tools/test/test-setup.sh - bazel - Git at Google

 #!/usr/bin/env bash

 # Copyright 2015 The Bazel Authors. All rights reserved.
 #
 # Licensed under the Apache License, Version 2.0 (the "License");
 # you may not use this file except in compliance with the License.
 # You may obtain a copy of the License at
 #
 #    http://www.apache.org/licenses/LICENSE-2.0
 #
 # Unless required by applicable law or agreed to in writing, software
 # distributed under the License is distributed on an "AS IS" BASIS,
 # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 # See the License for the specific language governing permissions and
 # limitations under the License.

 # shift stderr to stdout.
 exec 2>&1

 # Executing the test log will page it.
 echo 'exec ${PAGER:-/usr/bin/less} "$0" || exit 1'
 echo "Executing tests from ${TEST_TARGET}"

 function is_absolute {
   [[ "$1" = /* ]] || [[ "$1" =~ ^[a-zA-Z]:[/\\].* ]]
 }

 # The original execution root. Usually this script changes directory into the
 # runfiles directory, so using $PWD is not a reliable way to find the execution
 # root.
 EXEC_ROOT="$PWD"

 # Declare that the executable is running in a `bazel test` environment
 # This allows test frameworks to enable output to the unprefixed environment variable
 # For example, if `BAZEL_TEST` and `XML_OUTPUT_FILE` are defined, write JUnit output
 export BAZEL_TEST=1

 # Bazel sets some environment vars to relative paths to improve caching and
 # support remote execution, where the absolute path may not be known to Bazel.
 # Convert them to absolute paths here before running the actual test.
 is_absolute "$TEST_PREMATURE_EXIT_FILE" ||
   TEST_PREMATURE_EXIT_FILE="$PWD/$TEST_PREMATURE_EXIT_FILE"
 is_absolute "$TEST_WARNINGS_OUTPUT_FILE" ||
   TEST_WARNINGS_OUTPUT_FILE="$PWD/$TEST_WARNINGS_OUTPUT_FILE"
 is_absolute "$TEST_LOGSPLITTER_OUTPUT_FILE" ||
   TEST_LOGSPLITTER_OUTPUT_FILE="$PWD/$TEST_LOGSPLITTER_OUTPUT_FILE"
 is_absolute "$TEST_INFRASTRUCTURE_FAILURE_FILE" ||
   TEST_INFRASTRUCTURE_FAILURE_FILE="$PWD/$TEST_INFRASTRUCTURE_FAILURE_FILE"
 is_absolute "$TEST_UNUSED_RUNFILES_LOG_FILE" ||
   TEST_UNUSED_RUNFILES_LOG_FILE="$PWD/$TEST_UNUSED_RUNFILES_LOG_FILE"
 is_absolute "$TEST_UNDECLARED_OUTPUTS_DIR" ||
   TEST_UNDECLARED_OUTPUTS_DIR="$PWD/$TEST_UNDECLARED_OUTPUTS_DIR"
 is_absolute "$TEST_UNDECLARED_OUTPUTS_MANIFEST" ||
   TEST_UNDECLARED_OUTPUTS_MANIFEST="$PWD/$TEST_UNDECLARED_OUTPUTS_MANIFEST"
 if [[ -n "$TEST_UNDECLARED_OUTPUTS_ZIP" ]]; then
   is_absolute "$TEST_UNDECLARED_OUTPUTS_ZIP" ||
     TEST_UNDECLARED_OUTPUTS_ZIP="$PWD/$TEST_UNDECLARED_OUTPUTS_ZIP"
 fi
 is_absolute "$TEST_UNDECLARED_OUTPUTS_ANNOTATIONS" ||
   TEST_UNDECLARED_OUTPUTS_ANNOTATIONS="$PWD/$TEST_UNDECLARED_OUTPUTS_ANNOTATIONS"
 is_absolute "$TEST_UNDECLARED_OUTPUTS_ANNOTATIONS_DIR" ||
   TEST_UNDECLARED_OUTPUTS_ANNOTATIONS_DIR="$PWD/$TEST_UNDECLARED_OUTPUTS_ANNOTATIONS_DIR"

 is_absolute "$TEST_SRCDIR" || TEST_SRCDIR="$PWD/$TEST_SRCDIR"
 is_absolute "$TEST_TMPDIR" || TEST_TMPDIR="$PWD/$TEST_TMPDIR"
 is_absolute "$HOME" || HOME="$TEST_TMPDIR"
 is_absolute "$XML_OUTPUT_FILE" || XML_OUTPUT_FILE="$PWD/$XML_OUTPUT_FILE"

 # Set USER to the current user, unless passed by Bazel via --test_env.
 if [[ -z "$USER" ]]; then
   export USER=$(whoami)
 fi

 # The test shard status file is only set for sharded tests.
 if [[ -n "$TEST_SHARD_STATUS_FILE" ]]; then
   is_absolute "$TEST_SHARD_STATUS_FILE" || TEST_SHARD_STATUS_FILE="$PWD/$TEST_SHARD_STATUS_FILE"
   mkdir -p "$(dirname "$TEST_SHARD_STATUS_FILE")"
 fi

 is_absolute "$RUNFILES_DIR" || RUNFILES_DIR="$PWD/$RUNFILES_DIR"

 # TODO(ulfjack): Standardize on RUNFILES_DIR and remove the {JAVA,PYTHON}_RUNFILES vars.
 is_absolute "$JAVA_RUNFILES" || JAVA_RUNFILES="$PWD/$JAVA_RUNFILES"
 is_absolute "$PYTHON_RUNFILES" || PYTHON_RUNFILES="$PWD/$PYTHON_RUNFILES"

 # Create directories for undeclared outputs and their annotations
 mkdir -p "$(dirname "$XML_OUTPUT_FILE")" \
     "$TEST_UNDECLARED_OUTPUTS_DIR" \
     "$TEST_UNDECLARED_OUTPUTS_ANNOTATIONS_DIR"

 # Create the test temp directory, which may not exist on the remote host when
 # doing a remote build.
 mkdir -p "$TEST_TMPDIR"

 # Unexport environment variables related to undeclared test outputs that are
 # only supposed to be used in this script.
 export -n TEST_UNDECLARED_OUTPUTS_MANIFEST
 export -n TEST_UNDECLARED_OUTPUTS_ZIP
 export -n TEST_UNDECLARED_OUTPUTS_ANNOTATIONS

 # Tell googletest about Bazel sharding.
 if [[ -n "${TEST_TOTAL_SHARDS+x}" ]] && ((TEST_TOTAL_SHARDS != 0)); then
   export GTEST_SHARD_INDEX="${TEST_SHARD_INDEX}"
   export GTEST_TOTAL_SHARDS="${TEST_TOTAL_SHARDS}"
   export GTEST_SHARD_STATUS_FILE="${TEST_SHARD_STATUS_FILE}"
 fi
 export GTEST_TMP_DIR="${TEST_TMPDIR}"

 # TODO(ulfjack): Update Gunit to accept XML_OUTPUT_FILE and drop this env
 # variable.
 GUNIT_OUTPUT="xml:${XML_OUTPUT_FILE}"

 RUNFILES_MANIFEST_FILE="${TEST_SRCDIR}/MANIFEST"

 function rlocation() {
   if is_absolute "$1" ; then
     # If the file path is already fully specified, simply return it.
     echo "$1"
   elif [[ -e "$TEST_SRCDIR/$1" ]]; then
     # If the file exists in the $TEST_SRCDIR then just use it.
     echo "$TEST_SRCDIR/$1"
   elif [[ -e "$RUNFILES_MANIFEST_FILE" ]]; then
     # If a runfiles manifest file exists then use it.
     echo "$(grep "^$1 " "$RUNFILES_MANIFEST_FILE" | sed 's/[^ ]* //')"
   fi
 }

 export -f rlocation
 export -f is_absolute
 # If RUNFILES_MANIFEST_ONLY is set to 1 and the manifest file does exist,
 # then test programs should use manifest file to find runfiles.
 if [[ "${RUNFILES_MANIFEST_ONLY:-}" == "1" && -e "${RUNFILES_MANIFEST_FILE:-}" ]]; then
   export RUNFILES_MANIFEST_FILE
   export RUNFILES_MANIFEST_ONLY
 fi

 DIR="$TEST_SRCDIR"
 if [ ! -z "$TEST_WORKSPACE" ]; then
   DIR="$DIR"/"$TEST_WORKSPACE"
 fi
 [[ -n "$RUNTEST_PRESERVE_CWD" ]] && DIR="$PWD"


 # normal commands are run in the exec-root where they have access to
 # the entire source tree. By chdir'ing to the runfiles root, tests only
 # have direct access to their declared dependencies.
 if [ -z "$COVERAGE_DIR" ]; then
   cd "$DIR" || { echo "Could not chdir $DIR"; exit 1; }
 fi

 # This header marks where --test_output=streamed will start being printed.
 echo "-----------------------------------------------------------------------------"

 # Unused if EXPERIMENTAL_SPLIT_XML_GENERATION is set.
 function encode_stream {
   # See generate-xml.sh for documentation.
   LC_ALL=C sed -E \
       -e 's/.*/& /g' \
       -e 's/(('\
 "$(echo -e '[\x9\x20-\x7f]')|"\
 "$(echo -e '[\xc0-\xdf][\x80-\xbf]')|"\
 "$(echo -e '[\xe0-\xec][\x80-\xbf][\x80-\xbf]')|"\
 "$(echo -e '[\xed][\x80-\x9f][\x80-\xbf]')|"\
 "$(echo -e '[\xee-\xef][\x80-\xbf][\x80-\xbf]')|"\
 "$(echo -e '[\xf0][\x80-\x8f][\x80-\xbf][\x80-\xbf]')"\
 ')*)./\1?/g' \
       -e 's/(.*)\?/\1/g' \
       -e 's|]]>|]]>]]<![CDATA[>|g'
 }

 function encode_output_file {
   if [ -f "$1" ]; then
     cat "$1" | encode_stream
   fi
 }

 # Unused if EXPERIMENTAL_SPLIT_XML_GENERATION is set.
 # Keep this in sync with generate-xml.sh!
 function write_xml_output_file {
   local duration=$(expr $(date +%s) - $start)
   local errors=0
   local error_msg=
   local signal="${1-}"
   local test_name=
   if [ -n "${XML_OUTPUT_FILE-}" -a ! -f "${XML_OUTPUT_FILE-}" ]; then
     # Create a default XML output file if the test runner hasn't generated it
     if [ -n "${signal}" ]; then
       errors=1
       if [ "${signal}" = "SIGTERM" ]; then
         error_msg="<error message=\"Timed out\"></error>"
       else
         error_msg="<error message=\"Terminated by signal ${signal}\"></error>"
       fi
     elif (( $exitCode != 0 )); then
       errors=1
       error_msg="<error message=\"exited with error code $exitCode\"></error>"
     fi
     test_name="${TEST_BINARY#./}"
     test_name="${test_name#../}"
     # Ensure that test shards have unique names in the xml output.
     if [[ -n "${TEST_TOTAL_SHARDS+x}" ]] && ((TEST_TOTAL_SHARDS != 0)); then
       ((shard_num=TEST_SHARD_INDEX+1))
       test_name="${test_name}"_shard_"$shard_num"/"$TEST_TOTAL_SHARDS"
     fi
     cat <<EOF >${XML_OUTPUT_FILE}
 <?xml version="1.0" encoding="UTF-8"?>
 <testsuites>
   <testsuite name="$test_name" tests="1" failures="0" errors="${errors}">
     <testcase name="$test_name" status="run" duration="${duration}" time="${duration}">${error_msg}</testcase>
     <system-out>Generated test.log (if the file is not UTF-8, then this may be unreadable):
       <![CDATA[$(encode_output_file "${XML_OUTPUT_FILE}.log")]]>
     </system-out>
   </testsuite>
 </testsuites>
 EOF
   fi
   rm -f "${XML_OUTPUT_FILE}.log"
 }

 # The path of this command-line is usually relative to the exec-root,
 # but when using --run_under it can be a "/bin/bash -c" command-line.

 # If the test is at the top of the tree, we have to add '.' to $PATH,
 PATH=".:$PATH"

 if [ -z "$COVERAGE_DIR" ]; then
   EXE="${1#./}"
   shift
 else
   EXE="${2#./}"
 fi

 if is_absolute "$EXE"; then
   TEST_PATH="$EXE"
 else
   TEST_PATH="$(rlocation $TEST_WORKSPACE/$EXE)"
 fi

 # TODO(jsharpe): Use --test_env=TEST_SHORT_EXEC_PATH=true to activate this code
 # path to workaround a bug with long executable paths when executing remote
 # tests on Windows.
 if [ ! -z "$TEST_SHORT_EXEC_PATH" ]; then
   QUALIFIER=0
   BASE="${EXEC_ROOT}/t${QUALIFIER}"
   while [[ -e "${BASE}" || -e "${BASE}.exe" || -e "${BASE}.zip" ]]; do
     ((QUALIFIER++))
     BASE="${EXEC_ROOT}/t${QUALIFIER}"
   done

   # Note for the commands below: "ln -s" is equivalent to "cp" on Windows.

   # Needs to be in the same directory for sh_test. Ignore the error when it
   # doesn't exist.
   ln -s "${TEST_PATH%.*}" "${BASE}" 2>/dev/null
   # Needs to be in the same directory for py_test. Ignore the error when it
   # doesn't exist.
   ln -s "${TEST_PATH%.*}.zip" "${BASE}.zip" 2>/dev/null
   # Needed for all tests.
   ln -s "${TEST_PATH}" "${BASE}.exe"
   TEST_PATH="${BASE}.exe"
 fi

 # Helper to kill a process and its entire group.
 function kill_group {
   local signal="${1-}"
   local pid="${2-}"
   kill -$signal -$pid &> /dev/null
 }

 childPid=""
 function signal_children {
   local signal="${1-}"
   if [ "${signal}" = "SIGTERM" ]; then
     echo "-- Test timed out at $(date +"%F %T %Z") --"
   fi
   if [ ! -z "$childPid" ]; then
     # For consistency with historical bazel behaviour, send signal to all child
     # processes, not just the first one. We use the process group for this
     # purpose.
     kill_group $signal $childPid
   fi
 }

 exitCode=0
 signals="$(trap -l | sed -E 's/[0-9]+\)//g')"
 if [[ "${EXPERIMENTAL_SPLIT_XML_GENERATION}" == "1" ]]; then
   for signal in $signals; do
     # SIGCHLD is expected when a subprocess dies
     [ "${signal}" = "SIGCHLD" ] && continue
     trap "signal_children ${signal}" ${signal}
   done
 else
   for signal in $signals; do
     # SIGCHLD is expected when a subprocess dies
     [ "${signal}" = "SIGCHLD" ] && continue
     trap "write_xml_output_file ${signal}; signal_children ${signal}" ${signal}
   done
 fi
 start=$(date +%s)

 # We have a challenge here: we want to forward signals to our child processes,
 # but we also want them to send themselves signals like SIGINT. Catching signals
 # ourselves requires use of background processes, trap, and wait. But normally
 # background processes are themselves unable to receive signals like SIGINT,
 # since those signals are intended for interactive processes - the only way for
 # them to get SIGINT in bash is for us to run them in the foreground.
 # To achieve this, we have to use `set -m` to enable Job Control in bash. This
 # has the effect of putting the child processes and any children of their own
 # into their own process groups, which are then able to receive SIGINT, etc,
 # without our shell interfering. Of course, this has the new complication that
 # anyone trying to SIGKILL *us* by group (as we know bazel's legacy process
 # wrapper does) will only kill this process and not the children below it. Any
 # reasonable sandboxing uses at least a process namespace, but we don't have the
 # luxury of assuming one, so our children could be left behind in that
 # eventuality. So, what we do is spawn a *second* background process that
 # watches for us to be killed, and then chain-kills the test's process group.
 # Aren't processes fun?
 set -m
 if [[ "${EXPERIMENTAL_SPLIT_XML_GENERATION}" == "1" ]]; then
   if [ -z "$COVERAGE_DIR" ]; then
     ("${TEST_PATH}" "$@" 2>&1) <&0 &
   else
     ("$1" "$TEST_PATH" "${@:3}" 2>&1) <&0 &
   fi
 else
   set -o pipefail
   if [ -z "$COVERAGE_DIR" ]; then
     ("${TEST_PATH}" "$@" 2>&1 | tee -a "${XML_OUTPUT_FILE}.log") <&0 &
   else
     ("$1" "$TEST_PATH" "${@:3}" 2>&1 | tee -a "${XML_OUTPUT_FILE}.log") <&0 &
   fi
   set +o pipefail
 fi
 childPid=$!

 # Cleanup helper
 # It would be nice to use `kill -0 $PPID` here, but when whatever called this
 # is running as a different user (as happens in remote execution) that will
 # return an error, causing us to prematurely reap a running test.
 ( if ! (ps -p $$ &> /dev/null || [ "`pgrep -a -g $$ 2> /dev/null`" != "" ] ); then
    # `ps` is known to be unrunnable in the darwin sandbox-exec environment due
    # to being a set-uid root program. pgrep exists in most environments, but not
    # universally. In the event that we find ourselves running in an environment
    # where *neither* exists, we have no reliable way to check if our parent is
    # still alive - so simply disable this cleanup routine entirely.
    exit 0
  fi
  while ps -p $$ &> /dev/null || [ "`pgrep -a -g $$ 2> /dev/null`" != "" ]; do
     sleep 10
  done
  # Parent process not found - we've been abandoned! Clean up test processes.
  kill_group SIGKILL $childPid
 ) &
 cleanupPid=$!

 set +m

 # Wait until $childPid fully exits.
 # We need to wait in a loop because wait is interrupted by any incoming trapped
 # signal (https://www.gnu.org/software/bash/manual/bash.html#Signals).
 while kill -0 $childPid 2>/dev/null; do
   wait $childPid
 done
 # Wait one more time to retrieve the exit code.
 wait $childPid
 exitCode=$?

 # By this point, we have everything we're willing to wait for. Tidy up our own
 # processes and move on.
 kill_group SIGKILL $childPid
 kill_group SIGKILL $cleanupPid &> /dev/null
 wait $cleanupPid

 for signal in $signals; do
   trap - ${signal}
 done
 if [[ "${EXPERIMENTAL_SPLIT_XML_GENERATION}" != "1" ]]; then
   # This call to write_xml_output_file does nothing if a a test.xml already
   # exists, e.g., because we received SIGTERM and the trap handler created it.
   write_xml_output_file
 fi

 # Add all of the files from the undeclared outputs directory to the manifest.
 if [[ -n "$TEST_UNDECLARED_OUTPUTS_DIR" && -n "$TEST_UNDECLARED_OUTPUTS_MANIFEST" ]]; then
   undeclared_outputs="$(find -L "$TEST_UNDECLARED_OUTPUTS_DIR" -type f | sort)"
   # Only write the manifest if there are any undeclared outputs.
   if [[ ! -z "$undeclared_outputs" ]]; then
     # For each file, write a tab-separated line with name (relative to
     # TEST_UNDECLARED_OUTPUTS_DIR), size, and mime type to the manifest. e.g.
     # foo.txt	9	text/plain
     while read -r undeclared_output; do
       rel_path="${undeclared_output#$TEST_UNDECLARED_OUTPUTS_DIR/}"
       # stat has different flags for different systems. -c is supported by GNU,
       # and -f by BSD (and thus OSX). Try both.
       file_size="$(stat -f%z "$undeclared_output" 2>/dev/null || stat -c%s "$undeclared_output" 2>/dev/null || echo "Could not stat $undeclared_output")"
       file_type="$(file -L -b --mime-type "$undeclared_output")"

       printf "$rel_path\t$file_size\t$file_type\n"
     done <<< "$undeclared_outputs" \
       > "$TEST_UNDECLARED_OUTPUTS_MANIFEST"
     if [[ ! -s "$TEST_UNDECLARED_OUTPUTS_MANIFEST" ]]; then
       rm "$TEST_UNDECLARED_OUTPUTS_MANIFEST"
     fi
   fi
 fi

 # Add all of the custom manifest entries to the annotation file.
 if [[ -n "$TEST_UNDECLARED_OUTPUTS_ANNOTATIONS" && \
       -n "$TEST_UNDECLARED_OUTPUTS_ANNOTATIONS_DIR" && \
       -d "$TEST_UNDECLARED_OUTPUTS_ANNOTATIONS_DIR" ]]; then
   (
    shopt -s failglob
    cat "$TEST_UNDECLARED_OUTPUTS_ANNOTATIONS_DIR"/*.part > "$TEST_UNDECLARED_OUTPUTS_ANNOTATIONS"
   ) 2> /dev/null
   (
    # length-delimited proto files
    shopt -s failglob
    cat $TEST_UNDECLARED_OUTPUTS_ANNOTATIONS_DIR/*.pb > "${TEST_UNDECLARED_OUTPUTS_ANNOTATIONS}.pb"
   ) 2> /dev/null
 fi

 # Zip up undeclared outputs.
 if [[ -n "$TEST_UNDECLARED_OUTPUTS_ZIP" ]] && cd "$TEST_UNDECLARED_OUTPUTS_DIR"; then
   shopt -s dotglob
   if [[ "$(echo *)" != "*" ]]; then
     # If * found nothing, echo printed the literal *.
     # Otherwise echo printed the top-level files and directories.
     # Pass files to zip with *, so paths with spaces aren't broken up.
     # Remove original files after zipping them.
     zip -qrm "$TEST_UNDECLARED_OUTPUTS_ZIP" -- * 2>/dev/null || \
         echo >&2 "Could not create \"$TEST_UNDECLARED_OUTPUTS_ZIP\": zip not found or failed"
   fi
 fi

 # Raise the original signal if the test terminated abnormally.
 if [ $exitCode -gt 128 ]; then
   kill -$(($exitCode - 128)) $$ &> /dev/null
 fi
 exit $exitCode
	#!/usr/bin/env bash

	# Copyright 2015 The Bazel Authors. All rights reserved.
	#
	# Licensed under the Apache License, Version 2.0 (the "License");
	# you may not use this file except in compliance with the License.
	# You may obtain a copy of the License at
	#
	# http://www.apache.org/licenses/LICENSE-2.0
	#
	# Unless required by applicable law or agreed to in writing, software
	# distributed under the License is distributed on an "AS IS" BASIS,
	# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
	# See the License for the specific language governing permissions and
	# limitations under the License.

	# shift stderr to stdout.
	exec 2>&1

	# Executing the test log will page it.
	echo 'exec ${PAGER:-/usr/bin/less} "$0" \|\| exit 1'
	echo "Executing tests from ${TEST_TARGET}"

	function is_absolute {
	[[ "$1" = /* ]] \|\| [[ "$1" =~ ^[a-zA-Z]:[/\\].* ]]
	}

	# The original execution root. Usually this script changes directory into the
	# runfiles directory, so using $PWD is not a reliable way to find the execution
	# root.
	EXEC_ROOT="$PWD"

	# Declare that the executable is running in a `bazel test` environment
	# This allows test frameworks to enable output to the unprefixed environment variable
	# For example, if `BAZEL_TEST` and `XML_OUTPUT_FILE` are defined, write JUnit output
	export BAZEL_TEST=1

	# Bazel sets some environment vars to relative paths to improve caching and
	# support remote execution, where the absolute path may not be known to Bazel.
	# Convert them to absolute paths here before running the actual test.
	is_absolute "$TEST_PREMATURE_EXIT_FILE" \|\|
	TEST_PREMATURE_EXIT_FILE="$PWD/$TEST_PREMATURE_EXIT_FILE"
	is_absolute "$TEST_WARNINGS_OUTPUT_FILE" \|\|
	TEST_WARNINGS_OUTPUT_FILE="$PWD/$TEST_WARNINGS_OUTPUT_FILE"
	is_absolute "$TEST_LOGSPLITTER_OUTPUT_FILE" \|\|
	TEST_LOGSPLITTER_OUTPUT_FILE="$PWD/$TEST_LOGSPLITTER_OUTPUT_FILE"
	is_absolute "$TEST_INFRASTRUCTURE_FAILURE_FILE" \|\|
	TEST_INFRASTRUCTURE_FAILURE_FILE="$PWD/$TEST_INFRASTRUCTURE_FAILURE_FILE"
	is_absolute "$TEST_UNUSED_RUNFILES_LOG_FILE" \|\|
	TEST_UNUSED_RUNFILES_LOG_FILE="$PWD/$TEST_UNUSED_RUNFILES_LOG_FILE"
	is_absolute "$TEST_UNDECLARED_OUTPUTS_DIR" \|\|
	TEST_UNDECLARED_OUTPUTS_DIR="$PWD/$TEST_UNDECLARED_OUTPUTS_DIR"
	is_absolute "$TEST_UNDECLARED_OUTPUTS_MANIFEST" \|\|
	TEST_UNDECLARED_OUTPUTS_MANIFEST="$PWD/$TEST_UNDECLARED_OUTPUTS_MANIFEST"
	if [[ -n "$TEST_UNDECLARED_OUTPUTS_ZIP" ]]; then
	is_absolute "$TEST_UNDECLARED_OUTPUTS_ZIP" \|\|
	TEST_UNDECLARED_OUTPUTS_ZIP="$PWD/$TEST_UNDECLARED_OUTPUTS_ZIP"
	fi
	is_absolute "$TEST_UNDECLARED_OUTPUTS_ANNOTATIONS" \|\|
	TEST_UNDECLARED_OUTPUTS_ANNOTATIONS="$PWD/$TEST_UNDECLARED_OUTPUTS_ANNOTATIONS"
	is_absolute "$TEST_UNDECLARED_OUTPUTS_ANNOTATIONS_DIR" \|\|
	TEST_UNDECLARED_OUTPUTS_ANNOTATIONS_DIR="$PWD/$TEST_UNDECLARED_OUTPUTS_ANNOTATIONS_DIR"

	is_absolute "$TEST_SRCDIR" \|\| TEST_SRCDIR="$PWD/$TEST_SRCDIR"
	is_absolute "$TEST_TMPDIR" \|\| TEST_TMPDIR="$PWD/$TEST_TMPDIR"
	is_absolute "$HOME" \|\| HOME="$TEST_TMPDIR"
	is_absolute "$XML_OUTPUT_FILE" \|\| XML_OUTPUT_FILE="$PWD/$XML_OUTPUT_FILE"

	# Set USER to the current user, unless passed by Bazel via --test_env.
	if [[ -z "$USER" ]]; then
	export USER=$(whoami)
	fi

	# The test shard status file is only set for sharded tests.
	if [[ -n "$TEST_SHARD_STATUS_FILE" ]]; then
	is_absolute "$TEST_SHARD_STATUS_FILE" \|\| TEST_SHARD_STATUS_FILE="$PWD/$TEST_SHARD_STATUS_FILE"
	mkdir -p "$(dirname "$TEST_SHARD_STATUS_FILE")"
	fi

	is_absolute "$RUNFILES_DIR" \|\| RUNFILES_DIR="$PWD/$RUNFILES_DIR"

	# TODO(ulfjack): Standardize on RUNFILES_DIR and remove the {JAVA,PYTHON}_RUNFILES vars.
	is_absolute "$JAVA_RUNFILES" \|\| JAVA_RUNFILES="$PWD/$JAVA_RUNFILES"
	is_absolute "$PYTHON_RUNFILES" \|\| PYTHON_RUNFILES="$PWD/$PYTHON_RUNFILES"

	# Create directories for undeclared outputs and their annotations
	mkdir -p "$(dirname "$XML_OUTPUT_FILE")" \
	"$TEST_UNDECLARED_OUTPUTS_DIR" \
	"$TEST_UNDECLARED_OUTPUTS_ANNOTATIONS_DIR"

	# Create the test temp directory, which may not exist on the remote host when
	# doing a remote build.
	mkdir -p "$TEST_TMPDIR"

	# Unexport environment variables related to undeclared test outputs that are
	# only supposed to be used in this script.
	export -n TEST_UNDECLARED_OUTPUTS_MANIFEST
	export -n TEST_UNDECLARED_OUTPUTS_ZIP
	export -n TEST_UNDECLARED_OUTPUTS_ANNOTATIONS

	# Tell googletest about Bazel sharding.
	if [[ -n "${TEST_TOTAL_SHARDS+x}" ]] && ((TEST_TOTAL_SHARDS != 0)); then
	export GTEST_SHARD_INDEX="${TEST_SHARD_INDEX}"
	export GTEST_TOTAL_SHARDS="${TEST_TOTAL_SHARDS}"
	export GTEST_SHARD_STATUS_FILE="${TEST_SHARD_STATUS_FILE}"
	fi
	export GTEST_TMP_DIR="${TEST_TMPDIR}"

	# TODO(ulfjack): Update Gunit to accept XML_OUTPUT_FILE and drop this env
	# variable.
	GUNIT_OUTPUT="xml:${XML_OUTPUT_FILE}"

	RUNFILES_MANIFEST_FILE="${TEST_SRCDIR}/MANIFEST"

	function rlocation() {
	if is_absolute "$1" ; then
	# If the file path is already fully specified, simply return it.
	echo "$1"
	elif [[ -e "$TEST_SRCDIR/$1" ]]; then
	# If the file exists in the $TEST_SRCDIR then just use it.
	echo "$TEST_SRCDIR/$1"
	elif [[ -e "$RUNFILES_MANIFEST_FILE" ]]; then
	# If a runfiles manifest file exists then use it.
	echo "$(grep "^$1 " "$RUNFILES_MANIFEST_FILE" \| sed 's/[^ ]* //')"
	fi
	}

	export -f rlocation
	export -f is_absolute
	# If RUNFILES_MANIFEST_ONLY is set to 1 and the manifest file does exist,
	# then test programs should use manifest file to find runfiles.
	if [[ "${RUNFILES_MANIFEST_ONLY:-}" == "1" && -e "${RUNFILES_MANIFEST_FILE:-}" ]]; then
	export RUNFILES_MANIFEST_FILE
	export RUNFILES_MANIFEST_ONLY
	fi

	DIR="$TEST_SRCDIR"
	if [ ! -z "$TEST_WORKSPACE" ]; then
	DIR="$DIR"/"$TEST_WORKSPACE"
	fi
	[[ -n "$RUNTEST_PRESERVE_CWD" ]] && DIR="$PWD"


	# normal commands are run in the exec-root where they have access to
	# the entire source tree. By chdir'ing to the runfiles root, tests only
	# have direct access to their declared dependencies.
	if [ -z "$COVERAGE_DIR" ]; then
	cd "$DIR" \|\| { echo "Could not chdir $DIR"; exit 1; }
	fi

	# This header marks where --test_output=streamed will start being printed.
	echo "-----------------------------------------------------------------------------"

	# Unused if EXPERIMENTAL_SPLIT_XML_GENERATION is set.
	function encode_stream {
	# See generate-xml.sh for documentation.
	LC_ALL=C sed -E \
	-e 's/.*/& /g' \
	-e 's/(('\
	"$(echo -e '[\x9\x20-\x7f]')\|"\
	"$(echo -e '[\xc0-\xdf][\x80-\xbf]')\|"\
	"$(echo -e '[\xe0-\xec][\x80-\xbf][\x80-\xbf]')\|"\
	"$(echo -e '[\xed][\x80-\x9f][\x80-\xbf]')\|"\
	"$(echo -e '[\xee-\xef][\x80-\xbf][\x80-\xbf]')\|"\
	"$(echo -e '[\xf0][\x80-\x8f][\x80-\xbf][\x80-\xbf]')"\
	')*)./\1?/g' \
	-e 's/(.*)\?/\1/g' \
	-e 's\|]]>\|]]>]]<![CDATA[>\|g'
	}

	function encode_output_file {
	if [ -f "$1" ]; then
	cat "$1" \| encode_stream
	fi
	}

	# Unused if EXPERIMENTAL_SPLIT_XML_GENERATION is set.
	# Keep this in sync with generate-xml.sh!
	function write_xml_output_file {
	local duration=$(expr $(date +%s) - $start)
	local errors=0
	local error_msg=
	local signal="${1-}"
	local test_name=
	if [ -n "${XML_OUTPUT_FILE-}" -a ! -f "${XML_OUTPUT_FILE-}" ]; then
	# Create a default XML output file if the test runner hasn't generated it
	if [ -n "${signal}" ]; then
	errors=1
	if [ "${signal}" = "SIGTERM" ]; then
	error_msg="<error message=\"Timed out\"></error>"
	else
	error_msg="<error message=\"Terminated by signal ${signal}\"></error>"
	fi
	elif (( $exitCode != 0 )); then
	errors=1
	error_msg="<error message=\"exited with error code $exitCode\"></error>"
	fi
	test_name="${TEST_BINARY#./}"
	test_name="${test_name#../}"
	# Ensure that test shards have unique names in the xml output.
	if [[ -n "${TEST_TOTAL_SHARDS+x}" ]] && ((TEST_TOTAL_SHARDS != 0)); then
	((shard_num=TEST_SHARD_INDEX+1))
	test_name="${test_name}"_shard_"$shard_num"/"$TEST_TOTAL_SHARDS"
	fi
	cat <<EOF >${XML_OUTPUT_FILE}
	<?xml version="1.0" encoding="UTF-8"?>
	<testsuites>
	<testsuite name="$test_name" tests="1" failures="0" errors="${errors}">
	<testcase name="$test_name" status="run" duration="${duration}" time="${duration}">${error_msg}</testcase>
	<system-out>Generated test.log (if the file is not UTF-8, then this may be unreadable):
	<![CDATA[$(encode_output_file "${XML_OUTPUT_FILE}.log")]]>
	</system-out>
	</testsuite>
	</testsuites>
	EOF
	fi
	rm -f "${XML_OUTPUT_FILE}.log"
	}

	# The path of this command-line is usually relative to the exec-root,
	# but when using --run_under it can be a "/bin/bash -c" command-line.

	# If the test is at the top of the tree, we have to add '.' to $PATH,
	PATH=".:$PATH"

	if [ -z "$COVERAGE_DIR" ]; then
	EXE="${1#./}"
	shift
	else
	EXE="${2#./}"
	fi

	if is_absolute "$EXE"; then
	TEST_PATH="$EXE"
	else
	TEST_PATH="$(rlocation $TEST_WORKSPACE/$EXE)"
	fi

	# TODO(jsharpe): Use --test_env=TEST_SHORT_EXEC_PATH=true to activate this code
	# path to workaround a bug with long executable paths when executing remote
	# tests on Windows.
	if [ ! -z "$TEST_SHORT_EXEC_PATH" ]; then
	QUALIFIER=0
	BASE="${EXEC_ROOT}/t${QUALIFIER}"
	while [[ -e "${BASE}" \|\| -e "${BASE}.exe" \|\| -e "${BASE}.zip" ]]; do
	((QUALIFIER++))
	BASE="${EXEC_ROOT}/t${QUALIFIER}"
	done

	# Note for the commands below: "ln -s" is equivalent to "cp" on Windows.

	# Needs to be in the same directory for sh_test. Ignore the error when it
	# doesn't exist.
	ln -s "${TEST_PATH%.*}" "${BASE}" 2>/dev/null
	# Needs to be in the same directory for py_test. Ignore the error when it
	# doesn't exist.
	ln -s "${TEST_PATH%.*}.zip" "${BASE}.zip" 2>/dev/null
	# Needed for all tests.
	ln -s "${TEST_PATH}" "${BASE}.exe"
	TEST_PATH="${BASE}.exe"
	fi

	# Helper to kill a process and its entire group.
	function kill_group {
	local signal="${1-}"
	local pid="${2-}"
	kill -$signal -$pid &> /dev/null
	}

	childPid=""
	function signal_children {
	local signal="${1-}"
	if [ "${signal}" = "SIGTERM" ]; then
	echo "-- Test timed out at $(date +"%F %T %Z") --"
	fi
	if [ ! -z "$childPid" ]; then
	# For consistency with historical bazel behaviour, send signal to all child
	# processes, not just the first one. We use the process group for this
	# purpose.
	kill_group $signal $childPid
	fi
	}

	exitCode=0
	signals="$(trap -l \| sed -E 's/[0-9]+\)//g')"
	if [[ "${EXPERIMENTAL_SPLIT_XML_GENERATION}" == "1" ]]; then
	for signal in $signals; do
	# SIGCHLD is expected when a subprocess dies
	[ "${signal}" = "SIGCHLD" ] && continue
	trap "signal_children ${signal}" ${signal}
	done
	else
	for signal in $signals; do
	# SIGCHLD is expected when a subprocess dies
	[ "${signal}" = "SIGCHLD" ] && continue
	trap "write_xml_output_file ${signal}; signal_children ${signal}" ${signal}
	done
	fi
	start=$(date +%s)

	# We have a challenge here: we want to forward signals to our child processes,
	# but we also want them to send themselves signals like SIGINT. Catching signals
	# ourselves requires use of background processes, trap, and wait. But normally
	# background processes are themselves unable to receive signals like SIGINT,
	# since those signals are intended for interactive processes - the only way for
	# them to get SIGINT in bash is for us to run them in the foreground.
	# To achieve this, we have to use `set -m` to enable Job Control in bash. This
	# has the effect of putting the child processes and any children of their own
	# into their own process groups, which are then able to receive SIGINT, etc,
	# without our shell interfering. Of course, this has the new complication that
	# anyone trying to SIGKILL us by group (as we know bazel's legacy process
	# wrapper does) will only kill this process and not the children below it. Any
	# reasonable sandboxing uses at least a process namespace, but we don't have the
	# luxury of assuming one, so our children could be left behind in that
	# eventuality. So, what we do is spawn a second background process that
	# watches for us to be killed, and then chain-kills the test's process group.
	# Aren't processes fun?
	set -m
	if [[ "${EXPERIMENTAL_SPLIT_XML_GENERATION}" == "1" ]]; then
	if [ -z "$COVERAGE_DIR" ]; then
	("${TEST_PATH}" "$@" 2>&1) <&0 &
	else
	("$1" "$TEST_PATH" "${@:3}" 2>&1) <&0 &
	fi
	else
	set -o pipefail
	if [ -z "$COVERAGE_DIR" ]; then
	("${TEST_PATH}" "$@" 2>&1 \| tee -a "${XML_OUTPUT_FILE}.log") <&0 &
	else
	("$1" "$TEST_PATH" "${@:3}" 2>&1 \| tee -a "${XML_OUTPUT_FILE}.log") <&0 &
	fi
	set +o pipefail
	fi
	childPid=$!

	# Cleanup helper
	# It would be nice to use `kill -0 $PPID` here, but when whatever called this
	# is running as a different user (as happens in remote execution) that will
	# return an error, causing us to prematurely reap a running test.
	( if ! (ps -p $$ &> /dev/null \|\| [ "`pgrep -a -g $$ 2> /dev/null`" != "" ] ); then
	# `ps` is known to be unrunnable in the darwin sandbox-exec environment due
	# to being a set-uid root program. pgrep exists in most environments, but not
	# universally. In the event that we find ourselves running in an environment
	# where neither exists, we have no reliable way to check if our parent is
	# still alive - so simply disable this cleanup routine entirely.
	exit 0
	fi
	while ps -p $$ &> /dev/null \|\| [ "`pgrep -a -g $$ 2> /dev/null`" != "" ]; do
	sleep 10
	done
	# Parent process not found - we've been abandoned! Clean up test processes.
	kill_group SIGKILL $childPid
	) &
	cleanupPid=$!

	set +m

	# Wait until $childPid fully exits.
	# We need to wait in a loop because wait is interrupted by any incoming trapped
	# signal (https://www.gnu.org/software/bash/manual/bash.html#Signals).
	while kill -0 $childPid 2>/dev/null; do
	wait $childPid
	done
	# Wait one more time to retrieve the exit code.
	wait $childPid
	exitCode=$?

	# By this point, we have everything we're willing to wait for. Tidy up our own
	# processes and move on.
	kill_group SIGKILL $childPid
	kill_group SIGKILL $cleanupPid &> /dev/null
	wait $cleanupPid

	for signal in $signals; do
	trap - ${signal}
	done
	if [[ "${EXPERIMENTAL_SPLIT_XML_GENERATION}" != "1" ]]; then
	# This call to write_xml_output_file does nothing if a a test.xml already
	# exists, e.g., because we received SIGTERM and the trap handler created it.
	write_xml_output_file
	fi

	# Add all of the files from the undeclared outputs directory to the manifest.
	if [[ -n "$TEST_UNDECLARED_OUTPUTS_DIR" && -n "$TEST_UNDECLARED_OUTPUTS_MANIFEST" ]]; then
	undeclared_outputs="$(find -L "$TEST_UNDECLARED_OUTPUTS_DIR" -type f \| sort)"
	# Only write the manifest if there are any undeclared outputs.
	if [[ ! -z "$undeclared_outputs" ]]; then
	# For each file, write a tab-separated line with name (relative to
	# TEST_UNDECLARED_OUTPUTS_DIR), size, and mime type to the manifest. e.g.
	# foo.txt 9 text/plain
	while read -r undeclared_output; do
	rel_path="${undeclared_output#$TEST_UNDECLARED_OUTPUTS_DIR/}"
	# stat has different flags for different systems. -c is supported by GNU,
	# and -f by BSD (and thus OSX). Try both.
	file_size="$(stat -f%z "$undeclared_output" 2>/dev/null \|\| stat -c%s "$undeclared_output" 2>/dev/null \|\| echo "Could not stat $undeclared_output")"
	file_type="$(file -L -b --mime-type "$undeclared_output")"

	printf "$rel_path\t$file_size\t$file_type\n"
	done <<< "$undeclared_outputs" \
	> "$TEST_UNDECLARED_OUTPUTS_MANIFEST"
	if [[ ! -s "$TEST_UNDECLARED_OUTPUTS_MANIFEST" ]]; then
	rm "$TEST_UNDECLARED_OUTPUTS_MANIFEST"
	fi
	fi
	fi

	# Add all of the custom manifest entries to the annotation file.
	if [[ -n "$TEST_UNDECLARED_OUTPUTS_ANNOTATIONS" && \
	-n "$TEST_UNDECLARED_OUTPUTS_ANNOTATIONS_DIR" && \
	-d "$TEST_UNDECLARED_OUTPUTS_ANNOTATIONS_DIR" ]]; then
	(
	shopt -s failglob
	cat "$TEST_UNDECLARED_OUTPUTS_ANNOTATIONS_DIR"/*.part > "$TEST_UNDECLARED_OUTPUTS_ANNOTATIONS"
	) 2> /dev/null
	(
	# length-delimited proto files
	shopt -s failglob
	cat $TEST_UNDECLARED_OUTPUTS_ANNOTATIONS_DIR/*.pb > "${TEST_UNDECLARED_OUTPUTS_ANNOTATIONS}.pb"
	) 2> /dev/null
	fi

	# Zip up undeclared outputs.
	if [[ -n "$TEST_UNDECLARED_OUTPUTS_ZIP" ]] && cd "$TEST_UNDECLARED_OUTPUTS_DIR"; then
	shopt -s dotglob
	if [[ "$(echo )" != "" ]]; then
	# If * found nothing, echo printed the literal *.
	# Otherwise echo printed the top-level files and directories.
	# Pass files to zip with *, so paths with spaces aren't broken up.
	# Remove original files after zipping them.
	zip -qrm "$TEST_UNDECLARED_OUTPUTS_ZIP" -- * 2>/dev/null \|\| \
	echo >&2 "Could not create \"$TEST_UNDECLARED_OUTPUTS_ZIP\": zip not found or failed"
	fi
	fi

	# Raise the original signal if the test terminated abnormally.
	if [ $exitCode -gt 128 ]; then
	kill -$(($exitCode - 128)) $$ &> /dev/null
	fi
	exit $exitCode