grade.sh

CPATH='.:lib/hamcrest-core-1.3.jar:lib/junit-4.13.2.jar'

rm -rf student-submission
git clone $1 student-submission
echo 'Finished cloning'

if [[ -f student-submission/ListExamples.java ]]
then
  echo 'ListExamples.java found'
else
  echo 'ListExamples.java not found'
  echo 'Score: 0/4'
  exit 1
fi

cp student-submission/ListExamples.java ./

javac -cp $CPATH *.java

java -cp $CPATH org.junit.runner.JUnitCore TestListExamples > junit-output.txt

# The strategy used here relies on the last few lines of JUnit output, which
# looks like:

# FAILURES!!!
# Tests run: 4,  Failures: 2

# We check for "FAILURES!!!" and then do a bit of parsing of the last line to
# get the count
FAILURES=`grep -c FAILURES!!! junit-output.txt`


if [[ $FAILURES -eq 0 ]]
then
  RESULT_LINE=`grep "OK " junit-output.txt`
  PASSED=${RESULT_LINE:4:1}
  TOTAL=$PASSED
else
  # The ${VAR:N:M} syntax gets a substring of length M starting at index N
  # Note that since this is a precise character count into the "Tests run:..."
  # string, we'd need to update it if, say, we had a double-digit number of
  # tests. But it's nice and simple for the purposes of this script.

  # See, for example:
  # https://stackoverflow.com/questions/16484972/how-to-extract-a-substring-in-bash
  # https://www.gnu.org/savannah-checkouts/gnu/bash/manual/bash.html#Shell-Parameter-Expansion

  RESULT_LINE=`grep "Tests run:" junit-output.txt`
  COUNT=${RESULT_LINE:25:1}
  TOTAL=${RESULT_LINE:11:1}

  PASSED=$(echo "($TOTAL-$COUNT)" | bc)

  echo "JUnit output was:"
  cat junit-output.txt
fi

echo ""
echo "--------------"
echo "| Score: $PASSED/$TOTAL |"
echo "--------------"
echo ""