Guido van Rossum | 0af9a28 | 1992-01-01 18:38:21 +0000 | [diff] [blame] | 1 | #! /ufs/guido/bin/sgi/python |
| 2 | #! /usr/local/python |
| 3 | |
Guido van Rossum | 318a91c | 1992-01-01 19:22:25 +0000 | [diff] [blame^] | 4 | # Fix Python source files to use the new equality test operator, i.e., |
Guido van Rossum | 0af9a28 | 1992-01-01 18:38:21 +0000 | [diff] [blame] | 5 | # if x = y: ... |
Guido van Rossum | 318a91c | 1992-01-01 19:22:25 +0000 | [diff] [blame^] | 6 | # is changed to |
| 7 | # if x == y: ... |
| 8 | # The script correctly tokenizes the Python program to reliably |
| 9 | # distinguish between assignments and equality tests. |
Guido van Rossum | 0af9a28 | 1992-01-01 18:38:21 +0000 | [diff] [blame] | 10 | # |
| 11 | # Command line arguments are files or directories to be processed. |
| 12 | # Directories are searched recursively for files whose name looks |
| 13 | # like a python module. |
| 14 | # Symbolic links are always ignored (except as explicit directory |
| 15 | # arguments). Of course, the original file is kept as a back-up |
| 16 | # (with a "~" attached to its name). |
| 17 | # |
| 18 | # Changes made are reported to stdout in a diff-like format. |
| 19 | # |
| 20 | # Undoubtedly you can do this using find and sed or perl, but this is |
| 21 | # a nice example of Python code that recurses down a directory tree |
| 22 | # and uses regular expressions. Also note several subtleties like |
| 23 | # preserving the file's mode and avoiding to even write a temp file |
| 24 | # when no changes are needed for a file. |
| 25 | # |
| 26 | # NB: by changing only the function fixline() you can turn this |
| 27 | # into a program for a different change to Python programs... |
| 28 | |
| 29 | import sys |
| 30 | import regex |
| 31 | import posix |
| 32 | import path |
| 33 | from stat import * |
| 34 | |
| 35 | err = sys.stderr.write |
| 36 | dbg = err |
| 37 | rep = sys.stdout.write |
| 38 | |
| 39 | def main(): |
| 40 | bad = 0 |
| 41 | if not sys.argv[1:]: # No arguments |
| 42 | err('usage: ' + argv[0] + ' file-or-directory ...\n') |
| 43 | sys.exit(2) |
| 44 | for arg in sys.argv[1:]: |
| 45 | if path.isdir(arg): |
| 46 | if recursedown(arg): bad = 1 |
| 47 | elif path.islink(arg): |
| 48 | err(arg + ': will not process symbolic links\n') |
| 49 | bad = 1 |
| 50 | else: |
| 51 | if fix(arg): bad = 1 |
| 52 | sys.exit(bad) |
| 53 | |
| 54 | ispythonprog = regex.compile('^[a-zA-Z0-9_]+\.py$') |
| 55 | def ispython(name): |
| 56 | return ispythonprog.match(name) >= 0 |
| 57 | |
| 58 | def recursedown(dirname): |
| 59 | dbg('recursedown(' + `dirname` + ')\n') |
| 60 | bad = 0 |
| 61 | try: |
| 62 | names = posix.listdir(dirname) |
| 63 | except posix.error, msg: |
| 64 | err(dirname + ': cannot list directory: ' + `msg` + '\n') |
| 65 | return 1 |
| 66 | names.sort() |
| 67 | subdirs = [] |
| 68 | for name in names: |
| 69 | if name in ('.', '..'): continue |
| 70 | fullname = path.join(dirname, name) |
| 71 | if path.islink(fullname): pass |
| 72 | elif path.isdir(fullname): |
| 73 | subdirs.append(fullname) |
| 74 | elif ispython(name): |
| 75 | if fix(fullname): bad = 1 |
| 76 | for fullname in subdirs: |
| 77 | if recursedown(fullname): bad = 1 |
| 78 | return bad |
| 79 | |
| 80 | def fix(filename): |
Guido van Rossum | 318a91c | 1992-01-01 19:22:25 +0000 | [diff] [blame^] | 81 | ## dbg('fix(' + `filename` + ')\n') |
Guido van Rossum | 0af9a28 | 1992-01-01 18:38:21 +0000 | [diff] [blame] | 82 | try: |
| 83 | f = open(filename, 'r') |
| 84 | except IOError, msg: |
| 85 | err(filename + ': cannot open: ' + `msg` + '\n') |
| 86 | return 1 |
| 87 | head, tail = path.split(filename) |
| 88 | tempname = path.join(head, '@' + tail) |
| 89 | g = None |
| 90 | # If we find a match, we rewind the file and start over but |
| 91 | # now copy everything to a temp file. |
| 92 | lineno = 0 |
| 93 | while 1: |
| 94 | line = f.readline() |
| 95 | if not line: break |
| 96 | lineno = lineno + 1 |
| 97 | while line[-2:] == '\\\n': |
| 98 | nextline = f.readline() |
| 99 | if not nextline: break |
| 100 | line = line + nextline |
| 101 | lineno = lineno + 1 |
| 102 | newline = fixline(line) |
| 103 | if newline != line: |
| 104 | if g is None: |
| 105 | try: |
| 106 | g = open(tempname, 'w') |
| 107 | except IOError, msg: |
| 108 | f.close() |
| 109 | err(tempname+': cannot create: '+\ |
| 110 | `msg`+'\n') |
| 111 | return 1 |
| 112 | f.seek(0) |
| 113 | lineno = 0 |
| 114 | rep(filename + ':\n') |
| 115 | continue # restart from the beginning |
| 116 | rep(`lineno` + '\n') |
| 117 | rep('< ' + line) |
| 118 | rep('> ' + newline) |
| 119 | if g is not None: |
| 120 | g.write(newline) |
| 121 | |
| 122 | # End of file |
| 123 | f.close() |
| 124 | if not g: return 0 # No changes |
| 125 | |
| 126 | # Finishing touch -- move files |
| 127 | |
| 128 | # First copy the file's mode to the temp file |
| 129 | try: |
| 130 | statbuf = posix.stat(filename) |
| 131 | posix.chmod(tempname, statbuf[ST_MODE] & 07777) |
| 132 | except posix.error, msg: |
| 133 | err(tempname + ': warning: chmod failed (' + `msg` + ')\n') |
| 134 | # Then make a backup of the original file as filename~ |
| 135 | try: |
| 136 | posix.rename(filename, filename + '~') |
| 137 | except posix.error, msg: |
| 138 | err(filename + ': warning: backup failed (' + `msg` + ')\n') |
| 139 | # Now move the temp file to the original file |
| 140 | try: |
| 141 | posix.rename(tempname, filename) |
| 142 | except posix.error, msg: |
| 143 | err(filename + ': rename failed (' + `msg` + ')\n') |
| 144 | return 1 |
| 145 | # Return succes |
| 146 | return 0 |
| 147 | |
Guido van Rossum | 318a91c | 1992-01-01 19:22:25 +0000 | [diff] [blame^] | 148 | |
| 149 | from tokenize import tokenprog |
| 150 | |
| 151 | match = {'if':':', 'elif':':', 'while':':', 'return':'\n', \ |
| 152 | '(':')', '[':']', '{':'}', '`':'`'} |
Guido van Rossum | 0af9a28 | 1992-01-01 18:38:21 +0000 | [diff] [blame] | 153 | |
| 154 | def fixline(line): |
Guido van Rossum | 318a91c | 1992-01-01 19:22:25 +0000 | [diff] [blame^] | 155 | # Quick check for easy case |
| 156 | if '=' not in line: return line |
| 157 | |
| 158 | i, n = 0, len(line) |
| 159 | stack = [] |
| 160 | while i < n: |
| 161 | j = tokenprog.match(line, i) |
| 162 | if j < 0: |
| 163 | # A bad token; forget about the rest of this line |
| 164 | print '(Syntax error:)' |
| 165 | print line, |
| 166 | return line |
| 167 | a, b = tokenprog.regs[3] # Location of the token proper |
| 168 | token = line[a:b] |
| 169 | i = i+j |
| 170 | if stack and token == stack[-1]: |
| 171 | del stack[-1] |
| 172 | elif match.has_key(token): |
| 173 | stack.append(match[token]) |
| 174 | elif token == '=' and stack: |
| 175 | line = line[:a] + '==' + line[b:] |
| 176 | i, n = a + len('=='), len(line) |
| 177 | elif token == '==' and not stack: |
| 178 | print '(Warning: \'==\' at top level:)' |
| 179 | print line, |
Guido van Rossum | 0af9a28 | 1992-01-01 18:38:21 +0000 | [diff] [blame] | 180 | return line |
| 181 | |
Guido van Rossum | 318a91c | 1992-01-01 19:22:25 +0000 | [diff] [blame^] | 182 | |
Guido van Rossum | 0af9a28 | 1992-01-01 18:38:21 +0000 | [diff] [blame] | 183 | main() |