#! /usr/bin/env python

import difflib, getopt, os, re, stat, sys, time

PROG_NAME = "srep"
PROG_VERSION = "0.01 (2005/10/02)"
PROG_COPYRIGHT = "Copyright (C) Laurence Tratt 2003-2005"

PATCH_NONE = 0
PATCH_UNIFIED = 1
PATCH_NDIFF = 2





class SREP:

	def __init__(self):

		opts, args = getopt.getopt(sys.argv[1 : ], "rivnuh")
	
		self._recurse = 0
		self._interactive = 0
		self._verbose = 0
		self._produce_patch = PATCH_NONE
		for opt in opts:
			if opt[0] == "-r":
				self._recurse = 1
			elif opt[0] == "-i":
				self._interactive = 1
			elif opt[0] == "-v":
				self._verbose = 1
			elif opt[0] == "-n":
				self._produce_patch = PATCH_NDIFF
			elif opt[0] == "-u":
				self._produce_patch = PATCH_UNIFIED
			elif opt[0] == "-h":
				self.usage()
			else:
				self.usage("Error: Unknown arg '%s'" % opt[0])

		if len(args) < 3:
			self.usage("Insufficient number of args")

		self._pattern = re.compile(args[0], re.DOTALL | re.MULTILINE)
		self._replace = args[1]
		
		self.log("Replacing '%s' with '%s' in %s.\n" % (sys.argv[1], self._replace, args[2 : ]))

		for file_name in args[2 : ]:
			if os.path.isdir(file_name):
				self.recurse(file_name)
			else:
				self.process_file(file_name)



	def usage(self, msg = None):

		if msg is None:
			msg = ""
		else:
			msg = msg + "\n\n"

		sys.stderr.write("""%s%s [-rivnu] <regular expression> <replacement> <file_1> ... <file_n>\n""" % (msg, PROG_NAME))
		sys.exit(1)



	def log(self, msg):
	
		if self._verbose:
			sys.stderr.write("%s" % msg)



	def recurse(self, dir_name):

		self.log("Recursing into %s.\n" % dir_name)
		for leaf_name in os.listdir(dir_name):
			file_name = os.path.join(dir_name, leaf_name)
			if os.path.isdir(file_name):
				self.recurse(file_name)
				continue
			else:
				self.process_file(file_name)



	def process_file(self, file_name):

		self.log("Processing %s.\n" % file_name)
		
		file = open(file_name, "r")
		file_data = file.read(-1)
		file.close()
		if self._produce_patch:
			original_file_data = file_data
		
		i = 0
		changed = 0
		while 1:
			match = self._pattern.search(file_data, i)
			if not match:
				break
			file_data = file_data[ : match.start()] + self._replace + file_data[match.end() : ]
			i = match.start() + len(self._replace)
			changed += 1

		file_mdate = time.ctime(os.stat(file_name)[stat.ST_MTIME])
		if changed:
			self.log("(%s replace operations).\n" % `changed`)
			if self._produce_patch == PATCH_NDIFF:
				line = "--- %s " % file_name
				line += " " * (15 - len(line))
				line += file_mdate
				print line
				line = "+++ %s " % file_name
				line += " " * (15 - len(line))
				line += file_mdate
				print line
				ndiff = ''.join(difflib.ndiff(original_file_data.splitlines(1), file_data.splitlines(1)))
				print self.ndiff_to_unified_diff(ndiff)
			elif self._produce_patch == PATCH_UNIFIED:
				print ''.join(difflib.unified_diff(original_file_data.splitlines(1), file_data.splitlines(1), file_name, file_name, file_mdate, file_mdate))
			else:
				file = open(file_name, "w")
				file.write(file_data)
				file.close()
		else:
			self.log("(not changed).\n")



	def ndiff_to_unified_diff(self, ndiff):
	
		output = []
		
		# ndiff adds an extra space after the "+- " chars which confuses patch. They need to be
		# removed; we also strip the extra space after "?" to ensure that things line up
		# correctly.
		
		tlines = ndiff.splitlines(1)
		i = 0
		lines = []
		while i < len(tlines):
			if tlines[i][0 : 2] == "+ ":
				lines.append("+" + tlines[i][2 : ])
			elif tlines[i][0 : 2] == "- ":
				lines.append("-" + tlines[i][2 : ])
			elif tlines[i][0 : 2] == "  ":
				lines.append(" " + tlines[i][2 : ])
			elif tlines[i][0 : 2] == "? ":
				lines.append("?" + tlines[i][2 : ])
			else:
				lines.append(tlines[i])
			i += 1

		i = 0
		last = 0
		in_real_line = 1
		out_real_line = 1
		while i < len(lines):
			if lines[i][0] in ["+", "-"]:
				j = i - 1
				while j > i - 3 and lines[j][0] == " ":
					j -= 1
				context = i - j
				while in_real_line - context < 0 or out_real_line - context < 0:
					context -= 1

				chunk = []
				chunk.extend(lines[i - context : i])
				original_in_real_line = in_real_line
				original_out_real_line = out_real_line
				while i < len(lines) and lines[i][0] in ["+", "-", "?"]:
					if lines[i][0] == "+":
						out_real_line += 1
					elif lines[i][0] == "-":
						in_real_line += 1
					chunk.append(lines[i])
					i += 1
				while i < len(lines):
					j = i
					while i < len(lines) and i < j + 3 and lines[i][0] == " ":
						in_real_line += 1
						out_real_line += 1
						chunk.append(lines[i])
						i += 1
					if not (i < len(lines) and i < j + 3):
						k = i
						while k < len(lines) and k < j + 6 and lines[k][0] == " ":
							k += 1
						if k == j + 6:
							break
						j = i
						while i < len(lines) and i < j + 3 and lines[i][0] == " ":
							in_real_line += 1
							out_real_line += 1
							chunk.append(lines[i])
							i += 1
					while i < len(lines) and lines[i][0] != " ":
						if lines[i][0] == "+":
							out_real_line += 1
						elif lines[i][0] == "-":
							in_real_line += 1
						chunk.append(lines[i])
						i += 1
				output.append("@@ -%s,%s +%s,%s @@\n" % (original_in_real_line - context, in_real_line - original_in_real_line + context, original_out_real_line - context, out_real_line - original_out_real_line + context))
				output.extend(chunk)
				last = in_real_line
			elif len(lines[i]) > 2 and lines[i][0] == " ":
				in_real_line += 1
				out_real_line += 1
				i += 1
			else:
				i += 1
		
		return "".join(output)
			



if __name__ == "__main__":
	SREP()