Replace method calls with original checks using constants

[dlmarion]

I removed the methods from ByteUtils and reverted the places where the methods were called back to their original logic, but retained the use of the ByteUtils static variables. This yielded no significant performance difference. I re-ran the benchmarks on main before #27 was merged, the HEAD of main, and this branch. I ran mvn clean verify on each branch and ran the benchmark using the command:

taskset -c 1 mvn exec:exec -Dexec.executable="java" -Dexec.classpathScope=test -Dexec.args="-classpath %classpath org.apache.accumulo.access.AccessExpressionBenchmark"

Here are the results:

main @c959026c8631da8691764b3f20a9148b84bafc99:
AccessExpressionBenchmark.measureBytesParsing          thrpt   12  12.665 ± 0.225  ops/us
AccessExpressionBenchmark.measureEvaluation            thrpt   12  16.663 ± 0.441  ops/us
AccessExpressionBenchmark.measureEvaluationAndParsing  thrpt   12   8.435 ± 0.117  ops/us
AccessExpressionBenchmark.measureStringParsing         thrpt   12  11.709 ± 0.225  ops/us

main:
AccessExpressionBenchmark.measureBytesParsing          thrpt   12  12.864 ± 0.278  ops/us
AccessExpressionBenchmark.measureEvaluation            thrpt   12  16.849 ± 0.312  ops/us
AccessExpressionBenchmark.measureEvaluationAndParsing  thrpt   12   8.480 ± 0.094  ops/us
AccessExpressionBenchmark.measureStringParsing         thrpt   12  11.482 ± 0.033  ops/us


test-perf-fixes:
AccessExpressionBenchmark.measureBytesParsing          thrpt   12  12.555 ± 0.196  ops/us
AccessExpressionBenchmark.measureEvaluation            thrpt   12  16.651 ± 0.730  ops/us
AccessExpressionBenchmark.measureEvaluationAndParsing  thrpt   12   8.432 ± 0.094  ops/us
AccessExpressionBenchmark.measureStringParsing         thrpt   12  11.601 ± 0.209  ops/us

[dlmarion]

@DomGarguilo - I'm going to close this I think as there is no value-add in these changes. I did find a place (AccessEvaluatorImpl::71) where a literal is being used instead of the static variable.