#! /usr/bin/env python # Copyright 2016 Google Inc. # # Use of this source code is governed by a BSD-style license that can be # found in the LICENSE file. import array, zlib, base64, sys ## Source: http://www.unicode.org/Public/9.0.0/ucd/UnicodeData.txt valid_codepoint_data = ''' eNrt3buPG0UcwHHPeNf22b67hPBIUpAc0iEgElAgIkSRXEkXUVOcKKj5F6BFQpSUqSgpkUAU+Rco +Av4PyLicD6ZzT5mZmfmtzPz/UhWFJ/X8356H/dmkHJGFgAAAADe6I73q55j1NXLJ0VRAAASHzN0 ouOcaol/zuPyKmJYdct7G9oMANqv01gKGScFpbUusN0ogbheCPSVb9KUkfFcQAceQ9XA3+YOfVJX nHdrtaon3FBprA3SWwnnc9u/uuO4trjuymnZ8r3LkfHeWI4pKsIYZJOXLvWrKw27409H1tllo0ym 3v7H5GVb3FeNOhyyH1gPpHHekw/zg/dV4715x3eYqCzjscuLRaD5em05FsVo2z6+ez4iHcuW92vH +rnynO6p5b9r2zXJ/2VHW7CJ18kE8zbW2tAkPZXg+tQlHF/lMg8cX9N5kh6IY9Vz7FAYC4cy3Bay 5qoEwjxvKXd1ML7bzskO68GY9qEG2nNfXdOO/cR9lv1B5z85tJdU80pFPvac5gK0tqV6xm+PAFDS fHno3E3XfaVU0lhCvivqddGk82IeKZwFRR21/tDGAAAAkMtaomQu58H27TGEvO7zyPP3bUemXY34 HoRbqzYtPH/Otk34CgPDVgnFtaa4Wq0Fw7bdP68cPp/KvEZHHt9DiHWOlk2/Xuq8lj1kmd/ndIB6 vBCoAxdUn0n2faz5kcN8XLr+llb3HyZcdxTreSBb7M0AkFyf68ziqxqvqcU/ZUrgWB0gLn1hqMD5 gGnUTT3xdiO5Tr85gfI8pUrDQkUWAAAAZLt2i+mOxWffoCgBAAAAAAAAAAAAdNCFh79ney7s0O/J dxv/v5FwHVHEHXipTrQNlPCspP09IE8Ch3Pk2N/k8DfI+JzxMhvcZxXw767l50u/nou+PgzuD8n6 bQzuHUh/mFo93lLuAGh7AAw9u1N4BnxJHQAAAAAAAMD/5fz78BHFCwBRvE4WIGMrsgCJGjo35lGG aTY9Jz/359XSb/nJF6l6wjUAfm2EwuVc/DBjl0ufP5X4psZ1T017/hziux05vLVlHdEOdbbEexes qcpFMu1bf3sgvyab9LryjLoEADmOf5D1bs/f2q4Nj3X9uup4TzX+znWoZblJFgBAMm4Unv5c58Lv OR73muHncjpXQAX8ripweMivP6J+oERzyzHa5P25Y7+/9jBXiNWOQ4xfY77TdU71Dk2gGPXAHCnV vgoyfdquDn2UaFq/p/46OY0c3tT6qZj3B07pGU1ScZ3ius21zn7AEAQBpmuHRSbtrJT4NOV2Ldbu WpqfaL4AMhTqHkJD12D83DNPSO13EnWw9tYRwoH73OPjAOUxhTV0zH2fXXofZ16XfM+zczkfxbTO K6F+q7IsQ5M+Y1l4v3o/o3aa6z0TJfb9Y/ZpscbYZlvv24t/wpRr0MOM0nKcUVq6xrRLwUXGdyxw JjFXNXVK1l87EWyzTTne7821a5jys3a5f7fMvNP3MLMRruNSLgbSojymz+bYRUFtoXbME2XYxkKt 8UzuS6Uc235b3PVVHtSO/Urb+voHuuJoY//uxXnk/UzPvfiFeUFU1Fv3OZPLWH6Ldv7Sp5HiM2bv eezexSV7Rt7s1srPCsjPpUEaU8wG1zhL3C9V+joVPbEy+IILzwAAAAAAHn1DFiSrbW9hQ7YA1/7k 4U8AAAAAAGDC/uKCSgAAgGu/n5EHAAAgP5L3GTW9bm1/ve6lUDy7rtvjtJ9x+aI81A0bPu/n+S3X Tr4i5fNCfTy7/vGL173b1AMAAAAgRcfn//37nAvekuGyxjdZypvukw3d4+mrW5QRkJsp7wVrigcF e4ssaPUHWRCMr+es/EhWesftm4fzRs+6n8ehHL5TUR5WdVKPqMNKOO6q4zNz6gAAsOYHQH8EAAAA ZKTt9kpj9jqf9Jzg5GMPtW+e3veM7L+XlDXgc837PllozPS8z/21xE8r/32nqcN+VFuE3fY72+61 uEqXmr36W2tbWKrls2pm9zuvMoybS7rGlr8KFMYnV/+eJPZjZUX3MClboTFFCR0biss5t5KX7Zd2 jkOJ/c5nAmEeZ1gv9svJXxkuAAAAAACw4rJP0LdfdkSWZmVLFnj3wPLzue+Zrgw/t8gkHb59PeJY ZZAm7fk7fdZvX49j+fDF69Fz+iYAABDHP/Vm9jbZIOpfwWoh0w== ''' if sys.version_info[0] < 3: def to_unicode(i): return ('\\U%08x' % i).decode('unicode-escape').encode('UTF-8') else: def to_unicode(i): return chr(i) o = sys.stdout o.write(to_unicode(0xFEFF)) pack = array.array('I', zlib.decompress( base64.b64decode(valid_codepoint_data))) i = 0 last_row = -1 plist = pack.tolist() for increment in plist: i += increment + 1 row = i - (i & 63) if last_row != row: if row: o.write('\n' if row % 1024 else '\n\n') o.write('U+%06x ' % row) last_row = row o.write(' ' + to_unicode(i)) o.write('\n')