Inconsistency in reported result and my running result on D4J 2.0.0 Chart 1

[BEbillionaireUSD]

Context

I refer to this code #30 (comment) and run Gzoltar on Defects4j 2.0

Steps to Reproduce

1. checkout -p Chart -v 1b -w checkouts/Chart_1_buggy
2. Run the following commands

JAVA_HOME=/usr/lib/jvm/java-8-openjdk-arm64
export GZOLTAR_AGENT_JAR="/gzoltar/com.gzoltar.agent.rt/target/com.gzoltar.agent.rt-1.7.3-SNAPSHOT-all.jar"
export GZOLTAR_CLI_JAR="/gzoltar/com.gzoltar.cli/target/com.gzoltar.cli-1.7.3-SNAPSHOT-jar-with-dependencies.jar"
export TZ='America/Los_Angeles'
export LC_ALL=en_US.UTF-8
export LANG=en_US.UTF-8
export LANGUAGE=en_US.UTF-8

export D4J_HOME="/defects4j"
proj_dir="/defects4j/checkouts/${PID}_${BID}_buggy"

cd "$proj_dir"
test_classpath=$($D4J_HOME/framework/bin/defects4j export -p cp.test)
src_classes_dir=$($D4J_HOME/framework/bin/defects4j export -p dir.bin.classes)
src_classes_dir="$proj_dir/$src_classes_dir"
test_classes_dir=$($D4J_HOME/framework/bin/defects4j export -p dir.bin.tests)
test_classes_dir="$proj_dir/$test_classes_dir"
echo "${PID}_${BID}_buggy's classpath: $test_classpath" >&2
echo "${PID}_${BID}_buggy's bin dir: $src_classes_dir" >&2
echo "${PID}_${BID}_buggy's test bin dir: $test_classes_dir" >&2

cd "$proj_dir"
unit_tests_file="$proj_dir/unit_tests.txt"
relevant_tests="*"

java -cp "$D4J_HOME/framework/projects/lib/junit-4.11.jar:$test_classpath:$test_classes_dir:$GZOLTAR_CLI_JAR" \
  com.gzoltar.cli.Main listTestMethods \
    "$test_classes_dir" \
    --outputFile "$unit_tests_file" \
    --includes "$relevant_tests"

echo ">>>>> Unit Test File!"
head "$unit_tests_file"

cd "$proj_dir"
loaded_classes_file="$D4J_HOME/framework/projects/$PID/loaded_classes/$BID.src"
normal_classes=$(cat "$loaded_classes_file" | sed 's/$/:/' | sed ':a;N;$!ba;s/\\n//g')
inner_classes=$(cat "$loaded_classes_file" | sed 's/$/$*:/' | sed ':a;N;$!ba;s/\\n//g')
classes_to_debug="$normal_classes$inner_classes"
echo "Likely faulty classes: $classes_to_debug" >&2

ser_file="$proj_dir/gzoltar.ser"
export _JAVA_OPTIONS="-Xmx6144M -XX:MaxHeapSize=4096M"
java -javaagent:$GZOLTAR_AGENT_JAR=destfile=$ser_file,buildlocation=$src_classes_dir,includes=$classes_to_debug,excludes="",inclnolocationclasses=false,output="FILE" \
 -cp "$src_classes_dir:$D4J_HOME/framework/projects/lib/junit-4.11.jar:$test_classpath:$GZOLTAR_CLI_JAR" \
 com.gzoltar.cli.Main runTestMethods \
 --testMethods "$unit_tests_file" \
 --collectCoverage

cd "$proj_dir"
java -cp "$src_classes_dir:$D4J_HOME/framework/projects/lib/junit-4.11.jar:$test_classpath:$GZOLTAR_CLI_JAR" \
 com.gzoltar.cli.Main faultLocalizationReport \
 --buildLocation "$src_classes_dir" \
 --granularity "line" \
 --inclPublicMethods \
 --inclStaticConstructors \
      --inclDeprecatedMethods \
      --dataFile "$ser_file" \
      --outputDirectory "$proj_dir" \
      --family "sfl" \
      --formula "ochiai" \
      --metric "entropy" \
      --formatter "txt"

if [ $? -ne 0 ]; then
    echo "Error: Failed to list test methods."
    exit 1
fi

3. check the sfl/txt.ochiai.ranking.csv

Expected behaviour

As reported in https://bitbucket.org/rjust/fault-localization-data/src/master/analysis/pipeline-scripts/buggy-lines/Chart-1.buggy.lines, the suspicious line should be 1797
Howeve, sfl/txt.ochiai.ranking.csv shows that the faulty line is 1793.

Environment (please complete the following information, if relevant):

• Linux

Additional comments

I'm wondering what causes such a gap. Is it natural, or do I make some mistakes? Thanks!