History

For years I was GUI calculator fanatic. Desperately searching for that perfect calculator that did all the things I wanted it to do. I wanted a programmers calculator that did hex and binary operations and didn’t treat them as second class citizens. I used PCalc on the Mac for a while, and I even liked the built in Windows 95 calculator (that was the last time I had a dedicated Windows machine for working). But I was still left wanting.

At some point I decided to write my own. My idea was that I wanted a text entry box where the equation you were typing would go. Then you could edit the text and produce a new result without retyping the whole thing. I only got as far as being able to parse the text and convert it to reverse polish for easy evaluation before I gave up. But I kept the idea in the back of my mind for years.

At some point the idea morphed and I realized I didn’t need any of the GUI buttons to click on when I always had a perfectly good keyboard sitting in front of me. I looked into bc and dc but they were both too complex for me. Finally one day it hit me that Perl’s eval() was really what I wanted. And so “pc”, Perl Calc, was born.

pc started of very simple:

#!/usr/bin/perl

while (<>) {
    $a = eval($_);
    printf("%d 0x%x 0%o %f\n",$a,$a,$a,$a);
}

But it has now grown many features over the years designed to make my life more pleasant.

Getting PC

You can download it here or you can check it out from darcs like this:

~~darcs get http://porkrind.org/pc~~

Edit (2013-05-30): I am no longer using darcs for this project. It is now available on github.

You can give it input one of 2 ways: from the command line (this way you get all the editing and history niceties of your shell), or by running with no command line and typing into its stdin (this way you don’t have to quote characters your shell would like to steal).

How to use pc

The basics

$ pc 11+3
14 0xe 016

The first number is the integer result, followed by the hex and octal representations. Simple enough.

Order of operations

This shows that pc uses Perl’s order of operations (operator precedence if you are in a programming mood):

$ pc 1+3*2
7 0x7 07

ASCII

$ pc 1+3*20
61 0x3d 075 '='

Here we see an extra field was printed. In this case pc detected the final integer value was in the ASCII range and printed the character represented by the value 61, an equal sign.

Bitwise Operations

We also get Perl (and C) style bitwise operators:

$ pc '1+3*20<<2 & 0xff'
244 0xf4 0364

Also notice that we had to quote it since (a) I put a space in the expression and (b) I used the ‘<’ and ‘&’ characters which mean something to the shell.

Floating point

Of course it’s not restricted to only integers, it can handle floats too:

$ pc 1+3*20/55
2 0x2 02 2.0909090909090 23/11

You’ll notice it shows the result of the floating point math in addition to the integer math.

Fractions

You might have noticed a fraction in the output of the previous example. pc uses Perl’s “bigrat” library to do fractions:

$ pc 3/4+1/2
0 0x0 01 1.25 5/4

Human readable

$ pc 1000*2000
2,000,000 2000000 0x1e,8480 0x1e8480 07502200 1.90MB

You’ll notice that the integer and hex results are printed twice—one with commas and one without. The one with commas is so that you the human can read the output easily. The one without commas is for copying and pasting. You should also notice that pc helpfully told you the number was 1.90MB. If a number is bigger than 1024 (1KB) it will print it in human readable byte quantity.

Power of 2 magnitude suffixes

It also accepts magnitude suffixes on the input:

$ pc 16m
16,777,216 16777216 0x100,0000 0x1000000 0100000000 16MB

The following suffixes are allowed: kmgtpezy (lowercase, no b). Note that the human readable output “16MB” doesn’t have a decimal point. It will remove the decimal point if the number is exactly that value. So:

$ pc 16m+1
16,777,217 16777217 0x100,0001 0x1000001 0100000001 16.0MB

Since “16.0MB” has a decimal point, we know the value isn’t exactly 16 megabytes.

Large numbers

pc uses Perl’s bigint so that it can handle numbers bigger than 32 bits:

$ pc '<<40'
1,099,511,627,776 1099511627776 0x100,0000,0000 0x10000000000 020000000000000 1TB

Random Perl code

$ pc 'ord("a")'
0x61 0141 97 'a'

Since pc uses Perl’s eval you can do arbitrary perl code too. Though frankly the only thing I’ve ever used is ord().

Conclusion

So there you have it. It’s designed to give you all the output you ever could want without having to memorize any stupid command line switches. There are none, in fact. It also doesn’t overload you with redundant data. If the floating point answer is the same as the integer answer then it doesn’t show it to you. Same with the fractions.

So, if you read this far, go get it already!

First off, I don’t like Applescript. In fact, I hate it with a passion. You would think the Apple would learn from Cobol that natural language programming languages just suck. But no, Apple created the abomination called Applescript anyway and foisted it upon us and now sometimes you just have to suck it up and use it. Preferably from a distance with gloves on (and maybe a 10 foot pole) so that it doesn’t dirty your soul too much. But I digress.

I needed to call some Applescript from perl and was quite proud of my end result:

sub osascript($) { system 'osascript', map { ('-e', $_) } split(/\n/, $_[0]); }

The usage looks like this:

osascript <<END;
 tell application "Finder"
 display dialog "Hello"
 end tell
END

So basically the Applescript sits right inside your perl program and gets launched without needing any temporary files on the disk.

How does it work?

If the perl line is incomprehensible to you, let me break it down. The inner most expression is:

split(/\n/, $_[0]);

That takes the first argument to the function, breaks it apart at the line endings and puts the parts into a list. If we were to dump the result of the split using the dialog example above it would look like this:

(' tell application "Finder"', ' display dialog "Hello"', ' end tell')

Next, it uses the map function to insert “-” before every item in the list. It does that by using the fact that perl flattens arrays. The dump of the output of the map function conceptually looks like:

(('-e', ' tell application "Finder"'), ('-e', ' display dialog "Hello"'), ('-e', ' end tell'))

but really perl immediately flattens it to:

('-e', ' tell application "Finder"', '-e', ' display dialog "Hello"', '-e', ' end tell')

At that point it calls the system function with “osascript” and that last array as arguments. When system is passed a list instead of a string it runs the command with the arguments directly instead of launching a shell. This way all the arguments are effectively quoted correctly but without ever having to worry about quotes. This is an extremely clean way of launching commands. You really should never pass system (or open) a string: (a) it needlessly launches a shell, and (b) you have to worry about quotes.

In the end, I like that it fits on one line and that it lets the Applescript sit in the program in its native form.

Edit: Also check out my Ruby version.

Tag: perl

PC – Perl Calculator