#!/usr/bin/runhugs

-- numsort
-- 
-- Sort stdin by lines, except that sequences of digits are sniffed
-- out and compared numerically rather than lexically.  Useful for
-- sorting things like 'foo1, foo2, ..., foo10, foo11, ...' which
-- normal sort wouldn't get right.
-- 
-- Bugs: Reads only from stdin, not filenames specified.
--       Doesn't work reading from the console.
-- 
-- Todo: use the ReadS parsing stuff rather than chunking by hand
-- 
-- -- Ed Avis, ed@membled.com, 2000-06-09
-- 

module Main where

import List
import "/home/ed/lib/hs/Misc.hs"

main = unix_filter numsort

numsort :: String -> String
numsort = wrap (lines, unlines) $ sortBy cmp_lines
	  where cmp_lines a b = compare (parse_line a) (parse_line b)

-- chunk : split up a string into chunks which all satisfy a certain
-- property.  The property is that f (first char in chunk) (char) is
-- True.  So you might use some sort of equality test for f.
-- 
chunk :: (a -> a -> Bool) -> [a] -> [[a]]
chunk _ [] = []
chunk f (x:xs) = (x : c) : chunk f rest
		 where (c, rest) = span (f x) xs

same_digitness a b = (da && db) || ((not da) && (not db))
    where (da, db) = (isDigit a, isDigit b)

-- toIS: parse a String and convert it to a list of IS.  The string 
-- must be entirely digits (becomes a single IS), or entirely 
-- nondigits (each char becomes its own IS), and not empty.
-- 
toIS :: String -> [Either Char Int]
toIS s = let r = readDec s in
	     if (length r == 0) then
		map Left s
	     else if (length r == 1) then
		  let (i, rest) = head r in
		      if (length rest == 0) then
			 [ Right i ]
		      else error "trailing junk after digits"
	     else error "too many results from readDec"

parse_line :: String -> [Either Char Int]
parse_line = concat . map toIS . chunk same_digitness