William Kurkian | ea86948 | 2019-04-09 15:16:11 -0400 | [diff] [blame] | 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 | "flag" |
| 12 | "fmt" |
| 13 | "io" |
| 14 | "log" |
| 15 | "reflect" |
| 16 | "strings" |
| 17 | "unicode" |
| 18 | |
| 19 | "golang.org/x/text/collate" |
| 20 | "golang.org/x/text/internal/gen" |
| 21 | "golang.org/x/text/internal/ucd" |
| 22 | "golang.org/x/text/language" |
| 23 | "golang.org/x/text/unicode/rangetable" |
| 24 | ) |
| 25 | |
| 26 | var versionList = flag.String("versions", "", |
| 27 | "list of versions for which to generate RangeTables") |
| 28 | |
| 29 | const bootstrapMessage = `No versions specified. |
| 30 | To bootstrap the code generation, run: |
| 31 | 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 |
| 32 | |
| 33 | and ensure that the latest versions are included by checking: |
Abhilash S.L | 3b49463 | 2019-07-16 15:51:09 +0530 | [diff] [blame] | 34 | https://www.unicode.org/Public/` |
William Kurkian | ea86948 | 2019-04-09 15:16:11 -0400 | [diff] [blame] | 35 | |
| 36 | func getVersions() []string { |
| 37 | if *versionList == "" { |
| 38 | log.Fatal(bootstrapMessage) |
| 39 | } |
| 40 | |
| 41 | c := collate.New(language.Und, collate.Numeric) |
| 42 | versions := strings.Split(*versionList, ",") |
| 43 | c.SortStrings(versions) |
| 44 | |
| 45 | // Ensure that at least the current version is included. |
| 46 | for _, v := range versions { |
| 47 | if v == gen.UnicodeVersion() { |
| 48 | return versions |
| 49 | } |
| 50 | } |
| 51 | |
| 52 | versions = append(versions, gen.UnicodeVersion()) |
| 53 | c.SortStrings(versions) |
| 54 | return versions |
| 55 | } |
| 56 | |
| 57 | func main() { |
| 58 | gen.Init() |
| 59 | |
| 60 | versions := getVersions() |
| 61 | |
| 62 | w := &bytes.Buffer{} |
| 63 | |
| 64 | fmt.Fprintf(w, "//go:generate go run gen.go --versions=%s\n\n", strings.Join(versions, ",")) |
| 65 | fmt.Fprintf(w, "import \"unicode\"\n\n") |
| 66 | |
| 67 | vstr := func(s string) string { return strings.Replace(s, ".", "_", -1) } |
| 68 | |
| 69 | fmt.Fprintf(w, "var assigned = map[string]*unicode.RangeTable{\n") |
| 70 | for _, v := range versions { |
| 71 | fmt.Fprintf(w, "\t%q: assigned%s,\n", v, vstr(v)) |
| 72 | } |
| 73 | fmt.Fprintf(w, "}\n\n") |
| 74 | |
| 75 | var size int |
| 76 | for _, v := range versions { |
| 77 | assigned := []rune{} |
| 78 | |
Abhilash S.L | 3b49463 | 2019-07-16 15:51:09 +0530 | [diff] [blame] | 79 | r := gen.Open("https://www.unicode.org/Public/", "", v+"/ucd/UnicodeData.txt") |
William Kurkian | ea86948 | 2019-04-09 15:16:11 -0400 | [diff] [blame] | 80 | ucd.Parse(r, func(p *ucd.Parser) { |
| 81 | assigned = append(assigned, p.Rune(0)) |
| 82 | }) |
| 83 | |
| 84 | rt := rangetable.New(assigned...) |
| 85 | sz := int(reflect.TypeOf(unicode.RangeTable{}).Size()) |
| 86 | sz += int(reflect.TypeOf(unicode.Range16{}).Size()) * len(rt.R16) |
| 87 | sz += int(reflect.TypeOf(unicode.Range32{}).Size()) * len(rt.R32) |
| 88 | |
| 89 | fmt.Fprintf(w, "// size %d bytes (%d KiB)\n", sz, sz/1024) |
| 90 | fmt.Fprintf(w, "var assigned%s = ", vstr(v)) |
| 91 | print(w, rt) |
| 92 | |
| 93 | size += sz |
| 94 | } |
| 95 | |
| 96 | fmt.Fprintf(w, "// Total size %d bytes (%d KiB)\n", size, size/1024) |
| 97 | |
| 98 | gen.WriteVersionedGoFile("tables.go", "rangetable", w.Bytes()) |
| 99 | } |
| 100 | |
| 101 | func print(w io.Writer, rt *unicode.RangeTable) { |
| 102 | fmt.Fprintln(w, "&unicode.RangeTable{") |
| 103 | fmt.Fprintln(w, "\tR16: []unicode.Range16{") |
| 104 | for _, r := range rt.R16 { |
| 105 | fmt.Fprintf(w, "\t\t{%#04x, %#04x, %d},\n", r.Lo, r.Hi, r.Stride) |
| 106 | } |
| 107 | fmt.Fprintln(w, "\t},") |
| 108 | fmt.Fprintln(w, "\tR32: []unicode.Range32{") |
| 109 | for _, r := range rt.R32 { |
| 110 | fmt.Fprintf(w, "\t\t{%#08x, %#08x, %d},\n", r.Lo, r.Hi, r.Stride) |
| 111 | } |
| 112 | fmt.Fprintln(w, "\t},") |
| 113 | fmt.Fprintf(w, "\tLatinOffset: %d,\n", rt.LatinOffset) |
| 114 | fmt.Fprintf(w, "}\n\n") |
| 115 | } |