| 1 | // Scintilla source code edit control |
| 2 | // Nimrod lexer |
| 3 | // (c) 2009 Andreas Rumpf |
| 4 | /** @file LexNimrod.cxx |
| 5 | ** Lexer for Nimrod. |
| 6 | **/ |
| 7 | // Copyright 1998-2002 by Neil Hodgson <neilh@scintilla.org> |
| 8 | // The License.txt file describes the conditions under which this software may be distributed. |
| 9 | |
| 10 | #include <stdlib.h> |
| 11 | #include <string.h> |
| 12 | #include <stdio.h> |
| 13 | #include <stdarg.h> |
| 14 | #include <assert.h> |
| 15 | #include <ctype.h> |
| 16 | |
| 17 | #include "ILexer.h" |
| 18 | #include "Scintilla.h" |
| 19 | #include "SciLexer.h" |
| 20 | |
| 21 | #include "WordList.h" |
| 22 | #include "LexAccessor.h" |
| 23 | #include "Accessor.h" |
| 24 | #include "StyleContext.h" |
| 25 | #include "CharacterSet.h" |
| 26 | #include "LexerModule.h" |
| 27 | |
| 28 | #ifdef SCI_NAMESPACE |
| 29 | using namespace Scintilla; |
| 30 | #endif |
| 31 | |
| 32 | static inline bool IsAWordChar(int ch) { |
| 33 | return (ch >= 0x80) || isalnum(ch) || ch == '_'; |
| 34 | } |
| 35 | |
| 36 | static int tillEndOfTripleQuote(Accessor &styler, int pos, int max) { |
| 37 | /* search for """ */ |
| 38 | for (;;) { |
| 39 | if (styler.SafeGetCharAt(pos, '\0') == '\0') return pos; |
| 40 | if (pos >= max) return pos; |
| 41 | if (styler.Match(pos, "\"\"\"")) { |
| 42 | return pos + 2; |
| 43 | } |
| 44 | pos++; |
| 45 | } |
| 46 | } |
| 47 | |
| 48 | #define CR 13 /* use both because Scite allows changing the line ending */ |
| 49 | #define LF 10 |
| 50 | |
| 51 | static bool inline isNewLine(int ch) { |
| 52 | return ch == CR || ch == LF; |
| 53 | } |
| 54 | |
| 55 | static int scanString(Accessor &styler, int pos, int max, bool rawMode) { |
| 56 | for (;;) { |
| 57 | if (pos >= max) return pos; |
| 58 | char ch = styler.SafeGetCharAt(pos, '\0'); |
| 59 | if (ch == CR || ch == LF || ch == '\0') return pos; |
| 60 | if (ch == '"') return pos; |
| 61 | if (ch == '\\' && !rawMode) { |
| 62 | pos += 2; |
| 63 | } else { |
| 64 | pos++; |
| 65 | } |
| 66 | } |
| 67 | } |
| 68 | |
| 69 | static int scanChar(Accessor &styler, int pos, int max) { |
| 70 | for (;;) { |
| 71 | if (pos >= max) return pos; |
| 72 | char ch = styler.SafeGetCharAt(pos, '\0'); |
| 73 | if (ch == CR || ch == LF || ch == '\0') return pos; |
| 74 | if (ch == '\'' && !isalnum(styler.SafeGetCharAt(pos+1, '\0')) ) |
| 75 | return pos; |
| 76 | if (ch == '\\') { |
| 77 | pos += 2; |
| 78 | } else { |
| 79 | pos++; |
| 80 | } |
| 81 | } |
| 82 | } |
| 83 | |
| 84 | static int scanIdent(Accessor &styler, int pos, WordList &keywords) { |
| 85 | char buf[100]; /* copy to lowercase and ignore underscores */ |
| 86 | int i = 0; |
| 87 | |
| 88 | for (;;) { |
| 89 | char ch = styler.SafeGetCharAt(pos, '\0'); |
| 90 | if (!IsAWordChar(ch)) break; |
| 91 | if (ch != '_' && i < ((int)sizeof(buf))-1) { |
| 92 | buf[i] = static_cast<char>(tolower(ch)); |
| 93 | i++; |
| 94 | } |
| 95 | pos++; |
| 96 | } |
| 97 | buf[i] = '\0'; |
| 98 | /* look for keyword */ |
| 99 | if (keywords.InList(buf)) { |
| 100 | styler.ColourTo(pos-1, SCE_P_WORD); |
| 101 | } else { |
| 102 | styler.ColourTo(pos-1, SCE_P_IDENTIFIER); |
| 103 | } |
| 104 | return pos; |
| 105 | } |
| 106 | |
| 107 | static int scanNumber(Accessor &styler, int pos) { |
| 108 | char ch, ch2; |
| 109 | ch = styler.SafeGetCharAt(pos, '\0'); |
| 110 | ch2 = styler.SafeGetCharAt(pos+1, '\0'); |
| 111 | if (ch == '0' && (ch2 == 'b' || ch2 == 'B')) { |
| 112 | /* binary number: */ |
| 113 | pos += 2; |
| 114 | for (;;) { |
| 115 | ch = styler.SafeGetCharAt(pos, '\0'); |
| 116 | if (ch == '_' || (ch >= '0' && ch <= '1')) ++pos; |
| 117 | else break; |
| 118 | } |
| 119 | } else if (ch == '0' && |
| 120 | (ch2 == 'o' || ch2 == 'O' || ch2 == 'c' || ch2 == 'C')) { |
| 121 | /* octal number: */ |
| 122 | pos += 2; |
| 123 | for (;;) { |
| 124 | ch = styler.SafeGetCharAt(pos, '\0'); |
| 125 | if (ch == '_' || (ch >= '0' && ch <= '7')) ++pos; |
| 126 | else break; |
| 127 | } |
| 128 | } else if (ch == '0' && (ch2 == 'x' || ch2 == 'X')) { |
| 129 | /* hexadecimal number: */ |
| 130 | pos += 2; |
| 131 | for (;;) { |
| 132 | ch = styler.SafeGetCharAt(pos, '\0'); |
| 133 | if (ch == '_' || (ch >= '0' && ch <= '9') |
| 134 | || (ch >= 'a' && ch <= 'f') |
| 135 | || (ch >= 'A' && ch <= 'F')) ++pos; |
| 136 | else break; |
| 137 | } |
| 138 | } else { |
| 139 | // skip decimal part: |
| 140 | for (;;) { |
| 141 | ch = styler.SafeGetCharAt(pos, '\0'); |
| 142 | if (ch == '_' || (ch >= '0' && ch <= '9')) ++pos; |
| 143 | else break; |
| 144 | } |
| 145 | ch2 = styler.SafeGetCharAt(pos+1, '\0'); |
| 146 | if (ch == '.' && ch2 >= '0' && ch2 <= '9') { |
| 147 | ++pos; // skip '.' |
| 148 | for (;;) { |
| 149 | ch = styler.SafeGetCharAt(pos, '\0'); |
| 150 | if (ch == '_' || (ch >= '0' && ch <= '9')) ++pos; |
| 151 | else break; |
| 152 | } |
| 153 | } |
| 154 | if (ch == 'e' || ch == 'E') { |
| 155 | ++pos; |
| 156 | ch = styler.SafeGetCharAt(pos, '\0'); |
| 157 | if (ch == '-' || ch == '+') ++pos; |
| 158 | for (;;) { |
| 159 | ch = styler.SafeGetCharAt(pos, '\0'); |
| 160 | if (ch == '_' || (ch >= '0' && ch <= '9')) ++pos; |
| 161 | else break; |
| 162 | } |
| 163 | } |
| 164 | } |
| 165 | if (ch == '\'') { |
| 166 | /* a type suffix: */ |
| 167 | pos++; |
| 168 | for (;;) { |
| 169 | ch = styler.SafeGetCharAt(pos); |
| 170 | if ((ch >= '0' && ch <= '9') || (ch >= 'A' && ch <= 'Z') |
| 171 | || (ch >= 'a' && ch <= 'z') || ch == '_') ++pos; |
| 172 | else break; |
| 173 | } |
| 174 | } |
| 175 | styler.ColourTo(pos-1, SCE_P_NUMBER); |
| 176 | return pos; |
| 177 | } |
| 178 | |
| 179 | /* rewritten from scratch, because I couldn't get rid of the bugs... |
| 180 | (A character based approach sucks!) |
| 181 | */ |
| 182 | static void ColouriseNimrodDoc(unsigned int startPos, int length, int initStyle, |
| 183 | WordList *keywordlists[], Accessor &styler) { |
| 184 | int pos = startPos; |
| 185 | int max = startPos + length; |
| 186 | char ch; |
| 187 | WordList &keywords = *keywordlists[0]; |
| 188 | |
| 189 | styler.StartAt(startPos); |
| 190 | styler.StartSegment(startPos); |
| 191 | |
| 192 | switch (initStyle) { |
| 193 | /* check where we are: */ |
| 194 | case SCE_P_TRIPLEDOUBLE: |
| 195 | pos = tillEndOfTripleQuote(styler, pos, max); |
| 196 | styler.ColourTo(pos, SCE_P_TRIPLEDOUBLE); |
| 197 | pos++; |
| 198 | break; |
| 199 | default: /* nothing to do: */ |
| 200 | break; |
| 201 | } |
| 202 | while (pos < max) { |
| 203 | ch = styler.SafeGetCharAt(pos, '\0'); |
| 204 | switch (ch) { |
| 205 | case '\0': return; |
| 206 | case '#': { |
| 207 | bool doccomment = (styler.SafeGetCharAt(pos+1) == '#'); |
| 208 | while (pos < max && !isNewLine(styler.SafeGetCharAt(pos, LF))) pos++; |
| 209 | if (doccomment) |
| 210 | styler.ColourTo(pos, SCE_C_COMMENTLINEDOC); |
| 211 | else |
| 212 | styler.ColourTo(pos, SCE_P_COMMENTLINE); |
| 213 | } break; |
| 214 | case 'r': case 'R': { |
| 215 | if (styler.SafeGetCharAt(pos+1) == '"') { |
| 216 | pos = scanString(styler, pos+2, max, true); |
| 217 | styler.ColourTo(pos, SCE_P_STRING); |
| 218 | pos++; |
| 219 | } else { |
| 220 | pos = scanIdent(styler, pos, keywords); |
| 221 | } |
| 222 | } break; |
| 223 | case '"': |
| 224 | if (styler.Match(pos+1, "\"\"")) { |
| 225 | pos = tillEndOfTripleQuote(styler, pos+3, max); |
| 226 | styler.ColourTo(pos, SCE_P_TRIPLEDOUBLE); |
| 227 | } else { |
| 228 | pos = scanString(styler, pos+1, max, false); |
| 229 | styler.ColourTo(pos, SCE_P_STRING); |
| 230 | } |
| 231 | pos++; |
| 232 | break; |
| 233 | case '\'': |
| 234 | pos = scanChar(styler, pos+1, max); |
| 235 | styler.ColourTo(pos, SCE_P_CHARACTER); |
| 236 | pos++; |
| 237 | break; |
| 238 | default: // identifers, numbers, operators, whitespace |
| 239 | if (ch >= '0' && ch <= '9') { |
| 240 | pos = scanNumber(styler, pos); |
| 241 | } else if (IsAWordChar(ch)) { |
| 242 | pos = scanIdent(styler, pos, keywords); |
| 243 | } else if (ch == '`') { |
| 244 | pos++; |
| 245 | while (pos < max) { |
| 246 | ch = styler.SafeGetCharAt(pos, LF); |
| 247 | if (ch == '`') { |
| 248 | ++pos; |
| 249 | break; |
| 250 | } |
| 251 | if (ch == CR || ch == LF) break; |
| 252 | ++pos; |
| 253 | } |
| 254 | styler.ColourTo(pos, SCE_P_IDENTIFIER); |
| 255 | } else if (strchr("()[]{}:=;-\\/&%$!+<>|^?,.*~@", ch)) { |
| 256 | styler.ColourTo(pos, SCE_P_OPERATOR); |
| 257 | pos++; |
| 258 | } else { |
| 259 | styler.ColourTo(pos, SCE_P_DEFAULT); |
| 260 | pos++; |
| 261 | } |
| 262 | break; |
| 263 | } |
| 264 | } |
| 265 | } |
| 266 | |
| 267 | static bool IsCommentLine(int line, Accessor &styler) { |
| 268 | int pos = styler.LineStart(line); |
| 269 | int eol_pos = styler.LineStart(line + 1) - 1; |
| 270 | for (int i = pos; i < eol_pos; i++) { |
| 271 | char ch = styler[i]; |
| 272 | if (ch == '#') |
| 273 | return true; |
| 274 | else if (ch != ' ' && ch != '\t') |
| 275 | return false; |
| 276 | } |
| 277 | return false; |
| 278 | } |
| 279 | |
| 280 | static bool IsQuoteLine(int line, Accessor &styler) { |
| 281 | int style = styler.StyleAt(styler.LineStart(line)) & 31; |
| 282 | return ((style == SCE_P_TRIPLE) || (style == SCE_P_TRIPLEDOUBLE)); |
| 283 | } |
| 284 | |
| 285 | |
| 286 | static void FoldNimrodDoc(unsigned int startPos, int length, |
| 287 | int /*initStyle - unused*/, |
| 288 | WordList *[], Accessor &styler) { |
| 289 | const int maxPos = startPos + length; |
| 290 | const int maxLines = styler.GetLine(maxPos - 1); // Requested last line |
| 291 | const int docLines = styler.GetLine(styler.Length() - 1); // Available last line |
| 292 | const bool foldComment = styler.GetPropertyInt("fold.comment.nimrod") != 0; |
| 293 | const bool foldQuotes = styler.GetPropertyInt("fold.quotes.nimrod") != 0; |
| 294 | |
| 295 | // Backtrack to previous non-blank line so we can determine indent level |
| 296 | // for any white space lines (needed esp. within triple quoted strings) |
| 297 | // and so we can fix any preceding fold level (which is why we go back |
| 298 | // at least one line in all cases) |
| 299 | int spaceFlags = 0; |
| 300 | int lineCurrent = styler.GetLine(startPos); |
| 301 | int indentCurrent = styler.IndentAmount(lineCurrent, &spaceFlags, NULL); |
| 302 | while (lineCurrent > 0) { |
| 303 | lineCurrent--; |
| 304 | indentCurrent = styler.IndentAmount(lineCurrent, &spaceFlags, NULL); |
| 305 | if (!(indentCurrent & SC_FOLDLEVELWHITEFLAG) && |
| 306 | (!IsCommentLine(lineCurrent, styler)) && |
| 307 | (!IsQuoteLine(lineCurrent, styler))) |
| 308 | break; |
| 309 | } |
| 310 | int indentCurrentLevel = indentCurrent & SC_FOLDLEVELNUMBERMASK; |
| 311 | |
| 312 | // Set up initial loop state |
| 313 | startPos = styler.LineStart(lineCurrent); |
| 314 | int prev_state = SCE_P_DEFAULT & 31; |
| 315 | if (lineCurrent >= 1) |
| 316 | prev_state = styler.StyleAt(startPos - 1) & 31; |
| 317 | int prevQuote = foldQuotes && ((prev_state == SCE_P_TRIPLE) || |
| 318 | (prev_state == SCE_P_TRIPLEDOUBLE)); |
| 319 | int prevComment = 0; |
| 320 | if (lineCurrent >= 1) |
| 321 | prevComment = foldComment && IsCommentLine(lineCurrent - 1, styler); |
| 322 | |
| 323 | // Process all characters to end of requested range or end of any triple quote |
| 324 | // or comment that hangs over the end of the range. Cap processing in all cases |
| 325 | // to end of document (in case of unclosed quote or comment at end). |
| 326 | while ((lineCurrent <= docLines) && ((lineCurrent <= maxLines) || |
| 327 | prevQuote || prevComment)) { |
| 328 | |
| 329 | // Gather info |
| 330 | int lev = indentCurrent; |
| 331 | int lineNext = lineCurrent + 1; |
| 332 | int indentNext = indentCurrent; |
| 333 | int quote = false; |
| 334 | if (lineNext <= docLines) { |
| 335 | // Information about next line is only available if not at end of document |
| 336 | indentNext = styler.IndentAmount(lineNext, &spaceFlags, NULL); |
| 337 | int style = styler.StyleAt(styler.LineStart(lineNext)) & 31; |
| 338 | quote = foldQuotes && ((style == SCE_P_TRIPLE) || (style == SCE_P_TRIPLEDOUBLE)); |
| 339 | } |
| 340 | const int quote_start = (quote && !prevQuote); |
| 341 | const int quote_continue = (quote && prevQuote); |
| 342 | const int comment = foldComment && IsCommentLine(lineCurrent, styler); |
| 343 | const int comment_start = (comment && !prevComment && (lineNext <= docLines) && |
| 344 | IsCommentLine(lineNext, styler) && |
| 345 | (lev > SC_FOLDLEVELBASE)); |
| 346 | const int comment_continue = (comment && prevComment); |
| 347 | if ((!quote || !prevQuote) && !comment) |
| 348 | indentCurrentLevel = indentCurrent & SC_FOLDLEVELNUMBERMASK; |
| 349 | if (quote) |
| 350 | indentNext = indentCurrentLevel; |
| 351 | if (indentNext & SC_FOLDLEVELWHITEFLAG) |
| 352 | indentNext = SC_FOLDLEVELWHITEFLAG | indentCurrentLevel; |
| 353 | |
| 354 | if (quote_start) { |
| 355 | // Place fold point at start of triple quoted string |
| 356 | lev |= SC_FOLDLEVELHEADERFLAG; |
| 357 | } else if (quote_continue || prevQuote) { |
| 358 | // Add level to rest of lines in the string |
| 359 | lev = lev + 1; |
| 360 | } else if (comment_start) { |
| 361 | // Place fold point at start of a block of comments |
| 362 | lev |= SC_FOLDLEVELHEADERFLAG; |
| 363 | } else if (comment_continue) { |
| 364 | // Add level to rest of lines in the block |
| 365 | lev = lev + 1; |
| 366 | } |
| 367 | |
| 368 | // Skip past any blank lines for next indent level info; we skip also |
| 369 | // comments (all comments, not just those starting in column 0) |
| 370 | // which effectively folds them into surrounding code rather |
| 371 | // than screwing up folding. |
| 372 | |
| 373 | while (!quote && |
| 374 | (lineNext < docLines) && |
| 375 | ((indentNext & SC_FOLDLEVELWHITEFLAG) || |
| 376 | (lineNext <= docLines && IsCommentLine(lineNext, styler)))) { |
| 377 | |
| 378 | lineNext++; |
| 379 | indentNext = styler.IndentAmount(lineNext, &spaceFlags, NULL); |
| 380 | } |
| 381 | |
| 382 | const int levelAfterComments = indentNext & SC_FOLDLEVELNUMBERMASK; |
| 383 | const int levelBeforeComments = |
| 384 | Maximum(indentCurrentLevel,levelAfterComments); |
| 385 | |
| 386 | // Now set all the indent levels on the lines we skipped |
| 387 | // Do this from end to start. Once we encounter one line |
| 388 | // which is indented more than the line after the end of |
| 389 | // the comment-block, use the level of the block before |
| 390 | |
| 391 | int skipLine = lineNext; |
| 392 | int skipLevel = levelAfterComments; |
| 393 | |
| 394 | while (--skipLine > lineCurrent) { |
| 395 | int skipLineIndent = styler.IndentAmount(skipLine, &spaceFlags, NULL); |
| 396 | |
| 397 | if ((skipLineIndent & SC_FOLDLEVELNUMBERMASK) > levelAfterComments) |
| 398 | skipLevel = levelBeforeComments; |
| 399 | |
| 400 | int whiteFlag = skipLineIndent & SC_FOLDLEVELWHITEFLAG; |
| 401 | |
| 402 | styler.SetLevel(skipLine, skipLevel | whiteFlag); |
| 403 | } |
| 404 | |
| 405 | // Set fold header on non-quote/non-comment line |
| 406 | if (!quote && !comment && !(indentCurrent & SC_FOLDLEVELWHITEFLAG) ) { |
| 407 | if ((indentCurrent & SC_FOLDLEVELNUMBERMASK) < |
| 408 | (indentNext & SC_FOLDLEVELNUMBERMASK)) |
| 409 | lev |= SC_FOLDLEVELHEADERFLAG; |
| 410 | } |
| 411 | |
| 412 | // Keep track of triple quote and block comment state of previous line |
| 413 | prevQuote = quote; |
| 414 | prevComment = comment_start || comment_continue; |
| 415 | |
| 416 | // Set fold level for this line and move to next line |
| 417 | styler.SetLevel(lineCurrent, lev); |
| 418 | indentCurrent = indentNext; |
| 419 | lineCurrent = lineNext; |
| 420 | } |
| 421 | |
| 422 | // NOTE: Cannot set level of last line here because indentCurrent doesn't have |
| 423 | // header flag set; the loop above is crafted to take care of this case! |
| 424 | //styler.SetLevel(lineCurrent, indentCurrent); |
| 425 | } |
| 426 | |
| 427 | static const char * const nimrodWordListDesc[] = { |
| 428 | "Keywords", |
| 429 | 0 |
| 430 | }; |
| 431 | |
| 432 | LexerModule lmNimrod(SCLEX_NIMROD, ColouriseNimrodDoc, "nimrod", FoldNimrodDoc, |
| 433 | nimrodWordListDesc); |