makeqstrdata.py 3.53 KB
Newer Older
1
2
from __future__ import print_function

3
4
import argparse
import re
5
import sys
6
7
8
9
10
11
12

# codepoint2name is different in Python 2 to Python 3
import platform
if platform.python_version_tuple()[0] == '2':
    from htmlentitydefs import codepoint2name
elif platform.python_version_tuple()[0] == '3':
    from html.entities import codepoint2name
13
codepoint2name[ord('-')] = 'hyphen';
14

15
16
# add some custom names to map characters that aren't in HTML
codepoint2name[ord('.')] = 'dot'
17
18
codepoint2name[ord(':')] = 'colon'
codepoint2name[ord('/')] = 'slash'
Damien George's avatar
Damien George committed
19
codepoint2name[ord('%')] = 'percent'
20
codepoint2name[ord('#')] = 'hash'
21
22
codepoint2name[ord('{')] = 'brace_open'
codepoint2name[ord('}')] = 'brace_close'
23
codepoint2name[ord('*')] = 'star'
24

25
26
# this must match the equivalent function in qstr.c
def compute_hash(qstr):
27
    hash = 5381
28
    for char in qstr:
29
        hash = (hash * 33) ^ ord(char)
30
31
    return hash & 0xffff

32
# given a list of (name,regex) pairs, find the first one that matches the given line
33
def re_match_first(regexs, line):
34
35
36
37
38
39
    for name, regex in regexs:
        match = re.match(regex, line)
        if match:
            return name, match
    return None, None

40
41
42
43
44
45
46
# regexs to recognise lines that the CPP emits
# use a list so that matching order is honoured
cpp_regexs = [
    ('qstr', r'Q\((.+)\)$'),
    ('cdecl', r'(typedef|extern) [A-Za-z0-9_* ]+;$')
]

47
48
def do_work(infiles):
    # read the qstrs in from the input files
49
    qstrs = {}
50
51
52
53
54
55
56
    for infile in infiles:
        with open(infile, 'rt') as f:
            line_number = 0
            for line in f:
                line_number += 1
                line = line.strip()

57
58
                # ignore blank lines, comments and preprocessor directives
                if len(line) == 0 or line.startswith('//') or line.startswith('#'):
59
60
                    continue

61
                # work out what kind of line it is
62
                match_kind, match = re_match_first(cpp_regexs, line)
63
64
                if match_kind is None:
                    # unknown line format
65
                    print('({}:{}) bad qstr format, got {}'.format(infile, line_number, line), file=sys.stderr)
66
                    return False
67
68
69
                elif match_kind != 'qstr':
                    # not a line with a qstr
                    continue
70
71
72

                # get the qstr value
                qstr = match.group(1)
73
                ident = re.sub(r'[^A-Za-z0-9_]', lambda s: "_" + codepoint2name[ord(s.group(0))] + "_", qstr)
74
75

                # don't add duplicates
76
                if ident in qstrs:
77
78
                    continue

79
                # add the qstr to the list, with order number to retain original order in file
80
                qstrs[ident] = (len(qstrs), ident, qstr)
81
82
83

    # process the qstrs, printing out the generated C header file
    print('// This file was automatically generated by makeqstrdata.py')
84
    print('')
85
    for order, ident, qstr in sorted(qstrs.values(), key=lambda x: x[0]):
86
87
        qhash = compute_hash(qstr)
        qlen = len(qstr)
88
        print('Q({}, (const byte*)"\\x{:02x}\\x{:02x}\\x{:02x}\\x{:02x}" "{}")'.format(ident, qhash & 0xff, (qhash >> 8) & 0xff, qlen & 0xff, (qlen >> 8) & 0xff, qstr))
89
90
91
92
93
94
95
96
97
98

    return True

def main():
    arg_parser = argparse.ArgumentParser(description='Process raw qstr file and output qstr data with length, hash and data bytes')
    arg_parser.add_argument('files', nargs='+', help='input file(s)')
    args = arg_parser.parse_args()

    result = do_work(args.files)
    if not result:
99
        print('exiting with error code', file=sys.stderr)
100
101
102
103
        exit(1)

if __name__ == "__main__":
    main()