2015-05-30 22:11:16 +00:00
|
|
|
"""
|
|
|
|
Process raw qstr file and output qstr data with length, hash and data bytes.
|
|
|
|
|
|
|
|
This script works with Python 2.6, 2.7, 3.3 and 3.4.
|
|
|
|
"""
|
|
|
|
|
2014-03-10 07:07:35 +00:00
|
|
|
from __future__ import print_function
|
|
|
|
|
2014-01-21 21:40:13 +00:00
|
|
|
import re
|
2014-03-08 15:03:25 +00:00
|
|
|
import sys
|
2014-01-24 22:22:00 +00:00
|
|
|
|
2016-04-14 13:37:04 +00:00
|
|
|
# Python 2/3 compatibility:
|
|
|
|
# - iterating through bytes is different
|
|
|
|
# - codepoint2name lives in a different module
|
2014-01-24 22:22:00 +00:00
|
|
|
import platform
|
2020-02-27 04:36:53 +00:00
|
|
|
|
|
|
|
if platform.python_version_tuple()[0] == "2":
|
2016-09-02 04:32:47 +00:00
|
|
|
bytes_cons = lambda val, enc=None: bytearray(val)
|
2014-01-24 22:22:00 +00:00
|
|
|
from htmlentitydefs import codepoint2name
|
2020-02-27 04:36:53 +00:00
|
|
|
elif platform.python_version_tuple()[0] == "3":
|
2016-09-02 04:32:47 +00:00
|
|
|
bytes_cons = bytes
|
2014-01-24 22:22:00 +00:00
|
|
|
from html.entities import codepoint2name
|
2016-09-02 04:32:47 +00:00
|
|
|
# end compatibility code
|
|
|
|
|
2020-02-27 04:36:53 +00:00
|
|
|
codepoint2name[ord("-")] = "hyphen"
|
2014-01-21 21:40:13 +00:00
|
|
|
|
2014-02-15 11:34:50 +00:00
|
|
|
# add some custom names to map characters that aren't in HTML
|
2020-02-27 04:36:53 +00:00
|
|
|
codepoint2name[ord(" ")] = "space"
|
|
|
|
codepoint2name[ord("'")] = "squot"
|
|
|
|
codepoint2name[ord(",")] = "comma"
|
|
|
|
codepoint2name[ord(".")] = "dot"
|
|
|
|
codepoint2name[ord(":")] = "colon"
|
|
|
|
codepoint2name[ord(";")] = "semicolon"
|
|
|
|
codepoint2name[ord("/")] = "slash"
|
|
|
|
codepoint2name[ord("%")] = "percent"
|
|
|
|
codepoint2name[ord("#")] = "hash"
|
|
|
|
codepoint2name[ord("(")] = "paren_open"
|
|
|
|
codepoint2name[ord(")")] = "paren_close"
|
|
|
|
codepoint2name[ord("[")] = "bracket_open"
|
|
|
|
codepoint2name[ord("]")] = "bracket_close"
|
|
|
|
codepoint2name[ord("{")] = "brace_open"
|
|
|
|
codepoint2name[ord("}")] = "brace_close"
|
|
|
|
codepoint2name[ord("*")] = "star"
|
|
|
|
codepoint2name[ord("!")] = "bang"
|
|
|
|
codepoint2name[ord("\\")] = "backslash"
|
|
|
|
codepoint2name[ord("+")] = "plus"
|
|
|
|
codepoint2name[ord("$")] = "dollar"
|
|
|
|
codepoint2name[ord("=")] = "equals"
|
|
|
|
codepoint2name[ord("?")] = "question"
|
|
|
|
codepoint2name[ord("@")] = "at_sign"
|
|
|
|
codepoint2name[ord("^")] = "caret"
|
|
|
|
codepoint2name[ord("|")] = "pipe"
|
|
|
|
codepoint2name[ord("~")] = "tilde"
|
2014-02-15 11:34:50 +00:00
|
|
|
|
2019-03-01 03:33:03 +00:00
|
|
|
# static qstrs, should be sorted
|
|
|
|
|
|
|
|
static_qstr_list = [
|
|
|
|
"",
|
2020-02-27 04:36:53 +00:00
|
|
|
"__dir__", # Put __dir__ after empty qstr for builtin dir() to work
|
2019-03-01 03:33:03 +00:00
|
|
|
"\n",
|
|
|
|
" ",
|
|
|
|
"*",
|
|
|
|
"/",
|
|
|
|
"<module>",
|
|
|
|
"_",
|
|
|
|
"__call__",
|
|
|
|
"__class__",
|
|
|
|
"__delitem__",
|
|
|
|
"__enter__",
|
|
|
|
"__exit__",
|
|
|
|
"__getattr__",
|
|
|
|
"__getitem__",
|
|
|
|
"__hash__",
|
|
|
|
"__init__",
|
|
|
|
"__int__",
|
|
|
|
"__iter__",
|
|
|
|
"__len__",
|
|
|
|
"__main__",
|
|
|
|
"__module__",
|
|
|
|
"__name__",
|
|
|
|
"__new__",
|
|
|
|
"__next__",
|
|
|
|
"__qualname__",
|
|
|
|
"__repr__",
|
|
|
|
"__setitem__",
|
|
|
|
"__str__",
|
|
|
|
"ArithmeticError",
|
|
|
|
"AssertionError",
|
|
|
|
"AttributeError",
|
|
|
|
"BaseException",
|
|
|
|
"EOFError",
|
|
|
|
"Ellipsis",
|
|
|
|
"Exception",
|
|
|
|
"GeneratorExit",
|
|
|
|
"ImportError",
|
|
|
|
"IndentationError",
|
|
|
|
"IndexError",
|
|
|
|
"KeyError",
|
|
|
|
"KeyboardInterrupt",
|
|
|
|
"LookupError",
|
|
|
|
"MemoryError",
|
|
|
|
"NameError",
|
|
|
|
"NoneType",
|
|
|
|
"NotImplementedError",
|
|
|
|
"OSError",
|
|
|
|
"OverflowError",
|
|
|
|
"RuntimeError",
|
|
|
|
"StopIteration",
|
|
|
|
"SyntaxError",
|
|
|
|
"SystemExit",
|
|
|
|
"TypeError",
|
|
|
|
"ValueError",
|
|
|
|
"ZeroDivisionError",
|
|
|
|
"abs",
|
|
|
|
"all",
|
|
|
|
"any",
|
|
|
|
"append",
|
|
|
|
"args",
|
|
|
|
"bool",
|
|
|
|
"builtins",
|
|
|
|
"bytearray",
|
|
|
|
"bytecode",
|
|
|
|
"bytes",
|
|
|
|
"callable",
|
|
|
|
"chr",
|
|
|
|
"classmethod",
|
|
|
|
"clear",
|
|
|
|
"close",
|
|
|
|
"const",
|
|
|
|
"copy",
|
|
|
|
"count",
|
|
|
|
"dict",
|
|
|
|
"dir",
|
|
|
|
"divmod",
|
|
|
|
"end",
|
|
|
|
"endswith",
|
|
|
|
"eval",
|
|
|
|
"exec",
|
|
|
|
"extend",
|
|
|
|
"find",
|
|
|
|
"format",
|
|
|
|
"from_bytes",
|
|
|
|
"get",
|
|
|
|
"getattr",
|
|
|
|
"globals",
|
|
|
|
"hasattr",
|
|
|
|
"hash",
|
|
|
|
"id",
|
|
|
|
"index",
|
|
|
|
"insert",
|
|
|
|
"int",
|
|
|
|
"isalpha",
|
|
|
|
"isdigit",
|
|
|
|
"isinstance",
|
|
|
|
"islower",
|
|
|
|
"isspace",
|
|
|
|
"issubclass",
|
|
|
|
"isupper",
|
|
|
|
"items",
|
|
|
|
"iter",
|
|
|
|
"join",
|
|
|
|
"key",
|
|
|
|
"keys",
|
|
|
|
"len",
|
|
|
|
"list",
|
|
|
|
"little",
|
|
|
|
"locals",
|
|
|
|
"lower",
|
|
|
|
"lstrip",
|
|
|
|
"main",
|
|
|
|
"map",
|
|
|
|
"micropython",
|
|
|
|
"next",
|
|
|
|
"object",
|
|
|
|
"open",
|
|
|
|
"ord",
|
|
|
|
"pop",
|
|
|
|
"popitem",
|
|
|
|
"pow",
|
|
|
|
"print",
|
|
|
|
"range",
|
|
|
|
"read",
|
|
|
|
"readinto",
|
|
|
|
"readline",
|
|
|
|
"remove",
|
|
|
|
"replace",
|
|
|
|
"repr",
|
|
|
|
"reverse",
|
|
|
|
"rfind",
|
|
|
|
"rindex",
|
|
|
|
"round",
|
|
|
|
"rsplit",
|
|
|
|
"rstrip",
|
|
|
|
"self",
|
|
|
|
"send",
|
|
|
|
"sep",
|
|
|
|
"set",
|
|
|
|
"setattr",
|
|
|
|
"setdefault",
|
|
|
|
"sort",
|
|
|
|
"sorted",
|
|
|
|
"split",
|
|
|
|
"start",
|
|
|
|
"startswith",
|
|
|
|
"staticmethod",
|
|
|
|
"step",
|
|
|
|
"stop",
|
|
|
|
"str",
|
|
|
|
"strip",
|
|
|
|
"sum",
|
|
|
|
"super",
|
|
|
|
"throw",
|
|
|
|
"to_bytes",
|
|
|
|
"tuple",
|
|
|
|
"type",
|
|
|
|
"update",
|
|
|
|
"upper",
|
|
|
|
"utf-8",
|
|
|
|
"value",
|
|
|
|
"values",
|
|
|
|
"write",
|
|
|
|
"zip",
|
|
|
|
]
|
|
|
|
|
2014-01-21 21:40:13 +00:00
|
|
|
# this must match the equivalent function in qstr.c
|
2015-07-20 11:03:13 +00:00
|
|
|
def compute_hash(qstr, bytes_hash):
|
2014-03-25 15:27:15 +00:00
|
|
|
hash = 5381
|
2016-09-02 04:32:47 +00:00
|
|
|
for b in qstr:
|
|
|
|
hash = (hash * 33) ^ b
|
2014-06-06 20:55:27 +00:00
|
|
|
# Make sure that valid hash is never zero, zero means "hash not computed"
|
2015-07-20 11:03:13 +00:00
|
|
|
return (hash & ((1 << (8 * bytes_hash)) - 1)) or 1
|
2014-01-21 21:40:13 +00:00
|
|
|
|
2020-02-27 04:36:53 +00:00
|
|
|
|
2016-01-31 12:59:59 +00:00
|
|
|
def qstr_escape(qst):
|
2016-04-13 21:12:39 +00:00
|
|
|
def esc_char(m):
|
|
|
|
c = ord(m.group(0))
|
|
|
|
try:
|
|
|
|
name = codepoint2name[c]
|
|
|
|
except KeyError:
|
2020-02-27 04:36:53 +00:00
|
|
|
name = "0x%02x" % c
|
|
|
|
return "_" + name + "_"
|
|
|
|
|
|
|
|
return re.sub(r"[^A-Za-z0-9_]", esc_char, qst)
|
|
|
|
|
2016-01-31 12:59:59 +00:00
|
|
|
|
|
|
|
def parse_input_headers(infiles):
|
2015-01-11 17:52:45 +00:00
|
|
|
qcfgs = {}
|
2014-01-23 22:22:00 +00:00
|
|
|
qstrs = {}
|
2019-03-01 03:33:03 +00:00
|
|
|
|
|
|
|
# add static qstrs
|
|
|
|
for qstr in static_qstr_list:
|
|
|
|
# work out the corresponding qstr name
|
|
|
|
ident = qstr_escape(qstr)
|
|
|
|
|
|
|
|
# don't add duplicates
|
|
|
|
assert ident not in qstrs
|
|
|
|
|
|
|
|
# add the qstr to the list, with order number to retain original order in file
|
|
|
|
order = len(qstrs) - 300000
|
|
|
|
qstrs[ident] = (order, ident, qstr)
|
|
|
|
|
|
|
|
# read the qstrs in from the input files
|
2014-01-21 21:40:13 +00:00
|
|
|
for infile in infiles:
|
2020-02-27 04:36:53 +00:00
|
|
|
with open(infile, "rt") as f:
|
2014-01-21 21:40:13 +00:00
|
|
|
for line in f:
|
2015-01-11 17:52:45 +00:00
|
|
|
line = line.strip()
|
|
|
|
|
|
|
|
# is this a config line?
|
2020-02-27 04:36:53 +00:00
|
|
|
match = re.match(r"^QCFG\((.+), (.+)\)", line)
|
2015-01-11 17:52:45 +00:00
|
|
|
if match:
|
|
|
|
value = match.group(2)
|
2020-02-27 04:36:53 +00:00
|
|
|
if value[0] == "(" and value[-1] == ")":
|
2015-01-11 17:52:45 +00:00
|
|
|
# strip parenthesis from config value
|
|
|
|
value = value[1:-1]
|
|
|
|
qcfgs[match.group(1)] = value
|
|
|
|
continue
|
|
|
|
|
2014-05-02 19:10:47 +00:00
|
|
|
# is this a QSTR line?
|
2020-02-27 04:36:53 +00:00
|
|
|
match = re.match(r"^Q\((.*)\)$", line)
|
2014-05-02 19:10:47 +00:00
|
|
|
if not match:
|
2014-04-13 12:16:51 +00:00
|
|
|
continue
|
2014-01-21 21:40:13 +00:00
|
|
|
|
|
|
|
# get the qstr value
|
|
|
|
qstr = match.group(1)
|
2016-04-14 14:22:36 +00:00
|
|
|
|
2019-07-07 00:48:01 +00:00
|
|
|
# special cases to specify control characters
|
2020-02-27 04:36:53 +00:00
|
|
|
if qstr == "\\n":
|
|
|
|
qstr = "\n"
|
|
|
|
elif qstr == "\\r\\n":
|
|
|
|
qstr = "\r\n"
|
2016-04-14 14:22:36 +00:00
|
|
|
|
|
|
|
# work out the corresponding qstr name
|
2016-01-31 12:59:59 +00:00
|
|
|
ident = qstr_escape(qstr)
|
2014-01-21 21:40:13 +00:00
|
|
|
|
|
|
|
# don't add duplicates
|
2014-01-23 22:22:00 +00:00
|
|
|
if ident in qstrs:
|
2014-01-21 21:40:13 +00:00
|
|
|
continue
|
|
|
|
|
2014-01-24 22:22:00 +00:00
|
|
|
# add the qstr to the list, with order number to retain original order in file
|
2017-10-21 08:06:32 +00:00
|
|
|
order = len(qstrs)
|
|
|
|
# but put special method names like __add__ at the top of list, so
|
|
|
|
# that their id's fit into a byte
|
|
|
|
if ident == "":
|
|
|
|
# Sort empty qstr above all still
|
|
|
|
order = -200000
|
2018-05-10 13:10:46 +00:00
|
|
|
elif ident == "__dir__":
|
|
|
|
# Put __dir__ after empty qstr for builtin dir() to work
|
|
|
|
order = -190000
|
2017-10-21 08:06:32 +00:00
|
|
|
elif ident.startswith("__"):
|
|
|
|
order -= 100000
|
|
|
|
qstrs[ident] = (order, ident, qstr)
|
2014-01-21 21:40:13 +00:00
|
|
|
|
2015-10-11 08:09:57 +00:00
|
|
|
if not qcfgs:
|
|
|
|
sys.stderr.write("ERROR: Empty preprocessor output - check for errors above\n")
|
|
|
|
sys.exit(1)
|
|
|
|
|
2016-01-31 12:59:59 +00:00
|
|
|
return qcfgs, qstrs
|
|
|
|
|
2020-02-27 04:36:53 +00:00
|
|
|
|
2021-05-03 18:17:36 +00:00
|
|
|
def escape_bytes(qstr, qbytes):
|
2020-02-27 04:36:53 +00:00
|
|
|
if all(32 <= ord(c) <= 126 and c != "\\" and c != '"' for c in qstr):
|
2016-04-14 13:20:25 +00:00
|
|
|
# qstr is all printable ASCII so render it as-is (for easier debugging)
|
2021-05-03 18:17:36 +00:00
|
|
|
return qstr
|
2016-04-14 13:20:25 +00:00
|
|
|
else:
|
|
|
|
# qstr contains non-printable codes so render entire thing as hex pairs
|
2021-05-03 18:17:36 +00:00
|
|
|
return "".join(("\\x%02x" % b) for b in qbytes)
|
|
|
|
|
|
|
|
|
|
|
|
def make_bytes(cfg_bytes_len, cfg_bytes_hash, qstr):
|
|
|
|
qbytes = bytes_cons(qstr, "utf8")
|
|
|
|
qlen = len(qbytes)
|
|
|
|
qhash = compute_hash(qbytes, cfg_bytes_hash)
|
2016-01-31 12:59:59 +00:00
|
|
|
if qlen >= (1 << (8 * cfg_bytes_len)):
|
2020-02-27 04:36:53 +00:00
|
|
|
print("qstr is too long:", qstr)
|
2016-01-31 12:59:59 +00:00
|
|
|
assert False
|
2021-05-03 18:17:36 +00:00
|
|
|
qdata = escape_bytes(qstr, qbytes)
|
|
|
|
return '%d, %d, "%s"' % (qhash, qlen, qdata)
|
2016-01-31 12:59:59 +00:00
|
|
|
|
2020-02-27 04:36:53 +00:00
|
|
|
|
2016-01-31 12:59:59 +00:00
|
|
|
def print_qstr_data(qcfgs, qstrs):
|
2015-01-11 22:27:30 +00:00
|
|
|
# get config variables
|
2020-02-27 04:36:53 +00:00
|
|
|
cfg_bytes_len = int(qcfgs["BYTES_IN_LEN"])
|
|
|
|
cfg_bytes_hash = int(qcfgs["BYTES_IN_HASH"])
|
2015-01-11 22:27:30 +00:00
|
|
|
|
2015-07-31 11:57:36 +00:00
|
|
|
# print out the starter of the generated C header file
|
2020-02-27 04:36:53 +00:00
|
|
|
print("// This file was automatically generated by makeqstrdata.py")
|
|
|
|
print("")
|
2015-01-11 22:27:30 +00:00
|
|
|
|
2015-01-11 17:52:45 +00:00
|
|
|
# add NULL qstr with no hash or data
|
2021-05-03 18:17:36 +00:00
|
|
|
print('QDEF(MP_QSTRnull, 0, 0, "")')
|
2015-01-11 22:27:30 +00:00
|
|
|
|
|
|
|
# go through each qstr and print it out
|
2014-04-11 17:36:08 +00:00
|
|
|
for order, ident, qstr in sorted(qstrs.values(), key=lambda x: x[0]):
|
2016-01-31 12:59:59 +00:00
|
|
|
qbytes = make_bytes(cfg_bytes_len, cfg_bytes_hash, qstr)
|
2020-02-27 04:36:53 +00:00
|
|
|
print("QDEF(MP_QSTR_%s, %s)" % (ident, qbytes))
|
|
|
|
|
2016-01-31 12:59:59 +00:00
|
|
|
|
|
|
|
def do_work(infiles):
|
|
|
|
qcfgs, qstrs = parse_input_headers(infiles)
|
|
|
|
print_qstr_data(qcfgs, qstrs)
|
2014-01-21 21:40:13 +00:00
|
|
|
|
2020-02-27 04:36:53 +00:00
|
|
|
|
2014-01-21 21:40:13 +00:00
|
|
|
if __name__ == "__main__":
|
2015-05-30 22:11:16 +00:00
|
|
|
do_work(sys.argv[1:])
|