -gt 2 ]; then
    _shunit_error "assertNull() requires one or two arguments; $# given"
    _shunit_assertFail
    return ${SHUNIT_ERROR}
  fi
  _shunit_shouldSkip && return ${SHUNIT_TRUE}

  shunit_message_=${__shunit_lineno}
  if \[ $# -eq 2 ]; then
    shunit_message_="${shunit_message_}$1"
    shift
  fi
  assertTrue "${shunit_message_}" "[ -z '$1' ]"
  shunit_return=$?

  unset shunit_message_
  return ${shunit_return}
}
# shellcheck disable=SC2016,SC2034
_ASSERT_NULL_='eval assertNull --lineno "${LINENO:-}"'

# Assert that a value is not null (i.e. a non-empty string)
#
# Args:
#   message: string: failure message [optional]
#   actual: string: actual value
# Returns:
#   integer: success (TRUE/FALSE/ERROR constant)
assertNotNull() {
  # shellcheck disable=SC2090
  ${_SHUNIT_LINENO_}
  if \[ $# -gt 2 ]; then  # allowing 0 arguments as $1 might actually be null
    _shunit_error "assertNotNull() requires one or two arguments; $# given"
    _shunit_assertFail
    return ${SHUNIT_ERROR}
  fi
  _shunit_shouldSkip && return ${SHUNIT_TRUE}

  shunit_message_=${__shunit_lineno}
  if \[ $# -eq 2 ]; then
    shunit_message_="${shunit_message_}$1"
    shift
  fi
  shunit_actual_=`_shunit_escapeCharactersInString "${1:-}"`
  test -n "${shunit_actual_}"
  assertTrue "${shunit_message_}" $?
  shunit_return=$?

  unset shunit_actual_ shunit_message_
  return ${shunit_return}
}
# shellcheck disable=SC2016,SC2034
_ASSERT_NOT_NULL_='eval assertNotNull --lineno "${LINENO:-}"'

# Assert that two values are the same (i.e. equal to one another).
#
# Args:
#   message: string: failure message [optional]
#   expected: string: expected value
#   actual: string: actual value
# Returns:
#   integer: success (TRUE/FALSE/ERROR constant)
assertSame() {
  # shellcheck disable=SC2090
  ${_SHUNIT_LINENO_}
  if \[ $# -lt 2 -o $# -gt 3 ]; then
    _shunit_error "assertSame() requires two or three arguments; $# given"
    _shunit_assertFail
    return ${SHUNIT_ERROR}
  fi
  _shunit_shouldSkip && return ${SHUNIT_TRUE}

  shunit_message_=${__shunit_lineno}
  if \[ $# -eq 3 ]; then
    shunit_message_="${shunit_message_}$1"
    shift
  fi
  assertEquals "${shunit_message_}" "$1" "$2"
  shunit_return=$?

  unset shunit_message_
  return ${shunit_return}
}
# shellcheck disable=SC2016,SC2034
_ASSERT_SAME_='eval assertSame --lineno "${LINENO:-}"'

# Assert that two values are not the same (i.e. not equal to one another).
#
# Args:
#   message: string: failure message [optional]
#   expected: string: expected value
#   actual: string: actual value
# Returns:
#   integer: success (TRUE/FALSE/ERROR constant)
assertNotSame() {
  # shellcheck disable=SC2090
  ${_SHUNIT_LINENO_}
  if \[ $# -lt 2 -o $# -gt 3 ]; then
    _shunit_error "assertNotSame() requires two or three arguments; $# given"
    _shunit_assertFail
    return ${SHUNIT_ERROR}
  fi
  _shunit_shouldSkip && return ${SHUNIT_TRUE}

  shunit_message_=${__shunit_lineno}
  if \[ $# -eq 3 ]; then
    shunit_message_="${shunit_message_:-}$1"
    shift
  fi
  assertNotEquals "${shunit_message_}" "$1" "$2"
  shunit_return=$?

  unset shunit_message_
  return ${shunit_return}
}
# shellcheck disable=SC2016,SC2034
_ASSERT_NOT_SAME_='eval assertNotSame --lineno "${LINENO:-}"'

# Assert that a value or shell test condition is true.
#
# In shell, a value of 0 is true and a non-zero value is false. Any integer
# value passed can thereby be tested.
#
# Shell supports much more complicated tests though, and a means to support
# them was needed. As such, this function tests that conditions are true or
# false through evaluation rather than just looking for a true or false.
#
# The following test will succeed:
#   assertTrue 0
#   assertTrue "[ 34 -gt 23 ]"
# The following test will fail with a message:
#   assertTrue 123
#   assertTrue "test failed" "[ -r '/non/existent/file' ]"
#
# Args:
#   message: string: failure message [optional]
#   condition: string: integer value or shell conditional statement
# Returns:
#   integer: success (TRUE/FALSE/ERROR constant)
assertTrue() {
  # shellcheck disable=SC2090
  ${_SHUNIT_LINENO_}
  if \[ $# -lt 1 -o $# -gt 2 ]; then
    _shunit_error "assertTrue() takes one or two arguments; $# given"
    _shunit_assertFail
    return ${SHUNIT_ERROR}
  fi
  _shunit_shouldSkip && return ${SHUNIT_TRUE}

  shunit_message_=${__shunit_lineno}
  if \[ $# -eq 2 ]; then
    shunit_message_="${shunit_message_}$1"
    shift
  fi
  shunit_condition_=$1

  # See if condition is an integer, i.e. a return value.
  shunit_match_=`expr "${shunit_condition_}" : '\([0-9]*\)'`
  shunit_return=${SHUNIT_TRUE}
  if \[ -z "${shunit_condition_}" ]; then
    # Null condition.
    shunit_return=${SHUNIT_FALSE}
  elif \[ -n "${shunit_match_}" -a "${shunit_condition_}" = "${shunit_match_}" ]
  then
    # Possible return value. Treating 0 as true, and non-zero as false.
    \[ "${shunit_condition_}" -ne 0 ] && shunit_return=${SHUNIT_FALSE}
  else
    # Hopefully... a condition.
    ( eval "${shunit_condition_}" ) >/dev/null 2>&1
    \[ $? -ne 0 ] && shunit_return=${SHUNIT_FALSE}
  fi

  # Record the test.
  if \[ ${shunit_return} -eq ${SHUNIT_TRUE} ]; then
    _shunit_assertPass
  else
    _shunit_assertFail "${shunit_message_}"
  fi

  unset shunit_message_ shunit_condition_ shunit_match_
  return ${shunit_return}
}
# shellcheck disable=SC2016,SC2034
_ASSERT_TRUE_='eval assertTrue --lineno "${LINENO:-}"'

# Assert that a value or shell test condition is false.
#
# In shell, a value of 0 is true and a non-zero value is false. Any integer
# value passed can thereby be tested.
#
# Shell supports much more complicated tests though, and a means to support
# them was needed. As such, this function tests that conditions are true or
# false through evaluation rather than just looking for a true or false.
#
# The following test will succeed:
#   assertFalse 1
#   assertFalse "[ 'apples' = 'oranges' ]"
# The following test will fail with a message:
#   assertFalse 0
#   assertFalse "test failed" "[ 1 -eq 1 -a 2 -eq 2 ]"
#
# Args:
#   message: string: failure message [optional]
#   condition: string: integer value or shell conditional statement
# Returns:
#   integer: success (TRUE/FALSE/ERROR constant)
assertFalse() {
  # shellcheck disable=SC2090
  ${_SHUNIT_LINENO_}
  if \[ $# -lt 1 -o $# -gt 2 ]; then
    _shunit_error "assertFalse() quires one or two arguments; $# given"
    _shunit_assertFail
    return ${SHUNIT_ERROR}
  fi
  _shunit_shouldSkip && return ${SHUNIT_TRUE}

  shunit_message_=${__shunit_lineno}
  if \[ $# -eq 2 ]; then
    shunit_message_="${shunit_message_}$1"
    shift
  fi
  shunit_condition_=$1

  # See if condition is an integer, i.e. a return value.
  shunit_match_=`expr "${shunit_condition_}" : '\([0-9]*\)'`
  shunit_return=${SHUNIT_TRUE}
  if \[ -z "${shunit_condition_}" ]; then
    # Null condition.
    shunit_return=${SHUNIT_FALSE}
  elif \[ -n "${shunit_match_}" -a "${shunit_condition_}" = "${shunit_match_}" ]
  then
    # Possible return value. Treating 0 as true, and non-zero as false.
    \[ "${shunit_condition_}" -eq 0 ] && shunit_return=${SHUNIT_FALSE}
  else
    # Hopefully... a condition.
    ( eval "${shunit_condition_}" ) >/dev/null 2>&1
    \[ $? -eq 0 ] && shunit_return=${SHUNIT_FALSE}
  fi

  # Record the test.
  if \[ "${shunit_return}" -eq "${SHUNIT_TRUE}" ]; then
    _shunit_assertPass
  else
    _shunit_assertFail "${shunit_message_}"
  fi

  unset shunit_message_ shunit_condition_ shunit_match_
  return "${shunit_return}"
}
# shellcheck disable=SC2016,SC2034
_ASSERT_FALSE_='eval assertFalse --lineno "${LINENO:-}"'

#-----------------------------------------------------------------------------
# Failure functions.
#

# Records a test failure.
#
# Args:
#   message: string: failure message [optional]
# Returns:
#   integer: success (TRUE/FALSE/ERROR constant)
fail() {
  # shellcheck disable=SC2090
  ${_SHUNIT_LINENO_}
  if \[ $# -gt 1 ]; then
    _shunit_error "fail() requires zero or one arguments; $# given"
    return ${SHUNIT_ERROR}
  fi
  _shunit_shouldSkip && return ${SHUNIT_TRUE}

  shunit_message_=${__shunit_lineno}
  if \[ $# -eq 1 ]; then
    shunit_message_="${shunit_message_}$1"
    shift
  fi

  _shunit_assertFail "${shunit_message_}"

  unset shunit_message_
  return ${SHUNIT_FALSE}
}
# shellcheck disable=SC2016,SC2034
_FAIL_='eval fail --lineno "${LINENO:-}"'

# Records a test failure, stating two values were not equal.
#
# Args:
#   message: string: failure message [optional]
#   expected: string: expected value
#   actual: string: actual value
# Returns:
#   integer: success (TRUE/FALSE/ERROR constant)
failNotEquals() {
  # shellcheck disable=SC2090
  ${_SHUNIT_LINENO_}
  if \[ $# -lt 2 -o $# -gt 3 ]; then
    _shunit_error "failNotEquals() requires one or two arguments; $# given"
    return ${SHUNIT_ERROR}
  fi
  _shunit_shouldSkip && return ${SHUNIT_TRUE}

  shunit_message_=${__shunit_lineno}
  if \[ $# -eq 3 ]; then
    shunit_message_="${shunit_message_}$1"
    shift
  fi
  shunit_expected_=$1
  shunit_actual_=$2

  _shunit_assertFail "${shunit_message_:+${shunit_message_} }expected:<${shunit_expected_}> but was:<${shunit_actual_}>"

  unset shunit_message_ shunit_expected_ shunit_actual_
  return ${SHUNIT_FALSE}
}
# shellcheck disable=SC2016,SC2034
_FAIL_NOT_EQUALS_='eval failNotEquals --lineno "${LINENO:-}"'

# Records a test failure, stating two values should have been the same.
#
# Args:
#   message: string: failure message [optional]
#   expected: string: expected value
#   actual: string: actual value
# Returns:
#   integer: success (TRUE/FALSE/ERROR constant)
failSame()
{
  # shellcheck disable=SC2090
  ${_SHUNIT_LINENO_}
  if \[ $# -lt 2 -o $# -gt 3 ]; then
    _shunit_error "failSame() requires two or three arguments; $# given"
    return ${SHUNIT_ERROR}
  fi
  _shunit_shouldSkip && return ${SHUNIT_TRUE}

  shunit_message_=${__shunit_lineno}
  if \[ $# -eq 3 ]; then
    shunit_message_="${shunit_message_}$1"
    shift
  fi

  _shunit_assertFail "${shunit_message_:+${shunit_message_} }expected not same"

  unset shunit_message_
  return ${SHUNIT_FALSE}
}
# shellcheck disable=SC2016,SC2034
_FAIL_SAME_='eval failSame --lineno "${LINENO:-}"'

# Records a test failure, stating two values were not equal.
#
# This is functionally equivalent to calling failNotEquals().
#
# Args:
#   message: string: failure message [optional]
#   expected: string: expected value
#   actual: string: actual value
# Returns:
#   integer: success (TRUE/FALSE/ERROR constant)
failNotSame() {
  # shellcheck disable=SC2090
  ${_SHUNIT_LINENO_}
  if \[ $# -lt 2 -o $# -gt 3 ]; then
    _shunit_error "failNotEquals() requires one or two arguments; $# given"
    return ${SHUNIT_ERROR}
  fi
  _shunit_shouldSkip && return ${SHUNIT_TRUE}

  shunit_message_=${__shunit_lineno}
  if \[ $# -eq 3 ]; then
    shunit_message_="${shunit_message_}$1"
    shift
  fi
  failNotEquals "${shunit_message_}" "$1" "$2"
  shunit_return=$?

  unset shunit_message_
  return ${shunit_return}
}
# shellcheck disable=SC2016,SC2034
_FAIL_NOT_SAME_='eval failNotSame --lineno "${LINENO:-}"'

#-----------------------------------------------------------------------------
# Skipping functions.
#

# Force remaining assert and fail functions to be "skipped".
#
# This function forces the remaining assert and fail functions to be "skipped",
# i.e. they will have no effect. Each function skipped will be recorded so that
# the total of asserts and fails will not be altered.
#
# Args:
#   None
startSkipping() { __shunit_skip=${SHUNIT_TRUE}; }

# Resume the normal recording behavior of assert and fail calls.
#
# Args:
#   None
endSkipping() { __shunit_skip=${SHUNIT_FALSE}; }

# Returns the state of assert and fail call skipping.
#
# Args:
#   None
# Returns:
#   boolean: (TRUE/FALSE constant)
isSkipping() { return ${__shunit_skip}; }

#-----------------------------------------------------------------------------
# Suite functions.
#

# Stub. This function should contains all unit test calls to be made.
#
# DEPRECATED (as of 2.1.0)
#
# This function can be optionally overridden by the user in their test suite.
#
# If this function exists, it will be called when shunit2 is sourced. If it
# does not exist, shunit2 will search the parent script for all functions
# beginning with the word 'test', and they will be added dynamically to the
# test suite.
#
# This function should be overridden by the user in their unit test suite.
# Note: see _shunit_mktempFunc() for actual implementation
#
# Args:
#   None
#suite() { :; }  # DO NOT UNCOMMENT THIS FUNCTION

# Adds a function name to the list of tests schedule for execution.
#
# This function should only be called from within the suite() function.
#
# Args:
#   function: string: name of a function to add to current unit test suite
suite_addTest() {
  shunit_func_=${1:-}

  __shunit_suite="${__shunit_suite:+${__shunit_suite} }${shunit_func_}"
  __shunit_testsTotal=`expr ${__shunit_testsTotal} + 1`

  unset shunit_func_
}

# Stub. This function will be called once before any tests are run.
#
# Common one-time environment preparation tasks shared by all tests can be
# defined here.
#
# This function should be overridden by the user in their unit test suite.
# Note: see _shunit_mktempFunc() for actual implementation
#
# Args:
#   None
#oneTimeSetUp() { :; }  # DO NOT UNCOMMENT THIS FUNCTION

# Stub. This function will be called once after all tests are finished.
#
# Common one-time environment cleanup tasks shared by all tests can be defined
# here.
#
# This function should be overridden by the user in their unit test suite.
# Note: see _shunit_mktempFunc() for actual implementation
#
# Args:
#   None
#oneTimeTearDown() { :; }  # DO NOT UNCOMMENT THIS FUNCTION

# Stub. This function will be called before each test is run.
#
# Common environment preparation tasks shared by all tests can be defined here.
#
# This function should be overridden by the user in their unit test suite.
# Note: see _shunit_mktempFunc() for actual implementation
#
# Args:
#   None
#setUp() { :; }  # DO NOT UNCOMMENT THIS FUNCTION

# Note: see _shunit_mktempFunc() for actual implementation
# Stub. This function will be called after each test is run.
#
# Common environment cleanup tasks shared by all tests can be defined here.
#
# This function should be overridden by the user in their unit test suite.
# Note: see _shunit_mktempFunc() for actual implementation
#
# Args:
#   None
#tearDown() { :; }  # DO NOT UNCOMMENT THIS FUNCTION

#------------------------------------------------------------------------------
# Internal shUnit2 functions.
#

# Create a temporary directory to store various run-time files in.
#
# This function is a cross-platform temporary directory creation tool. Not all
# OSes have the `mktemp` function, so one is included here.
#
# Args:
#   None
# Outputs:
#   string: the temporary directory that was created
_shunit_mktempDir() {
  # Try the standard `mktemp` function.
  ( exec mktemp -dqt shunit.XXXXXX 2>/dev/null ) && return

  # The standard `mktemp` didn't work. Use our own.
  # shellcheck disable=SC2039
  if \[ -r '/dev/urandom' -a -x '/usr/bin/od' ]; then
    _shunit_random_=`/usr/bin/od -vAn -N4 -tx4 </dev/urandom \
        |sed 's/^[^0-9a-f]*//'`
  elif \[ -n "${RANDOM:-}" ]; then
    # $RANDOM works
    _shunit_random_=${RANDOM}${RANDOM}${RANDOM}$$
  else
    # `$RANDOM` doesn't work.
    _shunit_date_=`date '+%Y%m%d%H%M%S'`
    _shunit_random_=`expr "${_shunit_date_}" / $$`
  fi

  _shunit_tmpDir_="${TMPDIR:-/tmp}/shunit.${_shunit_random_}"
  ( umask 077 && \mkdir "${_shunit_tmpDir_}" ) || \
      _shunit_fatal 'could not create temporary directory! exiting'

  echo "${_shunit_tmpDir_}"
  unset _shunit_date_ _shunit_random_ _shunit_tmpDir_
}

# This function is here to work around issues in Cygwin.
#
# Args:
#   None
_shunit_mktempFunc() {
  for _shunit_func_ in oneTimeSetUp oneTimeTearDown setUp tearDown suite noexec
  do
    _shunit_file_="${__shunit_tmpDir}/${_shunit_func_}"
    \cat <<EOF >"${_shunit_file_}"
#! /bin/sh
exit ${SHUNIT_TRUE}
EOF
    \chmod +x "${_shunit_file_}"
  done

  unset _shunit_file_
}

# Final cleanup function to leave things as we found them.
#
# Besides removing the temporary directory, this function is in charge of the
# final exit code of the unit test. The exit code is based on how the script
# was ended (e.g. normal exit, or via Ctrl-C).
#
# Args:
#   name: string: name of the trap called (specified when trap defined)
_shunit_cleanup() {
  _shunit_name_=$1

  case ${_shunit_name_} in
    EXIT) _shunit_signal_=0 ;;
    INT) _shunit_signal_=2 ;;
    TERM) _shunit_signal_=15 ;;
    *)
      _shunit_error "unrecognized trap value (${_shunit_name_})"
      _shunit_signal_=0
      ;;
  esac

  # Do our work.
  \rm -fr "${__shunit_tmpDir}"

  # Exit for all non-EXIT signals.
  if \[ "${_shunit_name_}" != 'EXIT' ]; then
    _shunit_warn "trapped and now handling the (${_shunit_name_}) signal"
    # Disable EXIT trap.
    trap 0
    # Add 128 to signal and exit.
    exit "`expr "${_shunit_signal_}" + 128`"
  elif \[ ${__shunit_reportGenerated} -eq ${SHUNIT_FALSE} ] ; then
    _shunit_assertFail 'Unknown failure encountered running a test'
    _shunit_generateReport
    exit ${SHUNIT_ERROR}
  fi

  unset _shunit_name_ _shunit_signal_
}

# configureOutput based on user preferences, e.g. color.
#
# Args:
#   color: string: color mode (one of `always`, `auto`, or `none`).
_shunit_configureColor() {
  _shunit_color_=${SHUNIT_FALSE}  # By default, no color.
  case $1 in
    'always') _shunit_color_=${SHUNIT_TRUE} ;;
    'auto')
      ( exec tput >/dev/null 2>&1 )  # Check for existence of tput command.
      if [ $? -lt 127 ]; then
        _shunit_tput_=`tput colors`
        # shellcheck disable=SC2166,SC2181
        [ $? -eq 0 -a "${_shunit_tput_}" -ge 16 ] && _shunit_color_=${SHUNIT_TRUE}
      fi
      ;;
    'none') ;;
    *) _shunit_fatal "unrecognized SHUNIT_COLOR option '${SHUNIT_COLOR}'" ;;
  esac

  case ${_shunit_color_} in
    ${SHUNIT_TRUE})
      __shunit_ansi_none=${__SHUNIT_ANSI_NONE}
      __shunit_ansi_red=${__SHUNIT_ANSI_RED}
      __shunit_ansi_green=${__SHUNIT_ANSI_GREEN}
      __shunit_ansi_yellow=${__SHUNIT_ANSI_YELLOW}
      __shunit_ansi_cyan=${__SHUNIT_ANSI_CYAN}
      ;;
    ${SHUNIT_FALSE})
      __shunit_ansi_none=''
      __shunit_ansi_red=''
      __shunit_ansi_green=''
      __shunit_ansi_yellow=''
      __shunit_ansi_cyan=''
      ;;
  esac

  unset _shunit_color_ _shunit_tput_
}

# The actual running of the tests happens here.
#
# Args:
#   None
_shunit_execSuite() {
  for _shunit_test_ in ${__shunit_suite}; do
    __shunit_testSuccess=${SHUNIT_TRUE}

    # disable skipping
    endSkipping

    # execute the per-test setup function
    setUp

    # execute the test
    echo "${_shunit_test_}"
    eval "${_shunit_test_}"

    # execute the per-test tear-down function
    tearDown

    # update stats
    if \[ ${__shunit_testSuccess} -eq ${SHUNIT_TRUE} ]; then
      __shunit_testsPassed=`expr ${__shunit_testsPassed} + 1`
    else
      __shunit_testsFailed=`expr ${__shunit_testsFailed} + 1`
    fi
  done

  unset _shunit_test_
}

# Generates the user friendly report with appropriate OK/FAILED message.
#
# Args:
#   None
# Output:
#   string: the report of successful and failed tests, as well as totals.
_shunit_generateReport() {
  _shunit_ok_=${SHUNIT_TRUE}

  # If no exit code was provided one, determine an appropriate one.
  \[ "${__shunit_testsFailed}" -gt 0 \
      -o ${__shunit_testSuccess} -eq ${SHUNIT_FALSE} ] \
          && _shunit_ok_=${SHUNIT_FALSE}

  echo
  _shunit_msg_="Ran ${__shunit_ansi_cyan}${__shunit_testsTotal}${__shunit_ansi_none}"
  if \[ "${__shunit_testsTotal}" -eq 1 ]; then
    ${__SHUNIT_CMD_ECHO_ESC} "${_shunit_msg_} test."
  else
    ${__SHUNIT_CMD_ECHO_ESC} "${_shunit_msg_} tests."
  fi

  _shunit_failures_=''
  _shunit_skipped_=''
  \[ ${__shunit_assertsFailed} -gt 0 ] \
      && _shunit_failures_="failures=${__shunit_assertsFailed}"
  \[ ${__shunit_assertsSkipped} -gt 0 ] \
      && _shunit_skipped_="skipped=${__shunit_assertsSkipped}"

  if \[ ${_shunit_ok_} -eq ${SHUNIT_TRUE} ]; then
    _shunit_msg_="${__shunit_ansi_green}OK${__shunit_ansi_none}"
    \[ -n "${_shunit_skipped_}" ] \
        && _shunit_msg_="${_shunit_msg_} (${__shunit_ansi_yellow}${_shunit_skipped_}${__shunit_ansi_none})"
  else
    _shunit_msg_="${__shunit_ansi_red}FAILED${__shunit_ansi_none}"
    _shunit_msg_="${_shunit_msg_} (${__shunit_ansi_red}${_shunit_failures_}${__shunit_ansi_none}"
    \[ -n "${_shunit_skipped_}" ] \
        && _shunit_msg_="${_shunit_msg_},${__shunit_ansi_yellow}${_shunit_skipped_}${__shunit_ansi_none}"
    _shunit_msg_="${_shunit_msg_})"
  fi

  echo
  ${__SHUNIT_CMD_ECHO_ESC} "${_shunit_msg_}"
  __shunit_reportGenerated=${SHUNIT_TRUE}

  unset _shunit_failures_ _shunit_msg_ _shunit_ok_ _shunit_skipped_
}

# Test for whether a function should be skipped.
#
# Args:
#   None
# Returns:
#   boolean: whether the test should be skipped (TRUE/FALSE constant)
_shunit_shouldSkip() {
  \[ ${__shunit_skip} -eq ${SHUNIT_FALSE} ] && return ${SHUNIT_FALSE}
  _shunit_assertSkip
}

# Records a successful test.
#
# Args:
#   None
_shunit_assertPass() {
  __shunit_assertsPassed=`expr ${__shunit_assertsPassed} + 1`
  __shunit_assertsTotal=`expr ${__shunit_assertsTotal} + 1`
}

# Records a test failure.
#
# Args:
#   message: string: failure message to provide user
_shunit_assertFail() {
  __shunit_testSuccess=${SHUNIT_FALSE}
  __shunit_assertsFailed=`expr "${__shunit_assertsFailed}" + 1`
  __shunit_assertsTotal=`expr "${__shunit_assertsTotal}" + 1`

  \[ $# -gt 0 ] && ${__SHUNIT_CMD_ECHO_ESC} \
      "${__shunit_ansi_red}ASSERT:${__shunit_ansi_none}$*"
}

# Records a skipped test.
#
# Args:
#   None
_shunit_assertSkip() {
  __shunit_assertsSkipped=`expr "${__shunit_assertsSkipped}" + 1`
  __shunit_assertsTotal=`expr "${__shunit_assertsTotal}" + 1`
}

# Prepare a script filename for sourcing.
#
# Args:
#   script: string: path to a script to source
# Returns:
#   string: filename prefixed with ./ (if necessary)
_shunit_prepForSourcing() {
  _shunit_script_=$1
  case "${_shunit_script_}" in
    /*|./*) echo "${_shunit_script_}" ;;
    *) echo "./${_shunit_script_}" ;;
  esac
  unset _shunit_script_
}

# Escape a character in a string.
#
# Args:
#   c: string: unescaped character
#   s: string: to escape character in
# Returns:
#   string: with escaped character(s)
_shunit_escapeCharInStr() {
  \[ -n "$2" ] || return  # No point in doing work on an empty string.

  # Note: using shorter variable names to prevent conflicts with
  # _shunit_escapeCharactersInString().
  _shunit_c_=$1
  _shunit_s_=$2


  # Escape the character.
  # shellcheck disable=SC1003,SC2086
  echo ''${_shunit_s_}'' |sed 's/\'${_shunit_c_}'/\\\'${_shunit_c_}'/g'

  unset _shunit_c_ _shunit_s_
}

# Escape a character in a string.
#
# Args:
#   str: string: to escape characters in
# Returns:
#   string: with escaped character(s)
_shunit_escapeCharactersInString() {
  \[ -n "$1" ] || return  # No point in doing work on an empty string.

  _shunit_str_=$1

  # Note: using longer variable names to prevent conflicts with
  # _shunit_escapeCharInStr().
  for _shunit_char_ in '"' '$' "'" '`'; do
    _shunit_str_=`_shunit_escapeCharInStr "${_shunit_char_}" "${_shunit_str_}"`
  done

  echo "${_shunit_str_}"
  unset _shunit_char_ _shunit_str_
}

# Extract list of functions to run tests against.
#
# Args:
#   script: string: name of script to extract functions from
# Returns:
#   string: of function names
_shunit_extractTestFunctions() {
  _shunit_script_=$1

  # Extract the lines with test function names, strip of anything besides the
  # function name, and output everything on a single line.
  _shunit_regex_='^[ 	]*(function )*test[A-Za-z0-9_]* *\(\)'
  # shellcheck disable=SC2196
  egrep "${_shunit_regex_}" "${_shunit_script_}" \
  |sed 's/^[^A-Za-z0-9_]*//;s/^function //;s/\([A-Za-z0-9_]*\).*/\1/g' \
  |xargs

  unset _shunit_regex_ _shunit_script_
}

#------------------------------------------------------------------------------
# Main.
#

# Determine the operating mode.
if \[ $# -eq 0 ]; then
  __shunit_script=${__SHUNIT_PARENT}
  __shunit_mode=${__SHUNIT_MODE_SOURCED}
else
  __shunit_script=$1
  \[ -r "${__shunit_script}" ] || \
      _shunit_fatal "unable to read from ${__shunit_script}"
  __shunit_mode=${__SHUNIT_MODE_STANDALONE}
fi

# Create a temporary storage location.
__shunit_tmpDir=`_shunit_mktempDir`

# Provide a public temporary directory for unit test scripts.
# TODO(kward): document this.
SHUNIT_TMPDIR="${__shunit_tmpDir}/tmp"
\mkdir "${SHUNIT_TMPDIR}"

# Setup traps to clean up after ourselves.
trap '_shunit_cleanup EXIT' 0
trap '_shunit_cleanup INT' 2
trap '_shunit_cleanup TERM' 15

# Create phantom functions to work around issues with Cygwin.
_shunit_mktempFunc
PATH="${__shunit_tmpDir}:${PATH}"

# Make sure phantom functions are executable. This will bite if `/tmp` (or the
# current `$TMPDIR`) points to a path on a partition that was mounted with the
# 'noexec' option. The noexec command was created with `_shunit_mktempFunc()`.
noexec 2>/dev/null || _shunit_fatal \
    'Please declare TMPDIR with path on partition with exec permission.'

# We must manually source the tests in standalone mode.
if \[ "${__shunit_mode}" = "${__SHUNIT_MODE_STANDALONE}" ]; then
  # shellcheck disable=SC1090
  . "`_shunit_prepForSourcing \"${__shunit_script}\"`"
fi

# Configure default output coloring behavior.
_shunit_configureColor "${SHUNIT_COLOR}"

# Execute the oneTimeSetUp function (if it exists).
oneTimeSetUp

# Execute the suite function defined in the parent test script.
# DEPRECATED as of 2.1.0.
suite

# If no suite function was defined, dynamically build a list of functions.
if \[ -z "${__shunit_suite}" ]; then
  shunit_funcs_=`_shunit_extractTestFunctions "${__shunit_script}"`
  for shunit_func_ in ${shunit_funcs_}; do
    suite_addTest "${shunit_func_}"
  done
fi
unset shunit_func_ shunit_funcs_

_shunit_execSuite
oneTimeTearDown
_shunit_generateReport

# That's it folks.
\[ "${__shunit_testsFailed}" -eq 0 ]
exit $?