def _startswith_tagtable_rest_of_line(text):
return (
# Ensure the text starts with the given word
("begin", TT.Word, text, TT.MatchFail, +1),
# Read to the end of line
(None, TT.AllInSet, TT.invset('\r\n'), +1, +1),
# Read the end of line
(None, TT.Is, '\n', +1, +4), # matches '\n' or
(None, TT.Is, '\r', +2, +1), # '\r' followed by
(None, TT.Is, '\n', +2, +2), # optional '\n'
# Check if EOF (allow EOF if no EOL found)
(None, TT.EOF, TT.Here, +1, TT.MatchOk),
# Not EOF, so look for the next line starting with text
("begin", TT.Word, text, +1, -5),
# Not what I am looking for, so read to the end of line
(None, TT.AllInSet, TT.invset('\r\n'), +1, +1),
# Read the end of line then test the next line
(None, TT.Is, '\n', +1, -2), # '\n'
(None, TT.Is, '\r', +2, +1), # '\r' followed by
(None, TT.Is, '\n', -4, -4), # optional '\n'
# Allow termination at EOF
(None, TT.EOF, TT.Here, TT.MatchFail, TT.MatchOk),
)
def _endswith_tagtable_rest_of_line(text):
return (
# Is the current line the end of record marker?
(None, TT.Word, text, +8, +1),
# Read whatever else is on that line (could be nothing)
(None, TT.AllInSet, TT.invset('\r\n'), +1, +1),
# Get the end of line
("end", TT.Is, '\n', +1, -2), # matches '\n'
(None, TT.Is, '\r', +4, +1),
("end", TT.Is, '\n', +1, -4),
(None, TT.Skip, -1, +1, +1),
("end", TT.Skip, +1, -6, -6),
# Check if EOF (only tests when the end of record line has no \n)
# Only time this should fail is with a bug in TT.
("end", TT.EOF, TT.Here, TT.MatchFail, TT.MatchOk),
# Not the end of record marker, so read to the end of line
(None, TT.AllInSet, TT.invset('\r\n'), +1, +1),
# Check if EOF
(None, TT.EOF, TT.Here, +1, TT.MatchOk),
# Not EOF, so scarf any newlines and try again
(None, TT.AllInSet, TT.set('\r\n'), TT.MatchFail, -10),
)
def _startswith_tagtable_newline(text):
return (
# Ensure the text starts with the given word ...
("begin", TT.Word, text, TT.MatchFail, +1),
# ... followed by the end of line
(None, TT.Is, '\n', +1, +4), # matches '\n' or
(None, TT.Is, '\r', +2, +1), # '\r' followed by
(None, TT.Is, '\n', +2, +2), # optional '\n'
# Check if EOF instead of a newline (allow EOF if found)
# Otherwise, this means the line starts with the text but
# doesn't have a successive newline.
# XXX BUG! When looking for "A\n" should not fail on "AA\n"!
(None, TT.EOF, TT.Here, TT.MatchFail, TT.MatchOk),
# Look for the next line starting with text
("begin", TT.Word, text, +1, -4),
# Not what I am looking for, so read to the end of line
(None, TT.AllInSet, TT.invset('\r\n'), +1, +1),
# Read the end of line then test the next line
(None, TT.Is, '\n', +1, -2), # '\n'
(None, TT.Is, '\r', +2, +1), # '\r' followed by
(None, TT.Is, '\n', -4, -4), # optional '\n'
# Allow termination at EOF
(None, TT.EOF, TT.Here, TT.MatchFail, TT.MatchOk),
)
def _endswith_tagtable_newline(text):
return (
# Is the current line the end of record marker?
(None, TT.Word, text, +6, +1),
# Make sure it ends the line
("end", TT.Is, '\n', +1, -1), # matches '\n'
(None, TT.Is, '\r', +4, +1),
("end", TT.Is, '\n', +1, -3),
(None, TT.Skip, -1, +1, +1),
("end", TT.Skip, +1, -5, -5),
# Not the end of record marker, so read to the end of line
(None, TT.AllInSet, TT.invset('\r\n'), +1, +1),
# Check if EOF
(None, TT.EOF, TT.Here, +1, TT.MatchOk),
# Not EOF, so scarf any newlines
(None, TT.AllInSet, TT.set('\r\n'), TT.MatchFail, -8),
)
fake = self.text + "\n"
reader = StartsWith(self.infile, self.text, self.sizehint,
fake + self.lookahead)
rec = reader.next()
rec = rec[len(fake):] # remove the fake data
self.infile, self.lookahead = reader.remainder()
self.found = 1
return rec
def remainder(self):
return self.infile, self.lookahead
# Tag the last byte of every newline
_tag_lines_tagtable = (
# Skip non-newline characters
(None, TT.AllInSet, TT.invset('\r\n'), +1, +1),
# Check if newline
("newline", TT.Is, '\n', +1, -1), # can be '\n'
(None, TT.Is, '\r', +3, +1), # or start a '\r' followed by ..
("newline", TT.Is, '\n', +1, -3), # .. an optional '\n'
("newline", TT.Skip, 0, -4, -4), # get here with just an '\r'
(None, TT.EOF, TT.Here, -5, TT.MatchOk), # stop at end of text
)
class CountLines(RecordReader):
"""Read a specified (fixed) number of lines"""
def __init__(self, infile, count, sizehint = SIZEHINT, lookahead = ""):
assert count > 0, "CountLines reader must read at least one line"
assert lookahead > 0, "Must read at least a character at a time"
def generate_dot(expression, genstate):
return [
(None, TT.IsInSet, TT.invset('\n')),
]
def generate_dot(expression, genstate):
return [(None, TT.IsInSet, TT.invset('\n')), ]
