1# -----------------------------------------------------------------------------
2# hedit.py
3#
4# Paring of Fortran H Edit descriptions (Contributed by Pearu Peterson)
5#
6# These tokens can't be easily tokenized because they are of the following
7# form:
8#
9#   nHc1...cn
10#
11# where n is a positive integer and c1 ... cn are characters.
12#
13# This example shows how to modify the state of the lexer to parse
14# such tokens
15# -----------------------------------------------------------------------------
16import sys
17if ".." not in sys.path: sys.path.insert(0,"..")
18
19import ply.lex as lex
20
21tokens = (
22    'H_EDIT_DESCRIPTOR',
23    )
24
25# Tokens
26t_ignore = " \t\n"
27
28def t_H_EDIT_DESCRIPTOR(t):
29    r"\d+H.*"                     # This grabs all of the remaining text
30    i = t.value.index('H')
31    n = eval(t.value[:i])
32
33    # Adjust the tokenizing position
34    t.lexer.lexpos -= len(t.value) - (i+1+n)
35    t.value = t.value[i+1:i+1+n]
36    return t
37
38def t_error(t):
39    print("Illegal character '%s'" % t.value[0])
40    t.lexer.skip(1)
41
42# Build the lexer
43lex.lex()
44lex.runmain(data="3Habc 10Habcdefghij 2Hxy")
45
46
47
48