mirror of https://github.com/python/cpython.git
119 lines
4.1 KiB
Python
119 lines
4.1 KiB
Python
# Module and documentation by Eric S. Raymond, 21 Dec 1998
|
|
|
|
import sys
|
|
|
|
class shlex:
|
|
"A lexical analyzer class for simple shell-like syntaxes."
|
|
def __init__(self, instream=None):
|
|
if instream:
|
|
self.instream = instream
|
|
else:
|
|
self.instream = sys.stdin
|
|
self.commenters = '#'
|
|
self.wordchars = 'abcdfeghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ0123456789_'
|
|
self.whitespace = ' \t\r\n'
|
|
self.quotes = '\'"'
|
|
self.state = ' '
|
|
self.pushback = [];
|
|
self.lineno = 1
|
|
self.debug = 0
|
|
self.token = ''
|
|
|
|
def push_token(self, tok):
|
|
"Push a token onto the stack popped by the get_token method"
|
|
if (self.debug >= 1):
|
|
print "Pushing " + tok
|
|
self.pushback = [tok] + self.pushback;
|
|
|
|
def get_token(self):
|
|
"Get a token from the input stream (or from stack if it's monempty)"
|
|
if self.pushback:
|
|
tok = self.pushback[0]
|
|
self.pushback = self.pushback[1:]
|
|
if (self.debug >= 1):
|
|
print "Popping " + tok
|
|
return tok
|
|
tok = ''
|
|
while 1:
|
|
nextchar = self.instream.read(1);
|
|
if nextchar == '\n':
|
|
self.lineno = self.lineno + 1
|
|
if self.debug >= 3:
|
|
print "In state " + repr(self.state) + " I see character: " + repr(nextchar)
|
|
if self.state == None:
|
|
return ''
|
|
elif self.state == ' ':
|
|
if not nextchar:
|
|
self.state = None; # end of file
|
|
break
|
|
elif nextchar in self.whitespace:
|
|
if self.debug >= 2:
|
|
print "I see whitespace in whitespace state"
|
|
if self.token:
|
|
break # emit current token
|
|
else:
|
|
continue
|
|
elif nextchar in self.commenters:
|
|
self.instream.readline()
|
|
self.lineno = self.lineno + 1
|
|
elif nextchar in self.wordchars:
|
|
self.token = nextchar
|
|
self.state = 'a'
|
|
elif nextchar in self.quotes:
|
|
self.token = nextchar
|
|
self.state = nextchar
|
|
else:
|
|
self.token = nextchar
|
|
if self.token:
|
|
break # emit current token
|
|
else:
|
|
continue
|
|
elif self.state in self.quotes:
|
|
self.token = self.token + nextchar
|
|
if nextchar == self.state:
|
|
self.state = ' '
|
|
break
|
|
elif self.state == 'a':
|
|
if not nextchar:
|
|
self.state = None; # end of file
|
|
break
|
|
elif nextchar in self.whitespace:
|
|
if self.debug >= 2:
|
|
print "I see whitespace in word state"
|
|
self.state = ' '
|
|
if self.token:
|
|
break # emit current token
|
|
else:
|
|
continue
|
|
elif nextchar in self.commenters:
|
|
self.instream.readline()
|
|
self.lineno = self.lineno + 1
|
|
elif nextchar in self.wordchars or nextchar in self.quotes:
|
|
self.token = self.token + nextchar
|
|
else:
|
|
self.pushback = [nextchar] + self.pushback
|
|
if self.debug >= 2:
|
|
print "I see punctuation in word state"
|
|
self.state = ' '
|
|
if self.token:
|
|
break # emit current token
|
|
else:
|
|
continue
|
|
|
|
result = self.token
|
|
self.token = ''
|
|
if self.debug >= 1:
|
|
print "Token: " + result
|
|
return result
|
|
|
|
if __name__ == '__main__':
|
|
|
|
lexer = shlex()
|
|
while 1:
|
|
tt = lexer.get_token()
|
|
if tt != None:
|
|
print "Token: " + repr(tt)
|
|
else:
|
|
break
|
|
|