CLI RPN Calculator

The command-line Reverse Polish Notation (RPN) calculator. Was developed as a test task.

Specifications

• The calculator should use standard input and standard output
• It should implement the four standard arithmetic operators
• The calculator should handle errors and recover gracefully
• The calculator should exit when it receives a q command or an end of input indicator (EOF / Ctrl+D)

py op

1. A^2 + B^2 = C^2 5 12 13 = true

Build and run

You need a JDK 13 or later and Maven. To build run mvn clean package. The executable jar will be built and placed into target/rpn-calc.jar. To run it use:

java -jar target/rpn-calc.jar

Then it will prompt for an input with a line started with > .

Usage

For example to calculate an expression 5 / (9 - 1) = 0.625 write each number (operand) and press Enter then write an arithmetic command (operator):

> 5
5
> 9
9
> 1
1
> -
8
> /
0.625
>

The last number is your result. You can use a single line input:

> 5 9 1 - /
0.625

To exit the program you can input command quit or simply q.

You can also use other commands for example debug or d to see the values stack and clear or c to remove everything from the stack:

> 1
1
> 2
2
> 3
3
> d
3
2
1
> c
> d
>

The calculator also have other functions from java.lang.Math and to see them use a command help or h:

> h
*
+
-
/
IEEEremainder
abs
acos
asin
atan
atan2
cbrt
ceil
copySign
cos
cosh
exp
expm1
floor
hypot
log
log10
log1p
max
min
neg
nextAfter
nextDown
nextUp
pow
rint
signum
sin
sinh
sqrt
tan
tanh
toDegrees
toRadians
ulp
>

For example to calculate square root and negate it:

> 4
4
> sqrt
2
> neg
-2
>

Implementation

The core of the calculator is Calc class which contains the values stack and performs all the calculations. The class designed in a JavaBeans style and have getters and setters so can be easily used with DI. The actual parsing and interaction with a user via the command line performed by the StreamProcessor class. You can extend the calc and add your custom operator by extending CalcUnaryOp or CalcBinaryOp. Then you'll need to add them into Ops.BASIC_OPS list or this can be done with DI.

Internally the calculator uses a double type so all the calculations have the same issues as IEEE-754 float numbers have.