class TreeSerializer(IterObject):
def __init__(self, tree):
self.max_look_behind = 10
super(TreeSerializer, self).__init__(tree)
def resetIter(self):
self.outData = LimLQueue(
self.max_look_behind) # limited record of yielded tokens
self.tokenStream = LQueue(self.inData.toListG())
self.tokenStreamIterable = iter(self.tokenStream)
super(TreeSerializer, self).resetIter()
##
# Peek n tokens behind
def lookbehind(self, n=1):
if n > self.max_look_behind:
raise SyntaxException(
"TokenStream: can only look %d elements behind" %
self.max_look_behind)
return self.outData[n]
##
# Peek n tokens ahead
#
# peek needs to circumvent __iter__ and access the LQueue directly
def peek(self, n=1):
toks = []
cnt = 0
# get the desired token
while cnt < n:
t = self.tokenStreamIterable.next()
toks.append(t)
if t['type'] == "eof":
break
cnt += 1
# put all retrieved tokens back
for t in toks[::-1]:
self.tokenStream.putBack(t)
return toks[-1]
def __iter__(self):
for tok in self.tokenStreamIterable:
self.outData.appendleft(tok)
yield tok
class TreeSerializer(IterObject):
def __init__(self, tree):
self.max_look_behind = 10
super(TreeSerializer, self).__init__(tree)
def resetIter(self):
self.outData = LimLQueue(self.max_look_behind) # limited record of yielded tokens
self.tokenStream = LQueue(self.inData.toListG())
self.tokenStreamIterable = iter(self.tokenStream)
super(TreeSerializer, self).resetIter()
##
# Peek n tokens behind
def lookbehind(self, n=1):
if n>self.max_look_behind:
raise SyntaxException("TokenStream: can only look %d elements behind" % self.max_look_behind)
return self.outData[n]
##
# Peek n tokens ahead
#
# peek needs to circumvent __iter__ and access the LQueue directly
def peek(self, n=1):
toks = []
cnt = 0
# get the desired token
while cnt < n:
t = self.tokenStreamIterable.next()
toks.append(t)
if t['type'] == "eof":
break
cnt += 1
# put all retrieved tokens back
for t in toks[::-1]:
self.tokenStream.putBack(t)
return toks[-1]
def __iter__(self):
for tok in self.tokenStreamIterable:
self.outData.appendleft(tok)
yield tok
def resetIter(self):
self.outData = LimLQueue(self.max_look_behind) # limited record of yielded tokens
self.tokenStream = LQueue(self.inData.toListG())
self.tokenStreamIterable = iter(self.tokenStream)
super(TreeSerializer, self).resetIter()
def resetIter(self):
self.outData = LimLQueue(
self.max_look_behind) # limited record of yielded tokens
self.tokenStream = LQueue(self.inData.toListG())
self.tokenStreamIterable = iter(self.tokenStream)
super(TreeSerializer, self).resetIter()
def resetIter(self):
self.tokenStream= LQueue(iter(self.inData))
self.tok_stream = iter(self.tokenStream)
super(TokenStream, self).resetIter()
class TokenStream(IterObject):
def __init__(self, inData):
self.line = 0
self.spos = 0 # current char pos
self.sol = 0 # last start-of-line char pos
super(TokenStream, self).__init__(inData)
def resetIter(self):
self.tokenStream= LQueue(iter(self.inData))
self.tok_stream = iter(self.tokenStream)
super(TokenStream, self).resetIter()
def peek(self, n=1):
"peek n tokens ahead"
toks = []
cnt = 0
# get the desired token
while cnt < n:
t = self.tok_stream.next()
toks.append(t)
if t['type'] == "eof":
break
while self._nonGrammaticalToken(t):
t = self.tok_stream.next()
toks.append(t)
cnt += 1
# put all retrieved tokens back
for t in toks[::-1]:
self.tokenStream.putBack(t)
return self._symbolFromToken(Token(toks[-1]))
def _nonGrammaticalToken(self, tok):
return tok['type'] in ['white', 'comment', 'eol']
def _symbolFromToken(self, tok):
s = None
# TODO: Stuff for another refac:
# The following huge dispatch could be avoided if the tokenizer already
# provided the tokens with the right attributes (esp. name, detail).
# tok isinstanceof Token()
if (tok.name == "white" # TODO: 'white'?!
or tok.name == 'comment'):
pass
elif tok.name == "eof":
symbol = symbol_table.get("eof")
s = symbol()
s.value = ""
elif tok.name == "eol":
self.line += 1 # increase line count
self.sol = tok.spos + tok.len # char pos of next line start
self.spos = tok.spos
pass # don't yield this (yet)
#s = symbol_table.get("eol")()
# 'operation' nodes
elif tok.detail in (
MULTI_TOKEN_OPERATORS
+ MULTI_PROTECTED_OPERATORS
+ SINGLE_RIGHT_OPERATORS
+ SINGLE_LEFT_OPERATORS
):
s = symbol_table[tok.value]()
s.type = "operation"
s.set('operator', tok.detail)
# 'assignment' nodes
elif tok.detail in ASSIGN_OPERATORS:
s = symbol_table[tok.value]()
s.type = "assignment"
s.set('operator', tok.detail)
# 'constant' nodes
elif tok.name in ('number', 'string', 'regexp'):
symbol = symbol_table["constant"]
s = symbol()
s.set('value', tok.value)
if tok.name == 'number':
s.set('constantType', 'number')
s.set('detail', tok.detail)
elif tok.name == 'string':
s.set('constantType', 'string')
s.set('detail', tok.detail)
elif tok.name == 'regexp':
s.set('constantType', 'regexp')
elif tok.name in ('reserved',) and tok.detail in ("TRUE", "FALSE"):
symbol = symbol_table["constant"]
s = symbol()
s.set('value', tok.value)
s.set('constantType', 'boolean')
elif tok.name in ('name',):
s = symbol_table["identifier"]()
s.set('value', tok.value)
else:
# TODO: token, reserved, builtin
# name or operator
symbol = symbol_table.get(tok.value)
if symbol:
s = symbol()
s.value = tok.value
else:
# don't make assumptions about correct tokens here, as we might be in the
# middle of a regexp
#raise SyntaxError("Unknown operator %r (pos %d)" % (tok.value, tok.spos))
symbol = symbol_table['(unknown)']
s = symbol()
s.value = tok.value
if s:
s.value = tok.value
s.set('column', tok.column)
s.set('line', tok.line)
# SPY-POINT
#print tok
return s
##
# yields syntax nodes as "tokens" (kind of a misnomer)
def __iter__(self):
for t in self.tok_stream:
tok = Token(t)
s = self._symbolFromToken(tok)
if not s:
continue
self.spos = tok.spos
s.spos = tok.spos
yield s
def resetIter(self):
self.tokenStream= LQueue(iter(self.inData))
self.tok_stream = iter(self.tokenStream)
super(TokenStream, self).resetIter()
class TokenStream(IterObject):
def __init__(self, inData):
self.line = 0
self.spos = 0 # current char pos
self.sol = 0 # last start-of-line char pos
super(TokenStream, self).__init__(inData)
def resetIter(self):
self.tokenStream= LQueue(iter(self.inData))
self.tok_stream = iter(self.tokenStream)
super(TokenStream, self).resetIter()
def peek(self, n=1):
"peek n tokens ahead"
toks = []
cnt = 0
# get the desired token
while cnt < n:
t = self.tok_stream.next()
toks.append(t)
if t['type'] == "eof":
break
while self._nonGrammaticalToken(t):
t = self.tok_stream.next()
toks.append(t)
cnt += 1
# put all retrieved tokens back
for t in toks[::-1]:
self.tokenStream.putBack(t)
return self._symbolFromToken(Token(toks[-1]))
def _nonGrammaticalToken(self, tok):
return tok['type'] in ['white', 'comment', 'eol']
def _symbolFromToken(self, tok):
s = None
# TODO: Stuff for another refac:
# The following huge dispatch could be avoided if the tokenizer already
# provided the tokens with the right attributes (esp. name, detail).
# tok isinstanceof Token()
if (tok.name == "white" # TODO: 'white'?!
or tok.name == 'comment'):
pass
elif tok.name == "eof":
symbol = symbol_table.get("eof")
s = symbol()
s.value = ""
elif tok.name == "eol":
self.line += 1 # increase line count
self.sol = tok.spos + tok.len # char pos of next line start
self.spos = tok.spos
pass # don't yield this (yet)
#s = symbol_table.get("eol")()
# 'operation' nodes
elif tok.detail in (
MULTI_TOKEN_OPERATORS
+ MULTI_PROTECTED_OPERATORS
+ SINGLE_RIGHT_OPERATORS
+ SINGLE_LEFT_OPERATORS
):
s = symbol_table[tok.value]()
s.type = "operation"
s.set('operator', tok.detail)
# 'assignment' nodes
elif tok.detail in ASSIGN_OPERATORS:
s = symbol_table[tok.value]()
s.type = "assignment"
s.set('operator', tok.detail)
# 'constant' nodes
elif tok.name in ('number', 'string', 'regexp'):
symbol = symbol_table["constant"]
s = symbol()
s.set('value', tok.value)
if tok.name == 'number':
s.set('constantType', 'number')
s.set('detail', tok.detail)
elif tok.name == 'string':
s.set('constantType', 'string')
s.set('detail', tok.detail)
elif tok.name == 'regexp':
s.set('constantType', 'regexp')
elif tok.name in ('reserved',) and tok.detail in ("TRUE", "FALSE"):
symbol = symbol_table["constant"]
s = symbol()
s.set('value', tok.value)
s.set('constantType', 'boolean')
elif tok.name in ('name',):
s = symbol_table["identifier"]()
s.set('value', tok.value)
else:
# TODO: token, reserved, builtin
# name or operator
symbol = symbol_table.get(tok.value)
if symbol:
s = symbol()
s.value = tok.value
else:
# don't make assumptions about correct tokens here, as we might be in the
# middle of a regexp
#raise SyntaxError("Unknown operator %r (pos %d)" % (tok.value, tok.spos))
symbol = symbol_table['(unknown)']
s = symbol()
s.value = tok.value
if s:
s.value = tok.value
s.set('column', tok.column)
s.set('line', tok.line)
# SPY-POINT
#print tok
return s
##
# yields syntax nodes as "tokens" (kind of a misnomer)
def __iter__(self):
for t in self.tok_stream:
tok = Token(t)
s = self._symbolFromToken(tok)
if not s:
continue
self.spos = tok.spos
s.spos = tok.spos
yield s
