+++ /dev/null
-import re, unicodedata, sys\r
-\r
-if sys.maxunicode == 65535:\r
- raise RuntimeError("need UCS-4 Python")\r
-\r
-def gen_category(cats):\r
- for i in range(0, 0x110000):\r
- if unicodedata.category(unichr(i)) in cats:\r
- yield(i)\r
-\r
-def gen_bidirectional(cats):\r
- for i in range(0, 0x110000):\r
- if unicodedata.bidirectional(unichr(i)) in cats:\r
- yield(i)\r
-\r
-def compact_set(l):\r
- single = []\r
- tuple = []\r
- prev = None\r
- span = 0\r
- for e in l:\r
- if prev is None:\r
- prev = e\r
- span = 0\r
- continue\r
- if prev+span+1 != e:\r
- if span > 2:\r
- tuple.append((prev,prev+span+1))\r
- else:\r
- for i in range(prev, prev+span+1):\r
- single.append(i)\r
- prev = e\r
- span = 0\r
- else:\r
- span += 1\r
- if span:\r
- tuple.append((prev,prev+span+1))\r
- else:\r
- single.append(prev)\r
- tuple = " + ".join(["range(%d,%d)" % t for t in tuple])\r
- if not single:\r
- return "set(%s)" % tuple\r
- if not tuple:\r
- return "set(%s)" % repr(single)\r
- return "set(%s + %s)" % (repr(single),tuple)\r
-\r
-############## Read the tables in the RFC #######################\r
-\r
-data = open("rfc3454.txt").readlines()\r
-\r
-tables = []\r
-curname = None\r
-for l in data:\r
- l = l.strip()\r
- if not l:\r
- continue\r
- # Skip RFC page breaks\r
- if l.startswith("Hoffman & Blanchet") or\\r
- l.startswith("RFC 3454"):\r
- continue\r
- # Find start/end lines\r
- m = re.match("----- (Start|End) Table ([A-Z](.[0-9])+) -----", l)\r
- if m:\r
- if m.group(1) == "Start":\r
- if curname:\r
- raise RuntimeError("Double Start", (curname, l))\r
- curname = m.group(2)\r
- table = {}\r
- tables.append((curname, table))\r
- continue\r
- else:\r
- if not curname:\r
- raise RuntimeError("End without start", l)\r
- curname = None\r
- continue\r
- if not curname:\r
- continue\r
- # Now we are in a table\r
- fields = l.split(";")\r
- if len(fields) > 1:\r
- # Drop comment field\r
- fields = fields[:-1]\r
- if len(fields) == 1:\r
- fields = fields[0].split("-")\r
- if len(fields) > 1:\r
- # range\r
- try:\r
- start, end = fields\r
- except ValueError:\r
- raise RuntimeError("Unpacking problem", l)\r
- else:\r
- start = end = fields[0]\r
- start = int(start, 16)\r
- end = int(end, 16)\r
- for i in range(start, end+1):\r
- table[i] = i\r
- else:\r
- code, value = fields\r
- value = value.strip()\r
- if value:\r
- value = [int(v, 16) for v in value.split(" ")]\r
- else:\r
- # table B.1\r
- value = None\r
- table[int(code, 16)] = value\r
-\r
-########### Generate compact Python versions of the tables #############\r
-\r
-print """# This file is generated by mkstringprep.py. DO NOT EDIT.\r
-\"\"\"Library that exposes various tables found in the StringPrep RFC 3454.\r
-\r
-There are two kinds of tables: sets, for which a member test is provided,\r
-and mappings, for which a mapping function is provided.\r
-\"\"\"\r
-\r
-import unicodedata\r
-"""\r
-\r
-print "assert unicodedata.unidata_version == %s" % repr(unicodedata.unidata_version)\r
-\r
-# A.1 is the table of unassigned characters\r
-# XXX Plane 15 PUA is listed as unassigned in Python.\r
-name, table = tables[0]\r
-del tables[0]\r
-assert name == "A.1"\r
-table = set(table.keys())\r
-Cn = set(gen_category(["Cn"]))\r
-\r
-# FDD0..FDEF are process internal codes\r
-Cn -= set(range(0xFDD0, 0xFDF0))\r
-# not a character\r
-Cn -= set(range(0xFFFE, 0x110000, 0x10000))\r
-Cn -= set(range(0xFFFF, 0x110000, 0x10000))\r
-\r
-# assert table == Cn\r
-\r
-print """\r
-def in_table_a1(code):\r
- if unicodedata.category(code) != 'Cn': return False\r
- c = ord(code)\r
- if 0xFDD0 <= c < 0xFDF0: return False\r
- return (c & 0xFFFF) not in (0xFFFE, 0xFFFF)\r
-"""\r
-\r
-# B.1 cannot easily be derived\r
-name, table = tables[0]\r
-del tables[0]\r
-assert name == "B.1"\r
-table = sorted(table.keys())\r
-print """\r
-b1_set = """ + compact_set(table) + """\r
-def in_table_b1(code):\r
- return ord(code) in b1_set\r
-"""\r
-\r
-# B.2 and B.3 is case folding.\r
-# It takes CaseFolding.txt into account, which is\r
-# not available in the Python database. Since\r
-# B.2 is derived from B.3, we process B.3 first.\r
-# B.3 supposedly *is* CaseFolding-3.2.0.txt.\r
-\r
-name, table_b2 = tables[0]\r
-del tables[0]\r
-assert name == "B.2"\r
-\r
-name, table_b3 = tables[0]\r
-del tables[0]\r
-assert name == "B.3"\r
-\r
-# B.3 is mostly Python's .lower, except for a number\r
-# of special cases, e.g. considering canonical forms.\r
-\r
-b3_exceptions = {}\r
-\r
-for k,v in table_b2.items():\r
- if map(ord, unichr(k).lower()) != v:\r
- b3_exceptions[k] = u"".join(map(unichr,v))\r
-\r
-b3 = sorted(b3_exceptions.items())\r
-\r
-print """\r
-b3_exceptions = {"""\r
-for i,(k,v) in enumerate(b3):\r
- print "0x%x:%s," % (k, repr(v)),\r
- if i % 4 == 3:\r
- print\r
-print "}"\r
-\r
-print """\r
-def map_table_b3(code):\r
- r = b3_exceptions.get(ord(code))\r
- if r is not None: return r\r
- return code.lower()\r
-"""\r
-\r
-def map_table_b3(code):\r
- r = b3_exceptions.get(ord(code))\r
- if r is not None: return r\r
- return code.lower()\r
-\r
-# B.2 is case folding for NFKC. This is the same as B.3,\r
-# except where NormalizeWithKC(Fold(a)) !=\r
-# NormalizeWithKC(Fold(NormalizeWithKC(Fold(a))))\r
-\r
-def map_table_b2(a):\r
- al = map_table_b3(a)\r
- b = unicodedata.normalize("NFKC", al)\r
- bl = u"".join([map_table_b3(ch) for ch in b])\r
- c = unicodedata.normalize("NFKC", bl)\r
- if b != c:\r
- return c\r
- else:\r
- return al\r
-\r
-specials = {}\r
-for k,v in table_b2.items():\r
- if map(ord, map_table_b2(unichr(k))) != v:\r
- specials[k] = v\r
-\r
-# B.3 should not add any additional special cases\r
-assert specials == {}\r
-\r
-print """\r
-def map_table_b2(a):\r
- al = map_table_b3(a)\r
- b = unicodedata.normalize("NFKC", al)\r
- bl = u"".join([map_table_b3(ch) for ch in b])\r
- c = unicodedata.normalize("NFKC", bl)\r
- if b != c:\r
- return c\r
- else:\r
- return al\r
-"""\r
-\r
-# C.1.1 is a table with a single character\r
-name, table = tables[0]\r
-del tables[0]\r
-assert name == "C.1.1"\r
-assert table == {0x20:0x20}\r
-\r
-print """\r
-def in_table_c11(code):\r
- return code == u" "\r
-"""\r
-\r
-# C.1.2 is the rest of all space characters\r
-name, table = tables[0]\r
-del tables[0]\r
-assert name == "C.1.2"\r
-\r
-# table = set(table.keys())\r
-# Zs = set(gen_category(["Zs"])) - set([0x20])\r
-# assert Zs == table\r
-\r
-print """\r
-def in_table_c12(code):\r
- return unicodedata.category(code) == "Zs" and code != u" "\r
-\r
-def in_table_c11_c12(code):\r
- return unicodedata.category(code) == "Zs"\r
-"""\r
-\r
-# C.2.1 ASCII control characters\r
-name, table_c21 = tables[0]\r
-del tables[0]\r
-assert name == "C.2.1"\r
-\r
-Cc = set(gen_category(["Cc"]))\r
-Cc_ascii = Cc & set(range(128))\r
-table_c21 = set(table_c21.keys())\r
-assert Cc_ascii == table_c21\r
-\r
-print """\r
-def in_table_c21(code):\r
- return ord(code) < 128 and unicodedata.category(code) == "Cc"\r
-"""\r
-\r
-# C.2.2 Non-ASCII control characters. It also includes\r
-# a number of characters in category Cf.\r
-name, table_c22 = tables[0]\r
-del tables[0]\r
-assert name == "C.2.2"\r
-\r
-Cc_nonascii = Cc - Cc_ascii\r
-table_c22 = set(table_c22.keys())\r
-assert len(Cc_nonascii - table_c22) == 0\r
-\r
-specials = list(table_c22 - Cc_nonascii)\r
-specials.sort()\r
-\r
-print """c22_specials = """ + compact_set(specials) + """\r
-def in_table_c22(code):\r
- c = ord(code)\r
- if c < 128: return False\r
- if unicodedata.category(code) == "Cc": return True\r
- return c in c22_specials\r
-\r
-def in_table_c21_c22(code):\r
- return unicodedata.category(code) == "Cc" or \\\r
- ord(code) in c22_specials\r
-"""\r
-\r
-# C.3 Private use\r
-name, table = tables[0]\r
-del tables[0]\r
-assert name == "C.3"\r
-\r
-Co = set(gen_category(["Co"]))\r
-assert set(table.keys()) == Co\r
-\r
-print """\r
-def in_table_c3(code):\r
- return unicodedata.category(code) == "Co"\r
-"""\r
-\r
-# C.4 Non-character code points, xFFFE, xFFFF\r
-# plus process internal codes\r
-name, table = tables[0]\r
-del tables[0]\r
-assert name == "C.4"\r
-\r
-nonchar = set(range(0xFDD0,0xFDF0) +\r
- range(0xFFFE,0x110000,0x10000) +\r
- range(0xFFFF,0x110000,0x10000))\r
-table = set(table.keys())\r
-assert table == nonchar\r
-\r
-print """\r
-def in_table_c4(code):\r
- c = ord(code)\r
- if c < 0xFDD0: return False\r
- if c < 0xFDF0: return True\r
- return (ord(code) & 0xFFFF) in (0xFFFE, 0xFFFF)\r
-"""\r
-\r
-# C.5 Surrogate codes\r
-name, table = tables[0]\r
-del tables[0]\r
-assert name == "C.5"\r
-\r
-Cs = set(gen_category(["Cs"]))\r
-assert set(table.keys()) == Cs\r
-\r
-print """\r
-def in_table_c5(code):\r
- return unicodedata.category(code) == "Cs"\r
-"""\r
-\r
-# C.6 Inappropriate for plain text\r
-name, table = tables[0]\r
-del tables[0]\r
-assert name == "C.6"\r
-\r
-table = sorted(table.keys())\r
-\r
-print """\r
-c6_set = """ + compact_set(table) + """\r
-def in_table_c6(code):\r
- return ord(code) in c6_set\r
-"""\r
-\r
-# C.7 Inappropriate for canonical representation\r
-name, table = tables[0]\r
-del tables[0]\r
-assert name == "C.7"\r
-\r
-table = sorted(table.keys())\r
-\r
-print """\r
-c7_set = """ + compact_set(table) + """\r
-def in_table_c7(code):\r
- return ord(code) in c7_set\r
-"""\r
-\r
-# C.8 Change display properties or are deprecated\r
-name, table = tables[0]\r
-del tables[0]\r
-assert name == "C.8"\r
-\r
-table = sorted(table.keys())\r
-\r
-print """\r
-c8_set = """ + compact_set(table) + """\r
-def in_table_c8(code):\r
- return ord(code) in c8_set\r
-"""\r
-\r
-# C.9 Tagging characters\r
-name, table = tables[0]\r
-del tables[0]\r
-assert name == "C.9"\r
-\r
-table = sorted(table.keys())\r
-\r
-print """\r
-c9_set = """ + compact_set(table) + """\r
-def in_table_c9(code):\r
- return ord(code) in c9_set\r
-"""\r
-\r
-# D.1 Characters with bidirectional property "R" or "AL"\r
-name, table = tables[0]\r
-del tables[0]\r
-assert name == "D.1"\r
-\r
-RandAL = set(gen_bidirectional(["R","AL"]))\r
-assert set(table.keys()) == RandAL\r
-\r
-print """\r
-def in_table_d1(code):\r
- return unicodedata.bidirectional(code) in ("R","AL")\r
-"""\r
-\r
-# D.2 Characters with bidirectional property "L"\r
-name, table = tables[0]\r
-del tables[0]\r
-assert name == "D.2"\r
-\r
-L = set(gen_bidirectional(["L"]))\r
-assert set(table.keys()) == L\r
-\r
-print """\r
-def in_table_d2(code):\r
- return unicodedata.bidirectional(code) == "L"\r
-"""\r
projects / mirror_edk2.git / blobdiff