clerk/internal/lsp/lsputil/utf16.go (view raw)
Oleksandr Smirnov
Oleksandr Smirnov
olexsmir@gmail.com perf: lsp: resolve positions and entries via cached line index..., 1 month ago
olexsmir@gmail.com perf: lsp: resolve positions and entries via cached line index..., 1 month ago
| 1 | package lsputil |
| 2 | |
| 3 | import ( |
| 4 | "unicode/utf8" |
| 5 | |
| 6 | "go.lsp.dev/protocol" |
| 7 | ) |
| 8 | |
| 9 | // Utf16Col returns the UTF-16 code unit column (0-based) for a byte offset |
| 10 | // within the given line content (without newline). |
| 11 | func Utf16Col(line string, byteOffset int) int { |
| 12 | if byteOffset <= 0 { |
| 13 | return 0 |
| 14 | } |
| 15 | if byteOffset > len(line) { |
| 16 | byteOffset = len(line) |
| 17 | } |
| 18 | col := 0 |
| 19 | for i := 0; i < byteOffset; { |
| 20 | r, size := utf8.DecodeRuneInString(line[i:]) |
| 21 | if r == utf8.RuneError && size <= 1 { |
| 22 | break |
| 23 | } |
| 24 | col += utf16Len(r) |
| 25 | i += size |
| 26 | } |
| 27 | return col |
| 28 | } |
| 29 | |
| 30 | // Utf16ColBytes returns the UTF-16 code unit column of b (without newline). |
| 31 | func Utf16ColBytes(b []byte) int { |
| 32 | col := 0 |
| 33 | for i := 0; i < len(b); { |
| 34 | r, size := utf8.DecodeRune(b[i:]) |
| 35 | if r == utf8.RuneError && size <= 1 { |
| 36 | break |
| 37 | } |
| 38 | col += utf16Len(r) |
| 39 | i += size |
| 40 | } |
| 41 | return col |
| 42 | } |
| 43 | |
| 44 | // Utf16Len returns the UTF-16 code unit length of content[offset:end]. |
| 45 | func Utf16Len(content string, offset, end int) int { |
| 46 | if offset < 0 { |
| 47 | offset = 0 |
| 48 | } |
| 49 | if end > len(content) { |
| 50 | end = len(content) |
| 51 | } |
| 52 | if offset >= end { |
| 53 | return 0 |
| 54 | } |
| 55 | n := 0 |
| 56 | for i := offset; i < end; { |
| 57 | r, size := utf8.DecodeRuneInString(content[i:]) |
| 58 | if r == utf8.RuneError && size <= 1 { |
| 59 | break |
| 60 | } |
| 61 | n += utf16Len(r) |
| 62 | i += size |
| 63 | } |
| 64 | return n |
| 65 | } |
| 66 | |
| 67 | // Offset converts a 0-based line and UTF-16 code unit column to a byte offset |
| 68 | // in content, clamped to the content bounds. The inverse of LineCol. |
| 69 | func Offset(content string, line, col int) int { |
| 70 | if line < 0 { |
| 71 | line = 0 |
| 72 | } |
| 73 | off := 0 |
| 74 | for curLine := 0; curLine < line && off < len(content); { |
| 75 | switch content[off] { |
| 76 | case '\n': |
| 77 | off++ |
| 78 | curLine++ |
| 79 | case '\r': |
| 80 | off++ |
| 81 | if off < len(content) && content[off] == '\n' { |
| 82 | off++ |
| 83 | } |
| 84 | curLine++ |
| 85 | default: |
| 86 | for off < len(content) && content[off] != '\n' && content[off] != '\r' { |
| 87 | off++ |
| 88 | } |
| 89 | } |
| 90 | } |
| 91 | lineEnd := off |
| 92 | for lineEnd < len(content) && content[lineEnd] != '\n' && content[lineEnd] != '\r' { |
| 93 | lineEnd++ |
| 94 | } |
| 95 | units := 0 |
| 96 | for off < lineEnd && units < col { |
| 97 | r, size := utf8.DecodeRuneInString(content[off:]) |
| 98 | off += size |
| 99 | units += utf16Len(r) |
| 100 | } |
| 101 | return off |
| 102 | } |
| 103 | |
| 104 | // LineCol converts a byte offset in content to 0-based line number and |
| 105 | // UTF-16 code unit column. |
| 106 | func LineCol(content string, offset int) (line int, col int) { |
| 107 | if offset <= 0 { |
| 108 | return 0, 0 |
| 109 | } |
| 110 | if offset > len(content) { |
| 111 | offset = len(content) |
| 112 | } |
| 113 | lineStart := 0 |
| 114 | lineNum := 0 |
| 115 | for i := 0; i < len(content); { |
| 116 | r, size := utf8.DecodeRuneInString(content[i:]) |
| 117 | if r == utf8.RuneError && size <= 1 { |
| 118 | break |
| 119 | } |
| 120 | if r == '\n' { |
| 121 | if offset < i+size { |
| 122 | col = Utf16Col(content[lineStart:offset], offset-lineStart) |
| 123 | return lineNum, col |
| 124 | } |
| 125 | lineStart = i + size |
| 126 | lineNum++ |
| 127 | i += size |
| 128 | continue |
| 129 | } |
| 130 | if r == '\r' { |
| 131 | // skip \r\n |
| 132 | if i+1 < len(content) && content[i+1] == '\n' { |
| 133 | size++ |
| 134 | } |
| 135 | if offset < i+size { |
| 136 | col = Utf16Col(content[lineStart:offset], offset-lineStart) |
| 137 | return lineNum, col |
| 138 | } |
| 139 | lineStart = i + size |
| 140 | lineNum++ |
| 141 | i += size |
| 142 | continue |
| 143 | } |
| 144 | if i+size > offset { |
| 145 | col = Utf16Col(content[lineStart:offset], offset-lineStart) |
| 146 | return lineNum, col |
| 147 | } |
| 148 | i += size |
| 149 | } |
| 150 | col = Utf16Col(content[lineStart:], len(content)-lineStart) |
| 151 | return lineNum, col |
| 152 | } |
| 153 | |
| 154 | // Position converts a byte offset to an LSP position (0-based line, UTF-16 |
| 155 | // code unit column). The inverse of Offset at the protocol.Position level. |
| 156 | func Position(content string, offset int) protocol.Position { |
| 157 | line, col := LineCol(content, offset) |
| 158 | return protocol.Position{Line: uint32(line), Character: uint32(col)} |
| 159 | } |
| 160 | |
| 161 | // utf16Len returns the number of UTF-16 code units for a rune. |
| 162 | func utf16Len(r rune) int { |
| 163 | if r >= 0x10000 && r <= 0x10FFFF { |
| 164 | return 2 |
| 165 | } |
| 166 | return 1 |
| 167 | } |