src/etc/char_private.py

   1 #!/usr/bin/env python
   2 #
   3 # Copyright 2011-2016 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 # - UnicodeData.txt
  15
  16
  17 from collections import namedtuple
  18 import csv
  19 import os
  20 import subprocess
  21
  22 NUM_CODEPOINTS=0x110000
  23
  24 def to_ranges(iter):
  25     current = None
  26     for i in iter:
  27         if current is None or i != current[1] or i in (0x10000, 0x20000):
  28             if current is not None:
  29                 yield tuple(current)
  30             current = [i, i + 1]
  31         else:
  32             current[1] += 1
  33     if current is not None:
  34         yield tuple(current)
  35
  36 def get_escaped(codepoints):
  37     for c in codepoints:
  38         if (c.class_ or "Cn") in "Cc Cf Cs Co Cn Zl Zp Zs".split() and c.value != ord(' '):
  39             yield c.value
  40
  41 def get_file(f):
  42     try:
  43         return open(os.path.basename(f))
  44     except FileNotFoundError:
  45         subprocess.run(["curl", "-O", f], check=True)
  46         return open(os.path.basename(f))
  47
  48 Codepoint = namedtuple('Codepoint', 'value class_')
  49
  50 def get_codepoints(f):
  51     r = csv.reader(f, delimiter=";")
  52     prev_codepoint = 0
  53     class_first = None
  54     for row in r:
  55         codepoint = int(row[0], 16)
  56         name = row[1]
  57         class_ = row[2]
  58
  59         if class_first is not None:
  60             if not name.endswith("Last>"):
  61                 raise ValueError("Missing Last after First")
  62
  63         for c in range(prev_codepoint + 1, codepoint):
  64             yield Codepoint(c, class_first)
  65
  66         class_first = None
  67         if name.endswith("First>"):
  68             class_first = class_
  69
  70         yield Codepoint(codepoint, class_)
  71         prev_codepoint = codepoint
  72
  73     if class_first != None:
  74         raise ValueError("Missing Last after First")
  75
  76     for c in range(prev_codepoint + 1, NUM_CODEPOINTS):
  77         yield Codepoint(c, None)
  78
  79 def main():
  80     file = get_file("http://www.unicode.org/Public/UNIDATA/UnicodeData.txt")
  81
  82     codepoints = get_codepoints(file)
  83
  84     CUTOFF=0x10000
  85     singletons0 = []
  86     singletons1 = []
  87     normal0 = []
  88     normal1 = []
  89     extra = []
  90
  91     for a, b in to_ranges(get_escaped(codepoints)):
  92         if a > 2 * CUTOFF:
  93             extra.append((a, b - a))
  94         elif a == b - 1:
  95             if a & CUTOFF:
  96                 singletons1.append(a & ~CUTOFF)
  97             else:
  98                 singletons0.append(a)
  99         elif a == b - 2:
 100             if a & CUTOFF:
 101                 singletons1.append(a & ~CUTOFF)
 102                 singletons1.append((a + 1) & ~CUTOFF)
 103             else:
 104                 singletons0.append(a)
 105                 singletons0.append(a + 1)
 106         else:
 107             if a >= 2 * CUTOFF:
 108                 extra.append((a, b - a))
 109             elif a & CUTOFF:
 110                 normal1.append((a & ~CUTOFF, b - a))
 111             else:
 112                 normal0.append((a, b - a))
 113
 114     print("""\
 115 // Copyright 2012-2016 The Rust Project Developers. See the COPYRIGHT
 116 // file at the top-level directory of this distribution and at
 117 // http://rust-lang.org/COPYRIGHT.
 118 //
 119 // Licensed under the Apache License, Version 2.0 <LICENSE-APACHE or
 120 // http://www.apache.org/licenses/LICENSE-2.0> or the MIT license
 121 // <LICENSE-MIT or http://opensource.org/licenses/MIT>, at your
 122 // option. This file may not be copied, modified, or distributed
 123 // except according to those terms.
 124
 125 // NOTE: The following code was generated by "src/etc/char_private.py",
 126 //       do not edit directly!
 127
 128 use slice::SliceExt;
 129
 130 fn check(x: u16, singletons: &[u16], normal: &[u16]) -> bool {
 131     for &s in singletons {
 132         if x == s {
 133             return false;
 134         } else if x < s {
 135             break;
 136         }
 137     }
 138     for w in normal.chunks(2) {
 139         let start = w[0];
 140         let len = w[1];
 141         let difference = (x as i32) - (start as i32);
 142         if 0 <= difference {
 143             if difference < len as i32 {
 144                 return false;
 145             }
 146         } else {
 147             break;
 148         }
 149     }
 150     true
 151 }
 152
 153 pub fn is_printable(x: char) -> bool {
 154     let x = x as u32;
 155     let lower = x as u16;
 156     if x < 0x10000 {
 157         check(lower, SINGLETONS0, NORMAL0)
 158     } else if x < 0x20000 {
 159         check(lower, SINGLETONS1, NORMAL1)
 160     } else {\
 161 """)
 162     for a, b in extra:
 163         print("        if 0x{:x} <= x && x < 0x{:x} {{".format(a, a + b))
 164         print("            return false;")
 165         print("        }")
 166     print("""\
 167         true
 168     }
 169 }\
 170 """)
 171     print()
 172     print("const SINGLETONS0: &'static [u16] = &[")
 173     for s in singletons0:
 174         print("    0x{:x},".format(s))
 175     print("];")
 176     print("const SINGLETONS1: &'static [u16] = &[")
 177     for s in singletons1:
 178         print("    0x{:x},".format(s))
 179     print("];")
 180     print("const NORMAL0: &'static [u16] = &[")
 181     for a, b in normal0:
 182         print("    0x{:x}, 0x{:x},".format(a, b))
 183     print("];")
 184     print("const NORMAL1: &'static [u16] = &[")
 185     for a, b in normal1:
 186         print("    0x{:x}, 0x{:x},".format(a, b))
 187     print("];")
 188
 189 if __name__ == '__main__':
 190     main()