blob: fc4eee0a9fc7f15836de0f2008269c51b6402858 [file] [log] [blame]
Marcel van Lohuizen2acc1302015-06-30 20:24:20 +02001// 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
Russ Cox8f690f22021-02-19 18:54:44 -05005//go:build ignore
Marcel van Lohuizen2acc1302015-06-30 20:24:20 +02006// +build ignore
7
8package main
9
10import (
11 "bytes"
12 "flag"
13 "fmt"
14 "io"
15 "log"
16 "reflect"
Marcel van Lohuizen2acc1302015-06-30 20:24:20 +020017 "strings"
18 "unicode"
19
Marcel van Lohuizen996754b2017-09-12 18:24:34 +020020 "golang.org/x/text/collate"
Marcel van Lohuizen2acc1302015-06-30 20:24:20 +020021 "golang.org/x/text/internal/gen"
22 "golang.org/x/text/internal/ucd"
Marcel van Lohuizen996754b2017-09-12 18:24:34 +020023 "golang.org/x/text/language"
Marcel van Lohuizen2acc1302015-06-30 20:24:20 +020024 "golang.org/x/text/unicode/rangetable"
25)
26
27var versionList = flag.String("versions", "",
28 "list of versions for which to generate RangeTables")
29
30const bootstrapMessage = `No versions specified.
31To bootstrap the code generation, run:
32 go run gen.go --versions=4.1.0,5.0.0,6.0.0,6.1.0,6.2.0,6.3.0,7.0.0
33
34and ensure that the latest versions are included by checking:
Kevin Burke647d7ef2018-08-04 08:55:54 -070035 https://www.unicode.org/Public/`
Marcel van Lohuizen2acc1302015-06-30 20:24:20 +020036
37func getVersions() []string {
38 if *versionList == "" {
39 log.Fatal(bootstrapMessage)
40 }
41
Marcel van Lohuizen996754b2017-09-12 18:24:34 +020042 c := collate.New(language.Und, collate.Numeric)
Marcel van Lohuizen2acc1302015-06-30 20:24:20 +020043 versions := strings.Split(*versionList, ",")
Marcel van Lohuizen996754b2017-09-12 18:24:34 +020044 c.SortStrings(versions)
Marcel van Lohuizen2acc1302015-06-30 20:24:20 +020045
46 // Ensure that at least the current version is included.
47 for _, v := range versions {
48 if v == gen.UnicodeVersion() {
49 return versions
50 }
51 }
52
53 versions = append(versions, gen.UnicodeVersion())
Marcel van Lohuizen996754b2017-09-12 18:24:34 +020054 c.SortStrings(versions)
Marcel van Lohuizen2acc1302015-06-30 20:24:20 +020055 return versions
56}
57
58func main() {
59 gen.Init()
60
61 versions := getVersions()
62
63 w := &bytes.Buffer{}
64
65 fmt.Fprintf(w, "//go:generate go run gen.go --versions=%s\n\n", strings.Join(versions, ","))
66 fmt.Fprintf(w, "import \"unicode\"\n\n")
67
68 vstr := func(s string) string { return strings.Replace(s, ".", "_", -1) }
69
70 fmt.Fprintf(w, "var assigned = map[string]*unicode.RangeTable{\n")
71 for _, v := range versions {
72 fmt.Fprintf(w, "\t%q: assigned%s,\n", v, vstr(v))
73 }
74 fmt.Fprintf(w, "}\n\n")
75
76 var size int
77 for _, v := range versions {
78 assigned := []rune{}
79
Kevin Burke647d7ef2018-08-04 08:55:54 -070080 r := gen.Open("https://www.unicode.org/Public/", "", v+"/ucd/UnicodeData.txt")
Marcel van Lohuizen698d4932015-12-02 19:11:18 +010081 ucd.Parse(r, func(p *ucd.Parser) {
Marcel van Lohuizen2acc1302015-06-30 20:24:20 +020082 assigned = append(assigned, p.Rune(0))
83 })
84
85 rt := rangetable.New(assigned...)
86 sz := int(reflect.TypeOf(unicode.RangeTable{}).Size())
87 sz += int(reflect.TypeOf(unicode.Range16{}).Size()) * len(rt.R16)
88 sz += int(reflect.TypeOf(unicode.Range32{}).Size()) * len(rt.R32)
89
90 fmt.Fprintf(w, "// size %d bytes (%d KiB)\n", sz, sz/1024)
91 fmt.Fprintf(w, "var assigned%s = ", vstr(v))
92 print(w, rt)
93
94 size += sz
95 }
96
97 fmt.Fprintf(w, "// Total size %d bytes (%d KiB)\n", size, size/1024)
98
Marcel van Lohuizenc4d099d2017-12-11 13:26:16 +010099 gen.WriteVersionedGoFile("tables.go", "rangetable", w.Bytes())
Marcel van Lohuizen2acc1302015-06-30 20:24:20 +0200100}
101
Marcel van Lohuizen2acc1302015-06-30 20:24:20 +0200102func print(w io.Writer, rt *unicode.RangeTable) {
103 fmt.Fprintln(w, "&unicode.RangeTable{")
104 fmt.Fprintln(w, "\tR16: []unicode.Range16{")
105 for _, r := range rt.R16 {
106 fmt.Fprintf(w, "\t\t{%#04x, %#04x, %d},\n", r.Lo, r.Hi, r.Stride)
107 }
108 fmt.Fprintln(w, "\t},")
109 fmt.Fprintln(w, "\tR32: []unicode.Range32{")
110 for _, r := range rt.R32 {
111 fmt.Fprintf(w, "\t\t{%#08x, %#08x, %d},\n", r.Lo, r.Hi, r.Stride)
112 }
113 fmt.Fprintln(w, "\t},")
114 fmt.Fprintf(w, "\tLatinOffset: %d,\n", rt.LatinOffset)
115 fmt.Fprintf(w, "}\n\n")
116}