github.com/graybobo/golang.org-package-offline-cache@v0.0.0-20200626051047-6608995c132f/x/text/encoding/htmlindex/gen.go (about)

     1  // Copyright 2015 The Go Authors. All rights reserved.
     2  // Use of this source code is governed by a BSD-style
     3  // license that can be found in the LICENSE file.
     4  
     5  // +build ignore
     6  
     7  package main
     8  
     9  import (
    10  	"bytes"
    11  	"encoding/json"
    12  	"fmt"
    13  	"log"
    14  	"strings"
    15  
    16  	"golang.org/x/text/internal/gen"
    17  )
    18  
    19  type group struct {
    20  	Encodings []struct {
    21  		Labels []string
    22  		Name   string
    23  	}
    24  }
    25  
    26  func main() {
    27  	gen.Init()
    28  
    29  	r := gen.Open("http://www.w3.org/TR", "w3", "encoding/indexes/encodings.json")
    30  	var groups []group
    31  	if err := json.NewDecoder(r).Decode(&groups); err != nil {
    32  		log.Fatalf("Error reading encodings.json: %v", err)
    33  	}
    34  
    35  	w := &bytes.Buffer{}
    36  	fmt.Fprintln(w, "type htmlEncoding byte")
    37  	fmt.Fprintln(w, "const (")
    38  	for i, g := range groups {
    39  		for _, e := range g.Encodings {
    40  			name := consts[e.Name]
    41  			if name == "" {
    42  				log.Fatalf("No const defined for %s.", e.Name)
    43  			}
    44  			if i == 0 {
    45  				fmt.Fprintf(w, "%s htmlEncoding = iota\n", name)
    46  			} else {
    47  				fmt.Fprintf(w, "%s\n", name)
    48  			}
    49  		}
    50  	}
    51  	fmt.Fprintln(w, "numEncodings")
    52  	fmt.Fprint(w, ")\n\n")
    53  
    54  	fmt.Fprintln(w, "var canonical = [numEncodings]string{")
    55  	for _, g := range groups {
    56  		for _, e := range g.Encodings {
    57  			fmt.Fprintf(w, "%q,\n", e.Name)
    58  		}
    59  	}
    60  	fmt.Fprint(w, "}\n\n")
    61  
    62  	fmt.Fprintln(w, "var nameMap = map[string]htmlEncoding{")
    63  	for _, g := range groups {
    64  		for _, e := range g.Encodings {
    65  			for _, l := range e.Labels {
    66  				fmt.Fprintf(w, "%q: %s,\n", l, consts[e.Name])
    67  			}
    68  		}
    69  	}
    70  	fmt.Fprint(w, "}\n\n")
    71  
    72  	var tags []string
    73  	fmt.Fprintln(w, "var localeMap = []htmlEncoding{")
    74  	for _, loc := range locales {
    75  		tags = append(tags, loc.tag)
    76  		fmt.Fprintf(w, "%s, // %s \n", consts[loc.name], loc.tag)
    77  	}
    78  	fmt.Fprint(w, "}\n\n")
    79  
    80  	fmt.Fprintf(w, "const locales = %q\n", strings.Join(tags, " "))
    81  
    82  	gen.WriteGoFile("tables.go", "htmlindex", w.Bytes())
    83  }
    84  
    85  // consts maps canonical encoding name to internal constant.
    86  var consts = map[string]string{
    87  	"utf-8":          "utf8",
    88  	"ibm866":         "ibm866",
    89  	"iso-8859-2":     "iso8859_2",
    90  	"iso-8859-3":     "iso8859_3",
    91  	"iso-8859-4":     "iso8859_4",
    92  	"iso-8859-5":     "iso8859_5",
    93  	"iso-8859-6":     "iso8859_6",
    94  	"iso-8859-7":     "iso8859_7",
    95  	"iso-8859-8":     "iso8859_8",
    96  	"iso-8859-8-i":   "iso8859_8I",
    97  	"iso-8859-10":    "iso8859_10",
    98  	"iso-8859-13":    "iso8859_13",
    99  	"iso-8859-14":    "iso8859_14",
   100  	"iso-8859-15":    "iso8859_15",
   101  	"iso-8859-16":    "iso8859_16",
   102  	"koi8-r":         "koi8r",
   103  	"koi8-u":         "koi8u",
   104  	"macintosh":      "macintosh",
   105  	"windows-874":    "windows874",
   106  	"windows-1250":   "windows1250",
   107  	"windows-1251":   "windows1251",
   108  	"windows-1252":   "windows1252",
   109  	"windows-1253":   "windows1253",
   110  	"windows-1254":   "windows1254",
   111  	"windows-1255":   "windows1255",
   112  	"windows-1256":   "windows1256",
   113  	"windows-1257":   "windows1257",
   114  	"windows-1258":   "windows1258",
   115  	"x-mac-cyrillic": "macintoshCyrillic",
   116  	"gbk":            "gbk",
   117  	"gb18030":        "gb18030",
   118  	// "hz-gb-2312":     "hzgb2312", // Was removed from WhatWG
   119  	"big5":           "big5",
   120  	"euc-jp":         "eucjp",
   121  	"iso-2022-jp":    "iso2022jp",
   122  	"shift_jis":      "shiftJIS",
   123  	"euc-kr":         "euckr",
   124  	"replacement":    "replacement",
   125  	"utf-16be":       "utf16be",
   126  	"utf-16le":       "utf16le",
   127  	"x-user-defined": "xUserDefined",
   128  }
   129  
   130  // locales is taken from
   131  // https://html.spec.whatwg.org/multipage/syntax.html#encoding-sniffing-algorithm.
   132  var locales = []struct{ tag, name string }{
   133  	{"und", "windows-1252"}, // The default value.
   134  	{"ar", "windows-1256"},
   135  	{"ba", "windows-1251"},
   136  	{"be", "windows-1251"},
   137  	{"bg", "windows-1251"},
   138  	{"cs", "windows-1250"},
   139  	{"el", "iso-8859-7"},
   140  	{"et", "windows-1257"},
   141  	{"fa", "windows-1256"},
   142  	{"he", "windows-1255"},
   143  	{"hr", "windows-1250"},
   144  	{"hu", "iso-8859-2"},
   145  	{"ja", "shift_jis"},
   146  	{"kk", "windows-1251"},
   147  	{"ko", "euc-kr"},
   148  	{"ku", "windows-1254"},
   149  	{"ky", "windows-1251"},
   150  	{"lt", "windows-1257"},
   151  	{"lv", "windows-1257"},
   152  	{"mk", "windows-1251"},
   153  	{"pl", "iso-8859-2"},
   154  	{"ru", "windows-1251"},
   155  	{"sah", "windows-1251"},
   156  	{"sk", "windows-1250"},
   157  	{"sl", "iso-8859-2"},
   158  	{"sr", "windows-1251"},
   159  	{"tg", "windows-1251"},
   160  	{"th", "windows-874"},
   161  	{"tr", "windows-1254"},
   162  	{"tt", "windows-1251"},
   163  	{"uk", "windows-1251"},
   164  	{"vi", "windows-1258"},
   165  	{"zh-hans", "gb18030"},
   166  	{"zh-hant", "big5"},
   167  }