🚀 go-pugleaf

In article <C184vr.p5q@austin.ibm.com> tbates@austin.ibm.com (Tom Bates)
writes:

>Quoting "The AWK Programming Language",
>
>	"The cleanest way to do case conversion in awk is with an array that
>maps each letter...it's better to use...tr"
>
>I often use tr for this very reason. For example:
>
>	echo "HeLlO, wOrLd!" | tr [:upper:] [:lower]
>
>This gives you:
>
>	hello, world!
>
>To use tr from awk:
>
>#!/bin/awk -f
>   {
>   "echo "$0" | tr [:upper:] [:lower:]" | getline
>   print
>}

Note that this will give squirelly results if handed any characters
special to /bin/sh, like quotes, or <>|, or backslash, or backtick,
or $, or ...

In order to do this generally, one would have to quote the argument to
echo in single quotes, e.g.:

  gsub(/'/, "'\\''" $0);
  "echo '"$0"' | tr [A-Z] [a-z]" | getline

Doing this inside a shell script would make it look particularly ugly.

This approach, while undeniably clever, has a few other drawbacks:

1) It's really slow, if it has to be done repeatedly, since we fork a
   new subshell for each string we want to downcase.

2) It's very Un*x specific.


tolower and toupper are such innocuously handy functions that I have 
trouble understanding why they weren't included in the original nawk.

Here, just for fun, is an alternative implementation of tolower, in awk.
It's probably slower than one based on a mapping array, but it was faster
for me to type:

BEGIN {
  Aa = "AaBbCcDdEeFfGgHhIiJjKkLlMmNnOoPpQqRrSsTtUuVvWwXxYyZz";
}

function tolower(str,     s, i, c) {
  i = 1;
  s = str;
  while (match(substr(s, i), /[A-Z]/))
    {
      i += RSTART - 1;
      c = substr(s, i, 1);
      c = substr(Aa, index(Aa,c) + 1, 1);
      s = substr(s, 1, i - 1) c substr(s, i + 1);
      i++;
    }
  return s;
}

--Radey
-- 
Radey Shouman					rshouman@chpc.utexas.edu