[lldb] fix fd leak during lldb testsuite (#118093)
[llvm-project.git] / libcxx / utils / generate_escaped_output_table.py
blob59dd707ae6126d40985e17463d34078ff8e32554
1 #!/usr/bin/env python
2 # ===----------------------------------------------------------------------===##
4 # Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
5 # See https://llvm.org/LICENSE.txt for license information.
6 # SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
8 # ===----------------------------------------------------------------------===##
10 # The code is based on
11 # https://github.com/microsoft/STL/blob/main/tools/unicode_properties_parse/grapheme_break_property_data_gen.py
13 # Copyright (c) Microsoft Corporation.
14 # SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
16 from io import StringIO
17 from pathlib import Path
18 from dataclasses import dataclass
19 from typing import Optional
20 import re
21 import sys
24 @dataclass
25 class PropertyRange:
26 lower: int = -1
27 upper: int = -1
28 prop: str = None
31 @dataclass
32 class Entry:
33 lower: int = -1
34 offset: int = -1
37 LINE_REGEX = re.compile(
38 r"^(?P<lower>[0-9A-F]{4,6})(?:\.\.(?P<upper>[0-9A-F]{4,6}))?\s*;\s*(?P<prop>\w+)"
42 # https://www.unicode.org/reports/tr44/#GC_Values_Table
43 def filterGeneralProperty(element: PropertyRange) -> Optional[PropertyRange]:
44 if element.prop in ["Zs", "Zl", "Zp", "Cc", "Cf", "Cs", "Co", "Cn"]:
45 return element
46 return None
49 def parsePropertyLine(inputLine: str) -> Optional[PropertyRange]:
50 result = PropertyRange()
51 if m := LINE_REGEX.match(inputLine):
52 lower_str, upper_str, result.prop = m.group("lower", "upper", "prop")
53 result.lower = int(lower_str, base=16)
54 result.upper = result.lower
55 if upper_str is not None:
56 result.upper = int(upper_str, base=16)
57 return result
59 else:
60 return None
63 def compactPropertyRanges(input: list[PropertyRange]) -> list[PropertyRange]:
64 """
65 Merges overlapping and consecutive ranges to one range.
67 Since the input properties are filtered the exact property isn't
68 interesting anymore. The properties in the output are merged to aid
69 debugging.
70 Merging the ranges results in fewer ranges in the output table,
71 reducing binary and improving lookup performance.
72 """
73 result = list()
74 for x in input:
75 if (
76 len(result)
77 and x.lower > result[-1].lower
78 and x.lower <= result[-1].upper + 1
80 result[-1].upper = max(result[-1].upper, x.upper)
81 result[-1].prop += f" {x.prop}"
82 continue
83 result.append(x)
84 return result
87 DATA_ARRAY_TEMPLATE = r"""
88 /// The entries of the characters to escape in format's debug string.
89 ///
90 /// Contains the entries for [format.string.escaped]/2.2.1.2.1
91 /// CE is a Unicode encoding and C corresponds to a UCS scalar value whose
92 /// Unicode property General_Category has a value in the groups Separator (Z)
93 /// or Other (C), as described by table 12 of UAX #44
94 ///
95 /// Separator (Z) consists of General_Category
96 /// - Space_Separator,
97 /// - Line_Separator,
98 /// - Paragraph_Separator.
99 ///
100 /// Other (C) consists of General_Category
101 /// - Control,
102 /// - Format,
103 /// - Surrogate,
104 /// - Private_Use,
105 /// - Unassigned.
107 /// The data is generated from
108 /// - https://www.unicode.org/Public/UCD/latest/ucd/extracted/DerivedGeneralCategory.txt
110 /// The table is similar to the table
111 /// __extended_grapheme_custer_property_boundary::__entries
112 /// which explains the details of these classes. The only difference is this
113 /// table lacks a property, thus having more bits available for the size.
115 /// The data has 2 values:
116 /// - bits [0, 13] The size of the range, allowing 16384 elements.
117 /// - bits [14, 31] The lower bound code point of the range. The upper bound of
118 /// the range is lower bound + size. Note the code expects code units the fit
119 /// into 18 bits, instead of the 21 bits needed for the full Unicode range.
120 _LIBCPP_HIDE_FROM_ABI inline constexpr uint32_t __entries[{size}] = {{
121 {entries}}};
123 /// Returns whether the code unit needs to be escaped.
125 /// At the end of the valid Unicode code points space a lot of code points are
126 /// either reserved or a noncharacter. Adding all these entries to the
127 /// lookup table would greatly increase the size of the table. Instead these
128 /// entries are manually processed. In this large area of reserved code points,
129 /// there is a small area of extended graphemes that should not be escaped
130 /// unconditionally. This is also manually coded. See the generation script for
131 /// more details.
134 /// \\pre The code point is a valid Unicode code point.
135 [[nodiscard]] _LIBCPP_HIDE_FROM_ABI constexpr bool __needs_escape(const char32_t __code_point) noexcept {{
137 // The entries in the gap at the end.
138 if(__code_point >= 0x{gap_lower:08x} && __code_point <= 0x{gap_upper:08x})
139 return false;
141 // The entries at the end.
142 if (__code_point >= 0x{unallocated:08x})
143 return true;
145 ptrdiff_t __i = std::ranges::upper_bound(__entries, (__code_point << 14) | 0x3fffu) - __entries;
146 if (__i == 0)
147 return false;
149 --__i;
150 uint32_t __upper_bound = (__entries[__i] >> 14) + (__entries[__i] & 0x3fffu);
151 return __code_point <= __upper_bound;
155 TABLES_HPP_TEMPLATE = """
156 // -*- C++ -*-
157 //===----------------------------------------------------------------------===//
159 // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
160 // See https://llvm.org/LICENSE.txt for license information.
161 // SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
163 //===----------------------------------------------------------------------===//
165 // WARNING, this entire header is generated by
166 // utils/generate_escaped_output_table.py
167 // DO NOT MODIFY!
169 // UNICODE, INC. LICENSE AGREEMENT - DATA FILES AND SOFTWARE
171 // See Terms of Use <https://www.unicode.org/copyright.html>
172 // for definitions of Unicode Inc.'s Data Files and Software.
174 // NOTICE TO USER: Carefully read the following legal agreement.
175 // BY DOWNLOADING, INSTALLING, COPYING OR OTHERWISE USING UNICODE INC.'S
176 // DATA FILES ("DATA FILES"), AND/OR SOFTWARE ("SOFTWARE"),
177 // YOU UNEQUIVOCALLY ACCEPT, AND AGREE TO BE BOUND BY, ALL OF THE
178 // TERMS AND CONDITIONS OF THIS AGREEMENT.
179 // IF YOU DO NOT AGREE, DO NOT DOWNLOAD, INSTALL, COPY, DISTRIBUTE OR USE
180 // THE DATA FILES OR SOFTWARE.
182 // COPYRIGHT AND PERMISSION NOTICE
184 // Copyright (c) 1991-2022 Unicode, Inc. All rights reserved.
185 // Distributed under the Terms of Use in https://www.unicode.org/copyright.html.
187 // Permission is hereby granted, free of charge, to any person obtaining
188 // a copy of the Unicode data files and any associated documentation
189 // (the "Data Files") or Unicode software and any associated documentation
190 // (the "Software") to deal in the Data Files or Software
191 // without restriction, including without limitation the rights to use,
192 // copy, modify, merge, publish, distribute, and/or sell copies of
193 // the Data Files or Software, and to permit persons to whom the Data Files
194 // or Software are furnished to do so, provided that either
195 // (a) this copyright and permission notice appear with all copies
196 // of the Data Files or Software, or
197 // (b) this copyright and permission notice appear in associated
198 // Documentation.
200 // THE DATA FILES AND SOFTWARE ARE PROVIDED "AS IS", WITHOUT WARRANTY OF
201 // ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE
202 // WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
203 // NONINFRINGEMENT OF THIRD PARTY RIGHTS.
204 // IN NO EVENT SHALL THE COPYRIGHT HOLDER OR HOLDERS INCLUDED IN THIS
205 // NOTICE BE LIABLE FOR ANY CLAIM, OR ANY SPECIAL INDIRECT OR CONSEQUENTIAL
206 // DAMAGES, OR ANY DAMAGES WHATSOEVER RESULTING FROM LOSS OF USE,
207 // DATA OR PROFITS, WHETHER IN AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER
208 // TORTIOUS ACTION, ARISING OUT OF OR IN CONNECTION WITH THE USE OR
209 // PERFORMANCE OF THE DATA FILES OR SOFTWARE.
211 // Except as contained in this notice, the name of a copyright holder
212 // shall not be used in advertising or otherwise to promote the sale,
213 // use or other dealings in these Data Files or Software without prior
214 // written authorization of the copyright holder.
216 #ifndef _LIBCPP___FORMAT_ESCAPED_OUTPUT_TABLE_H
217 #define _LIBCPP___FORMAT_ESCAPED_OUTPUT_TABLE_H
219 #include <__algorithm/ranges_upper_bound.h>
220 #include <__config>
221 #include <__cstddef/ptrdiff_t.h>
222 #include <cstdint>
224 #if !defined(_LIBCPP_HAS_NO_PRAGMA_SYSTEM_HEADER)
225 # pragma GCC system_header
226 #endif
228 _LIBCPP_BEGIN_NAMESPACE_STD
230 #if _LIBCPP_STD_VER >= 23
232 namespace __escaped_output_table {{
233 // clang-format off
234 {content}
235 // clang-format on
236 }} // namespace __escaped_output_table
238 #endif // _LIBCPP_STD_VER >= 23
240 _LIBCPP_END_NAMESPACE_STD
242 #endif // _LIBCPP___FORMAT_ESCAPED_OUTPUT_TABLE_H"""
245 def property_ranges_to_table(ranges: list[PropertyRange]) -> list[Entry]:
246 result = list[Entry]()
247 high = -1
248 for range in sorted(ranges, key=lambda x: x.lower):
249 # Validate overlapping ranges
250 assert range.lower > high
251 high = range.upper
253 while True:
254 e = Entry(range.lower, range.upper - range.lower)
255 if e.offset <= 16383:
256 result.append(e)
257 break
258 e.offset = 16383
259 result.append(e)
260 range.lower += 16384
261 return result
264 cpp_entrytemplate = " 0x{:08x} /* {:08x} - {:08x} [{:>5}] */"
267 def generate_cpp_data(
268 ranges: list[PropertyRange], unallocated: int, gap_lower: int, gap_upper: int
269 ) -> str:
270 result = StringIO()
271 table = property_ranges_to_table(ranges)
272 # Validates all entries fit in 18 bits.
273 for x in table:
274 assert x.lower + x.offset < 0x3FFFF
275 result.write(
276 DATA_ARRAY_TEMPLATE.format(
277 size=len(table),
278 entries=",\n".join(
280 cpp_entrytemplate.format(
281 x.lower << 14 | x.offset,
282 x.lower,
283 x.lower + x.offset,
284 x.offset + 1,
286 for x in table
289 unallocated=unallocated,
290 gap_lower=gap_lower,
291 gap_upper=gap_upper,
295 return result.getvalue()
298 def generate_data_tables() -> str:
300 Generate Unicode data for [format.string.escaped]/2.2.1.2.1
302 derived_general_catagory_path = (
303 Path(__file__).absolute().parent
304 / "data"
305 / "unicode"
306 / "DerivedGeneralCategory.txt"
309 properties = list()
310 with derived_general_catagory_path.open(encoding="utf-8") as f:
311 properties.extend(
312 list(
313 filter(
314 filterGeneralProperty,
315 [x for line in f if (x := parsePropertyLine(line))],
320 data = compactPropertyRanges(sorted(properties, key=lambda x: x.lower))
322 # The output table has two large entries at the end, with a small "gap"
323 # E0100..E01EF ; Grapheme_Extend # Mn [240] VARIATION SELECTOR-17..VARIATION SELECTOR-256
324 # Based on Unicode 15.1.0:
325 # - Encoding all these entries in the table requires 1173 entries.
326 # - Manually handling these last two blocks reduces the size to 729 entries.
327 # This not only reduces the binary size, but also improves the performance
328 # by having fewer elements to search.
329 # The exact entries may differ between Unicode versions. When these numbers
330 # change the test needs to be updated too.
331 # libcxx/test/libcxx/utilities/format/format.string/format.string.std/escaped_output.pass.cpp
332 assert (data[-2].lower) == 0x323B0
333 assert (data[-2].upper) == 0xE00FF
334 assert (data[-1].lower) == 0xE01F0
335 assert (data[-1].upper) == 0x10FFFF
337 return "\n".join(
339 generate_cpp_data(
340 data[:-2], data[-2].lower, data[-2].upper + 1, data[-1].lower - 1
346 if __name__ == "__main__":
347 if len(sys.argv) == 2:
348 sys.stdout = open(sys.argv[1], "w")
349 print(TABLES_HPP_TEMPLATE.lstrip().format(content=generate_data_tables()))