Matthew Maurer | 19837bd | 2020-06-02 11:15:36 -0700 | [diff] [blame^] | 1 | #!/usr/bin/env python3 |
| 2 | # |
| 3 | # Copyright 2011-2015 The Rust Project Developers. See the COPYRIGHT |
| 4 | # file at the top-level directory of this distribution and at |
| 5 | # http://rust-lang.org/COPYRIGHT. |
| 6 | # |
| 7 | # Licensed under the Apache License, Version 2.0 <LICENSE-APACHE or |
| 8 | # http://www.apache.org/licenses/LICENSE-2.0> or the MIT license |
| 9 | # <LICENSE-MIT or http://opensource.org/licenses/MIT>, at your |
| 10 | # option. This file may not be copied, modified, or distributed |
| 11 | # except according to those terms. |
| 12 | |
| 13 | # This script uses the following Unicode tables: |
| 14 | # - EastAsianWidth.txt |
| 15 | # - ReadMe.txt |
| 16 | # - UnicodeData.txt |
| 17 | # |
| 18 | # Since this should not require frequent updates, we just store this |
| 19 | # out-of-line and check the unicode.rs file into git. |
| 20 | |
| 21 | import fileinput, re, os, sys, operator |
| 22 | |
| 23 | preamble = '''// Copyright 2012-2015 The Rust Project Developers. See the COPYRIGHT |
| 24 | // file at the top-level directory of this distribution and at |
| 25 | // http://rust-lang.org/COPYRIGHT. |
| 26 | // |
| 27 | // Licensed under the Apache License, Version 2.0 <LICENSE-APACHE or |
| 28 | // http://www.apache.org/licenses/LICENSE-2.0> or the MIT license |
| 29 | // <LICENSE-MIT or http://opensource.org/licenses/MIT>, at your |
| 30 | // option. This file may not be copied, modified, or distributed |
| 31 | // except according to those terms. |
| 32 | |
| 33 | // NOTE: The following code was generated by "scripts/unicode.py", do not edit directly |
| 34 | |
| 35 | #![allow(missing_docs, non_upper_case_globals, non_snake_case)] |
| 36 | ''' |
| 37 | |
| 38 | # Mapping taken from Table 12 from: |
| 39 | # http://www.unicode.org/reports/tr44/#General_Category_Values |
| 40 | expanded_categories = { |
| 41 | 'Lu': ['LC', 'L'], 'Ll': ['LC', 'L'], 'Lt': ['LC', 'L'], |
| 42 | 'Lm': ['L'], 'Lo': ['L'], |
| 43 | 'Mn': ['M'], 'Mc': ['M'], 'Me': ['M'], |
| 44 | 'Nd': ['N'], 'Nl': ['N'], 'No': ['No'], |
| 45 | 'Pc': ['P'], 'Pd': ['P'], 'Ps': ['P'], 'Pe': ['P'], |
| 46 | 'Pi': ['P'], 'Pf': ['P'], 'Po': ['P'], |
| 47 | 'Sm': ['S'], 'Sc': ['S'], 'Sk': ['S'], 'So': ['S'], |
| 48 | 'Zs': ['Z'], 'Zl': ['Z'], 'Zp': ['Z'], |
| 49 | 'Cc': ['C'], 'Cf': ['C'], 'Cs': ['C'], 'Co': ['C'], 'Cn': ['C'], |
| 50 | } |
| 51 | |
| 52 | # these are the surrogate codepoints, which are not valid rust characters |
| 53 | surrogate_codepoints = (0xd800, 0xdfff) |
| 54 | |
| 55 | def fetch(f): |
| 56 | if not os.path.exists(os.path.basename(f)): |
| 57 | os.system("curl -O http://www.unicode.org/Public/UNIDATA/%s" |
| 58 | % f) |
| 59 | |
| 60 | if not os.path.exists(os.path.basename(f)): |
| 61 | sys.stderr.write("cannot load %s" % f) |
| 62 | exit(1) |
| 63 | |
| 64 | def is_surrogate(n): |
| 65 | return surrogate_codepoints[0] <= n <= surrogate_codepoints[1] |
| 66 | |
| 67 | def load_unicode_data(f): |
| 68 | fetch(f) |
| 69 | gencats = {} |
| 70 | |
| 71 | udict = {} |
| 72 | range_start = -1 |
| 73 | for line in fileinput.input(f): |
| 74 | data = line.split(';') |
| 75 | if len(data) != 15: |
| 76 | continue |
| 77 | cp = int(data[0], 16) |
| 78 | if is_surrogate(cp): |
| 79 | continue |
| 80 | if range_start >= 0: |
| 81 | for i in range(range_start, cp): |
| 82 | udict[i] = data |
| 83 | range_start = -1 |
| 84 | if data[1].endswith(", First>"): |
| 85 | range_start = cp |
| 86 | continue |
| 87 | udict[cp] = data |
| 88 | |
| 89 | for code in udict: |
| 90 | [code_org, name, gencat, combine, bidi, |
| 91 | decomp, deci, digit, num, mirror, |
| 92 | old, iso, upcase, lowcase, titlecase ] = udict[code] |
| 93 | |
| 94 | # place letter in categories as appropriate |
| 95 | for cat in [gencat, "Assigned"] + expanded_categories.get(gencat, []): |
| 96 | if cat not in gencats: |
| 97 | gencats[cat] = [] |
| 98 | gencats[cat].append(code) |
| 99 | |
| 100 | gencats = group_cats(gencats) |
| 101 | |
| 102 | return gencats |
| 103 | |
| 104 | def group_cats(cats): |
| 105 | cats_out = {} |
| 106 | for cat in cats: |
| 107 | cats_out[cat] = group_cat(cats[cat]) |
| 108 | return cats_out |
| 109 | |
| 110 | def group_cat(cat): |
| 111 | cat_out = [] |
| 112 | letters = sorted(set(cat)) |
| 113 | cur_start = letters.pop(0) |
| 114 | cur_end = cur_start |
| 115 | for letter in letters: |
| 116 | assert letter > cur_end, \ |
| 117 | "cur_end: %s, letter: %s" % (hex(cur_end), hex(letter)) |
| 118 | if letter == cur_end + 1: |
| 119 | cur_end = letter |
| 120 | else: |
| 121 | cat_out.append((cur_start, cur_end)) |
| 122 | cur_start = cur_end = letter |
| 123 | cat_out.append((cur_start, cur_end)) |
| 124 | return cat_out |
| 125 | |
| 126 | def format_table_content(f, content, indent): |
| 127 | line = " "*indent |
| 128 | first = True |
| 129 | for chunk in content.split(","): |
| 130 | if len(line) + len(chunk) < 98: |
| 131 | if first: |
| 132 | line += chunk |
| 133 | else: |
| 134 | line += ", " + chunk |
| 135 | first = False |
| 136 | else: |
| 137 | f.write(line + ",\n") |
| 138 | line = " "*indent + chunk |
| 139 | f.write(line) |
| 140 | |
| 141 | # load all widths of want_widths, except those in except_cats |
| 142 | def load_east_asian_width(want_widths, except_cats): |
| 143 | f = "EastAsianWidth.txt" |
| 144 | fetch(f) |
| 145 | widths = {} |
| 146 | re1 = re.compile("^([0-9A-F]+);(\w+) +# (\w+)") |
| 147 | re2 = re.compile("^([0-9A-F]+)\.\.([0-9A-F]+);(\w+) +# (\w+)") |
| 148 | |
| 149 | for line in fileinput.input(f): |
| 150 | width = None |
| 151 | d_lo = 0 |
| 152 | d_hi = 0 |
| 153 | cat = None |
| 154 | m = re1.match(line) |
| 155 | if m: |
| 156 | d_lo = m.group(1) |
| 157 | d_hi = m.group(1) |
| 158 | width = m.group(2) |
| 159 | cat = m.group(3) |
| 160 | else: |
| 161 | m = re2.match(line) |
| 162 | if m: |
| 163 | d_lo = m.group(1) |
| 164 | d_hi = m.group(2) |
| 165 | width = m.group(3) |
| 166 | cat = m.group(4) |
| 167 | else: |
| 168 | continue |
| 169 | if cat in except_cats or width not in want_widths: |
| 170 | continue |
| 171 | d_lo = int(d_lo, 16) |
| 172 | d_hi = int(d_hi, 16) |
| 173 | if width not in widths: |
| 174 | widths[width] = [] |
| 175 | widths[width].append((d_lo, d_hi)) |
| 176 | return widths |
| 177 | |
| 178 | def escape_char(c): |
| 179 | return "'\\u{%x}'" % c |
| 180 | |
| 181 | def emit_table(f, name, t_data, t_type = "&'static [(char, char)]", is_pub=True, |
| 182 | pfun=lambda x: "(%s,%s)" % (escape_char(x[0]), escape_char(x[1])), is_const=True): |
| 183 | pub_string = "const" |
| 184 | if not is_const: |
| 185 | pub_string = "let" |
| 186 | if is_pub: |
| 187 | pub_string = "pub " + pub_string |
| 188 | f.write(" %s %s: %s = &[\n" % (pub_string, name, t_type)) |
| 189 | data = "" |
| 190 | first = True |
| 191 | for dat in t_data: |
| 192 | if not first: |
| 193 | data += "," |
| 194 | first = False |
| 195 | data += pfun(dat) |
| 196 | format_table_content(f, data, 8) |
| 197 | f.write("\n ];\n\n") |
| 198 | |
| 199 | def emit_charwidth_module(f, width_table): |
| 200 | f.write("pub mod charwidth {") |
| 201 | f.write(""" |
| 202 | use core::option::Option::{self, Some, None}; |
| 203 | use core::result::Result::{Ok, Err}; |
| 204 | |
| 205 | #[inline] |
| 206 | fn bsearch_range_value_table(c: char, is_cjk: bool, r: &'static [(char, char, u8, u8)]) -> u8 { |
| 207 | use core::cmp::Ordering::{Equal, Less, Greater}; |
| 208 | match r.binary_search_by(|&(lo, hi, _, _)| { |
| 209 | if lo <= c && c <= hi { Equal } |
| 210 | else if hi < c { Less } |
| 211 | else { Greater } |
| 212 | }) { |
| 213 | Ok(idx) => { |
| 214 | let (_, _, r_ncjk, r_cjk) = r[idx]; |
| 215 | if is_cjk { r_cjk } else { r_ncjk } |
| 216 | } |
| 217 | Err(_) => 1 |
| 218 | } |
| 219 | } |
| 220 | """) |
| 221 | |
| 222 | f.write(""" |
| 223 | #[inline] |
| 224 | pub fn width(c: char, is_cjk: bool) -> Option<usize> { |
| 225 | match c as usize { |
| 226 | _c @ 0 => Some(0), // null is zero width |
| 227 | cu if cu < 0x20 => None, // control sequences have no width |
| 228 | cu if cu < 0x7F => Some(1), // ASCII |
| 229 | cu if cu < 0xA0 => None, // more control sequences |
| 230 | _ => Some(bsearch_range_value_table(c, is_cjk, charwidth_table) as usize) |
| 231 | } |
| 232 | } |
| 233 | |
| 234 | """) |
| 235 | |
| 236 | f.write(" // character width table. Based on Markus Kuhn's free wcwidth() implementation,\n") |
| 237 | f.write(" // http://www.cl.cam.ac.uk/~mgk25/ucs/wcwidth.c\n") |
| 238 | emit_table(f, "charwidth_table", width_table, "&'static [(char, char, u8, u8)]", is_pub=False, |
| 239 | pfun=lambda x: "(%s,%s,%s,%s)" % (escape_char(x[0]), escape_char(x[1]), x[2], x[3])) |
| 240 | f.write("}\n\n") |
| 241 | |
| 242 | def remove_from_wtable(wtable, val): |
| 243 | wtable_out = [] |
| 244 | while wtable: |
| 245 | if wtable[0][1] < val: |
| 246 | wtable_out.append(wtable.pop(0)) |
| 247 | elif wtable[0][0] > val: |
| 248 | break |
| 249 | else: |
| 250 | (wt_lo, wt_hi, width, width_cjk) = wtable.pop(0) |
| 251 | if wt_lo == wt_hi == val: |
| 252 | continue |
| 253 | elif wt_lo == val: |
| 254 | wtable_out.append((wt_lo+1, wt_hi, width, width_cjk)) |
| 255 | elif wt_hi == val: |
| 256 | wtable_out.append((wt_lo, wt_hi-1, width, width_cjk)) |
| 257 | else: |
| 258 | wtable_out.append((wt_lo, val-1, width, width_cjk)) |
| 259 | wtable_out.append((val+1, wt_hi, width, width_cjk)) |
| 260 | if wtable: |
| 261 | wtable_out.extend(wtable) |
| 262 | return wtable_out |
| 263 | |
| 264 | |
| 265 | |
| 266 | def optimize_width_table(wtable): |
| 267 | wtable_out = [] |
| 268 | w_this = wtable.pop(0) |
| 269 | while wtable: |
| 270 | if w_this[1] == wtable[0][0] - 1 and w_this[2:3] == wtable[0][2:3]: |
| 271 | w_tmp = wtable.pop(0) |
| 272 | w_this = (w_this[0], w_tmp[1], w_tmp[2], w_tmp[3]) |
| 273 | else: |
| 274 | wtable_out.append(w_this) |
| 275 | w_this = wtable.pop(0) |
| 276 | wtable_out.append(w_this) |
| 277 | return wtable_out |
| 278 | |
| 279 | if __name__ == "__main__": |
| 280 | r = "tables.rs" |
| 281 | if os.path.exists(r): |
| 282 | os.remove(r) |
| 283 | with open(r, "w") as rf: |
| 284 | # write the file's preamble |
| 285 | rf.write(preamble) |
| 286 | |
| 287 | # download and parse all the data |
| 288 | fetch("ReadMe.txt") |
| 289 | with open("ReadMe.txt") as readme: |
| 290 | pattern = "for Version (\d+)\.(\d+)\.(\d+) of the Unicode" |
| 291 | unicode_version = re.search(pattern, readme.read()).groups() |
| 292 | rf.write(""" |
| 293 | /// The version of [Unicode](http://www.unicode.org/) |
| 294 | /// that this version of unicode-width is based on. |
| 295 | pub const UNICODE_VERSION: (u64, u64, u64) = (%s, %s, %s); |
| 296 | |
| 297 | """ % unicode_version) |
| 298 | gencats = load_unicode_data("UnicodeData.txt") |
| 299 | |
| 300 | ### character width module |
| 301 | width_table = [] |
| 302 | for zwcat in ["Me", "Mn", "Cf"]: |
| 303 | width_table.extend([(lo_hi[0], lo_hi[1], 0, 0) for lo_hi in gencats[zwcat]]) |
| 304 | width_table.append((4448, 4607, 0, 0)) |
| 305 | |
| 306 | # get widths, except those that are explicitly marked zero-width above |
| 307 | ea_widths = load_east_asian_width(["W", "F", "A"], ["Me", "Mn", "Cf"]) |
| 308 | # these are doublewidth |
| 309 | for dwcat in ["W", "F"]: |
| 310 | width_table.extend([(lo_hi1[0], lo_hi1[1], 2, 2) for lo_hi1 in ea_widths[dwcat]]) |
| 311 | width_table.extend([(lo_hi2[0], lo_hi2[1], 1, 2) for lo_hi2 in ea_widths["A"]]) |
| 312 | |
| 313 | width_table.sort(key=lambda w: w[0]) |
| 314 | |
| 315 | # soft hyphen is not zero width in preformatted text; it's used to indicate |
| 316 | # a hyphen inserted to facilitate a linebreak. |
| 317 | width_table = remove_from_wtable(width_table, 173) |
| 318 | |
| 319 | # optimize the width table by collapsing adjacent entities when possible |
| 320 | width_table = optimize_width_table(width_table) |
| 321 | emit_charwidth_module(rf, width_table) |