3 * Copyright 1998-2000 Bertho A. Stultiens (BS)
5 * This library is free software; you can redistribute it and/or
6 * modify it under the terms of the GNU Lesser General Public
7 * License as published by the Free Software Foundation; either
8 * version 2.1 of the License, or (at your option) any later version.
10 * This library is distributed in the hope that it will be useful,
11 * but WITHOUT ANY WARRANTY; without even the implied warranty of
12 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
13 * Lesser General Public License for more details.
15 * You should have received a copy of the GNU Lesser General Public
16 * License along with this library; if not, write to the Free Software
17 * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301, USA
20 * 21-May-2000 BS - Fixed the ident requirement of resource names
21 * which can be keywords.
22 * 30-Apr-2000 BS - Reintegration into the wine-tree
23 * 11-Jan-2000 BS - Very drastic cleanup because we don't have a
24 * preprocessor in here anymore.
25 * 02-Jan-2000 BS - Removed the preprocessor code
26 * 23-Dec-1999 BS - Removed the copyright for Martin von Loewis.
27 * There is really nothing left of his code in
29 * 20-Jun-1998 BS - Changed the filename conversion. Filenames are
30 * case-sensitive inder *nix, but not under dos.
31 * default behaviour is to convert to lower case.
32 * - All backslashes are converted to forward and
33 * both single and double slash is recognized as
35 * - Fixed a bug in 'yywf' case that prevented
36 * double quoted names to be scanned propperly.
38 * 19-May-1998 BS - Started to build a preprocessor.
39 * - Changed keyword processing completely to
42 * 20-Apr-1998 BS - Added ';' comment stripping
44 * 17-Apr-1998 BS - Made the win32 keywords optional when compiling in
47 * 15-Apr-1998 BS - Changed string handling to include escapes
48 * - Added unicode string handling (no codepage
49 * translation though).
50 * - 'Borrowed' the main idea of string scanning from
51 * the flex manual pages.
52 * - Added conditional handling of scanning depending
53 * on the state of the parser. This was mainly required
54 * to distinguish a file to load or raw data that
55 * follows. MS's definition of filenames is rather
56 * complex... It can be unquoted or double quoted. If
57 * double quoted, then the '\\' char is not automatically
58 * escaped according to Borland's rc compiler, but it
59 * accepts both "\\path\\file.rc" and "\path\file.rc".
60 * This makes life very hard! I go for the escaped
61 * version, as this seems to be the documented way...
62 * - Single quoted strings are now parsed and converted
64 * - Added comment stripping. The implementation is
65 * 'borrowed' from the flex manpages.
66 * - Rebuild string processing so that it may contain
70 /* Exclusive string handling */
72 /* Exclusive unicode string handling */
74 /* Exclusive rcdata single quoted data handling */
76 /* Exclusive comment eating... */
78 /* Set when stripping c-junk */
83 /* Set when scanning #line style directives */
85 /* Set when scanning #pragma */
90 %option nounput noyy_top_state
91 %option 8bit never-interactive
92 %option prefix="parser_"
94 /* Some shortcut definitions */
96 cident [a-zA-Z_][0-9a-zA-Z_]*
100 /*#define LEX_DEBUG*/
108 #include "wine/unicode.h"
112 #include "newstruc.h"
114 #include "parser.tab.h"
116 /* Always update the current character position within a line */
117 #define YY_USER_ACTION char_number+=yyleng; wanted_id = want_id; want_id = 0;
119 static void addcchar(char c);
120 static void addwchar(WCHAR s);
121 static string_t *get_buffered_cstring(void);
122 static string_t *get_buffered_wstring(void);
123 static string_t *make_string(char *s);
125 static char *cbuffer; /* Buffers for string collection */
127 static int cbufalloc = 0;
128 static WCHAR *wbuffer;
130 static int wbufalloc = 0;
131 static int stripslevel = 0; /* Count {} during pp_strips/pp_stripe mode */
132 static int stripplevel = 0; /* Count () during pp_strips mode */
133 static int cjunk_tagline; /* Where did we start stripping (helps error tracking) */
135 static int current_codepage = -1; /* use language default */
138 * This one is a bit tricky.
139 * We set 'want_id' in the parser to get the first
140 * identifier we get across in the scanner, but we
141 * also want it to be reset at nearly any token we
142 * see. Exceptions are:
147 * The scanner will automatically reset 'want_id'
148 * after *each* scanner reduction and puts is value
149 * into the var below. In this way we can see the
150 * state after the YY_RULE_SETUP (i.e. the user action;
151 * see above) and don't have to worry too much when
152 * it needs to be reset.
154 static int wanted_id = 0;
155 static int save_wanted_id; /* To save across comment reductions */
165 static struct keyword keywords[] = {
166 { "ACCELERATORS", tACCELERATORS, 0, 0, 0},
167 { "ALT", tALT, 0, 0, 0},
168 { "ASCII", tASCII, 0, 0, 0},
169 { "AUTO3STATE", tAUTO3STATE, 1, 0, 0},
170 { "AUTOCHECKBOX", tAUTOCHECKBOX, 1, 0, 0},
171 { "AUTORADIOBUTTON", tAUTORADIOBUTTON, 1, 0, 0},
172 { "BEGIN", tBEGIN, 0, 0, 0},
173 { "BITMAP", tBITMAP, 0, 0, 0},
174 { "BLOCK", tBLOCK, 0, 0, 0},
175 { "BUTTON", tBUTTON, 1, 0, 0},
176 { "CAPTION", tCAPTION, 0, 0, 0},
177 { "CHARACTERISTICS", tCHARACTERISTICS, 1, 0, 0},
178 { "CHECKBOX", tCHECKBOX, 0, 0, 0},
179 { "CHECKED", tCHECKED, 0, 0, 0},
180 { "CLASS", tCLASS, 0, 0, 0},
181 { "COMBOBOX", tCOMBOBOX, 0, 0, 0},
182 { "CONTROL", tCONTROL, 0, 0, 0},
183 { "CTEXT", tCTEXT, 0, 0, 0},
184 { "CURSOR", tCURSOR, 0, 0, 0},
185 { "DEFPUSHBUTTON", tDEFPUSHBUTTON, 0, 0, 0},
186 { "DIALOG", tDIALOG, 0, 0, 0},
187 { "DIALOGEX", tDIALOGEX, 1, 0, 0},
188 { "DISCARDABLE", tDISCARDABLE, 0, 0, 0},
189 { "DLGINIT", tDLGINIT, 0, 0, 0},
190 { "EDITTEXT", tEDITTEXT, 0, 0, 0},
191 { "END", tEND, 0, 0, 0},
192 { "EXSTYLE", tEXSTYLE, 0, 0, 0},
193 { "FILEFLAGS", tFILEFLAGS, 0, 0, 0},
194 { "FILEFLAGSMASK", tFILEFLAGSMASK, 0, 0, 0},
195 { "FILEOS", tFILEOS, 0, 0, 0},
196 { "FILESUBTYPE", tFILESUBTYPE, 0, 0, 0},
197 { "FILETYPE", tFILETYPE, 0, 0, 0},
198 { "FILEVERSION", tFILEVERSION, 0, 0, 0},
199 { "FIXED", tFIXED, 0, 0, 0},
200 { "FONT", tFONT, 0, 0, 0},
201 { "FONTDIR", tFONTDIR, 0, 0, 0}, /* This is a Borland BRC extension */
202 { "GRAYED", tGRAYED, 0, 0, 0},
203 { "GROUPBOX", tGROUPBOX, 0, 0, 0},
204 { "HELP", tHELP, 0, 0, 0},
205 { "HTML", tHTML, 0, 0, 0},
206 { "ICON", tICON, 0, 0, 0},
207 { "IMPURE", tIMPURE, 0, 0, 0},
208 { "INACTIVE", tINACTIVE, 0, 0, 0},
209 { "LANGUAGE", tLANGUAGE, 1, 0, 1},
210 { "LISTBOX", tLISTBOX, 0, 0, 0},
211 { "LOADONCALL", tLOADONCALL, 0, 0, 0},
212 { "LTEXT", tLTEXT, 0, 0, 0},
213 { "MENU", tMENU, 0, 0, 0},
214 { "MENUBARBREAK", tMENUBARBREAK, 0, 0, 0},
215 { "MENUBREAK", tMENUBREAK, 0, 0, 0},
216 { "MENUEX", tMENUEX, 1, 0, 0},
217 { "MENUITEM", tMENUITEM, 0, 0, 0},
218 { "MESSAGETABLE", tMESSAGETABLE, 1, 0, 0},
219 { "MOVEABLE", tMOVEABLE, 0, 0, 0},
220 { "NOINVERT", tNOINVERT, 0, 0, 0},
221 { "NOT", tNOT, 0, 0, 0},
222 { "POPUP", tPOPUP, 0, 0, 0},
223 { "PRELOAD", tPRELOAD, 0, 0, 0},
224 { "PRODUCTVERSION", tPRODUCTVERSION, 0, 0, 0},
225 { "PURE", tPURE, 0, 0, 0},
226 { "PUSHBUTTON", tPUSHBUTTON, 0, 0, 0},
227 { "RADIOBUTTON", tRADIOBUTTON, 0, 0, 0},
228 { "RCDATA", tRCDATA, 0, 0, 0},
229 { "RTEXT", tRTEXT, 0, 0, 0},
230 { "SCROLLBAR", tSCROLLBAR, 0, 0, 0},
231 { "SEPARATOR", tSEPARATOR, 0, 0, 0},
232 { "SHIFT", tSHIFT, 0, 0, 0},
233 { "STATE3", tSTATE3, 1, 0, 0},
234 { "STRING", tSTRING, 0, 0, 0},
235 { "STRINGTABLE", tSTRINGTABLE, 0, 0, 1},
236 { "STYLE", tSTYLE, 0, 0, 0},
237 { "TOOLBAR", tTOOLBAR, 1, 0, 0},
238 { "VALUE", tVALUE, 0, 0, 0},
239 { "VERSION", tVERSION, 1, 0, 0},
240 { "VERSIONINFO", tVERSIONINFO, 0, 0, 0},
241 { "VIRTKEY", tVIRTKEY, 0, 0, 0}
244 #define NKEYWORDS (sizeof(keywords)/sizeof(keywords[0]))
245 #define KWP(p) ((const struct keyword *)(p))
246 static int kw_cmp_func(const void *s1, const void *s2)
249 ret = strcasecmp(KWP(s1)->keyword, KWP(s2)->keyword);
250 if(!ret && (KWP(s1)->needcase || KWP(s2)->needcase))
251 return strcmp(KWP(s1)->keyword, KWP(s2)->keyword);
258 static struct keyword *iskeyword(char *kw)
266 /* Make sure that it is sorted for bsearsh */
267 static int sorted = 0;
270 qsort(keywords, NKEYWORDS, sizeof(keywords[0]), kw_cmp_func);
276 kwp = bsearch(&key, keywords, NKEYWORDS, sizeof(keywords[0]), kw_cmp_func);
280 for(i = 0; i < NKEYWORDS; i++)
282 if(!kw_cmp_func(&key, &keywords[i]))
292 if(kwp == NULL || (kwp->isextension && !extensions))
301 **************************************************************************
302 * The flexer starts here
303 **************************************************************************
307 * Catch the GCC-style line statements here and parse them.
308 * This has the advantage that you can #include at any
309 * stage in the resource file.
310 * The preprocessor generates line directives in the format:
311 * # <linenum> "filename" <codes>
313 * Codes can be a sequence of:
314 * - 1 start of new file
315 * - 2 returning to previous
317 * - 4 interpret as C-code
319 * 4 is not used and 1 mutually excludes 2
320 * Anyhow, we are not really interested in these at all
321 * because we only want to know the linenumber and
324 <INITIAL,pp_strips,pp_stripp>^{ws}*\#{ws}*pragma{ws}+ yy_push_state(pp_pragma);
325 <INITIAL,pp_strips,pp_stripp>^{ws}*\#{ws}* yy_push_state(pp_line);
331 lineno = (int)strtol(yytext, &cptr, 10);
333 parser_error("Malformed '#...' line-directive; invalid linenumber");
334 fname = strchr(cptr, '"');
336 parser_error("Malformed '#...' line-directive; missing filename");
338 cptr = strchr(fname, '"');
340 parser_error("Malformed '#...' line-directive; missing terminating \"");
342 line_number = lineno - 1; /* We didn't read the newline */
343 input_name = xstrdup(fname);
346 <pp_pragma>code_page[^\n]* yyless(9); yy_pop_state(); yy_push_state(pp_code_page);
347 <pp_pragma>[^\n]* yy_pop_state(); if (pedantic) parser_warning("Unrecognized #pragma directive '%s'",yytext);
349 <pp_code_page>\({ws}*default{ws}*\)[^\n]* current_codepage = -1; yy_pop_state();
350 <pp_code_page>\({ws}*[0-9]+{ws}*\)[^\n]* {
353 while (*p < '0' || *p > '9') p++;
354 current_codepage = strtol( p, NULL, 10 );
355 if (current_codepage && !wine_cp_get_table( current_codepage ))
357 parser_error("Codepage %d not supported", current_codepage);
358 current_codepage = 0;
361 <pp_code_page>[^\n]* yy_pop_state(); parser_error("Malformed #pragma code_page directive");
364 * Strip everything until a ';' taking
365 * into account braces {} for structures,
368 <pp_strips>\{ stripslevel++;
369 <pp_strips>\} stripslevel--;
370 <pp_strips>; if(!stripslevel) yy_pop_state();
371 <pp_strips>\/[^*\n] ; /* To catch comments */
372 <pp_strips>[^\{\};\n#/]* ; /* Ignore rest */
373 <pp_strips>\n line_number++; char_number = 1;
375 <pp_stripp>\( stripplevel++;
381 yy_push_state(pp_stripp_final);
384 <pp_stripp>\/[^*\n] ; /* To catch comments */
385 <pp_stripp>[^\(\);\n#/]* ; /* Ignore rest */
386 <pp_stripp>\n line_number++; char_number = 1;
388 <pp_stripp_final>{ws}* ; /* Ignore */
389 <pp_stripp_final>; yy_pop_state(); /* Kill the semicolon */
390 <pp_stripp_final>\n line_number++; char_number = 1; yy_pop_state();
391 <pp_stripp_final>. yyless(0); yy_pop_state();
396 [0-9]+[lL]? { parser_lval.num = strtoul(yytext, 0, 10); return toupper(yytext[yyleng-1]) == 'L' ? tLNUMBER : tNUMBER; }
397 0[xX][0-9A-Fa-f]+[lL]? { parser_lval.num = strtoul(yytext, 0, 16); return toupper(yytext[yyleng-1]) == 'L' ? tLNUMBER : tNUMBER; }
398 0[oO][0-7]+[lL]? { parser_lval.num = strtoul(yytext+2, 0, 8); return toupper(yytext[yyleng-1]) == 'L' ? tLNUMBER : tNUMBER; }
401 * The next two rules scan identifiers and filenames.
402 * This is achieved by using the priority ruling
403 * of the scanner where a '.' is valid in a filename
404 * and *only* in a filename. In this case, the second
405 * rule will be reduced because it is longer.
408 struct keyword *tok = iskeyword(yytext);
412 if(wanted_id && !tok->alwayskw)
414 parser_lval.str = make_string(yytext);
422 parser_lval.str = make_string(yytext);
426 [A-Za-z_0-9./\\]+ parser_lval.str = make_string(yytext); return tFILENAME;
429 * Wide string scanning
432 yy_push_state(tklstr);
435 parser_warning("16bit resource contains unicode strings\n");
440 parser_lval.str = get_buffered_wstring();
443 <tklstr>\\[0-7]{1,6} { /* octal escape sequence */
445 result = strtoul(yytext+1, 0, 8);
446 if ( result > 0xffff )
447 parser_error("Character constant out of range");
448 addwchar((WCHAR)result);
450 <tklstr>\\x[0-9a-fA-F]{4} { /* hex escape sequence */
452 result = strtoul(yytext+2, 0, 16);
453 addwchar((WCHAR)result);
455 <tklstr>\\x[0-9a-fA-F]{1,3} { parser_error("Invalid hex escape sequence '%s'", yytext); }
457 <tklstr>\\[0-9]+ parser_error("Bad escape sequence");
458 <tklstr>\\\n{ws}* line_number++; char_number = 1; /* backslash at EOL continues string after leading whitespace on next line */
459 <tklstr>\\a addwchar('\a');
460 <tklstr>\\b addwchar('\b');
461 <tklstr>\\f addwchar('\f');
462 <tklstr>\\n addwchar('\n');
463 <tklstr>\\r addwchar('\r');
464 <tklstr>\\t addwchar('\t');
465 <tklstr>\\v addwchar('\v');
466 <tklstr>\\. addwchar(yytext[1]);
467 <tklstr>\\\r\n addwchar(yytext[2]); line_number++; char_number = 1;
468 <tklstr>\"\" addwchar('\"'); /* "bla""bla" -> "bla\"bla" */
469 <tklstr>\\\"\" addwchar('\"'); /* "bla\""bla" -> "bla\"bla" */
470 <tklstr>\"{ws}+\" ; /* "bla" "bla" -> "blabla" */
473 while(*yptr) /* FIXME: codepage translation */
474 addwchar(*yptr++ & 0xff);
476 <tklstr>\n parser_error("Unterminated string");
479 * Normal string scanning
481 \" yy_push_state(tkstr); cbufidx = 0;
485 parser_lval.str = get_buffered_cstring();
488 <tkstr>\\[0-7]{1,3} { /* octal escape sequence */
490 result = strtol(yytext+1, 0, 8);
492 parser_error("Character constant out of range");
493 addcchar((char)result);
495 <tkstr>\\x[0-9a-fA-F]{2} { /* hex escape sequence */
497 result = strtol(yytext+2, 0, 16);
498 addcchar((char)result);
500 <tkstr>\\x[0-9a-fA-F] { parser_error("Invalid hex escape sequence '%s'", yytext); }
502 <tkstr>\\[0-9]+ parser_error("Bad escape sequence");
503 <tkstr>\\\n{ws}* line_number++; char_number = 1; /* backslash at EOL continues string after leading whitespace on next line */
504 <tkstr>\\a addcchar('\a');
505 <tkstr>\\b addcchar('\b');
506 <tkstr>\\f addcchar('\f');
507 <tkstr>\\n addcchar('\n');
508 <tkstr>\\r addcchar('\r');
509 <tkstr>\\t addcchar('\t');
510 <tkstr>\\v addcchar('\v');
511 <tkstr>\\. addcchar(yytext[1]);
512 <tkstr>\\\r\n addcchar(yytext[2]); line_number++; char_number = 1;
518 <tkstr>\"\" addcchar('\"'); /* "bla""bla" -> "bla\"bla" */
519 <tkstr>\\\"\" addcchar('\"'); /* "bla\""bla" -> "bla\"bla" */
520 <tkstr>\"{ws}+\" ; /* "bla" "bla" -> "blabla" */
521 <tkstr>\n parser_error("Unterminated string");
526 \' yy_push_state(tkrcd); cbufidx = 0;
529 parser_lval.raw = new_raw_data();
530 parser_lval.raw->size = cbufidx;
531 parser_lval.raw->data = xmalloc(parser_lval.raw->size);
532 memcpy(parser_lval.raw->data, cbuffer, parser_lval.raw->size);
535 <tkrcd>[0-9a-fA-F]{2} {
537 result = strtol(yytext, 0, 16);
538 addcchar((char)result);
540 <tkrcd>{ws}+ ; /* Ignore space */
541 <tkrcd>\n line_number++; char_number = 1;
542 <tkrcd>. parser_error("Malformed data-line");
546 * Should never occur after preprocessing
548 <INITIAL,pp_stripp,pp_strips>"/*" {
549 yy_push_state(comment);
550 save_wanted_id = wanted_id;
552 parser_warning("Found comments after preprocessing, please report");
555 <comment>"*"+[^*/\n]* ;
556 <comment>\n line_number++; char_number = 1;
557 <comment>"*"+"/" yy_pop_state(); want_id = save_wanted_id;
559 ;[^\n]* want_id = wanted_id; /* not really comment, but left-over c-junk */
560 "//"[^\n]* want_id = wanted_id; if(!no_preprocess) parser_warning("Found comments after preprocessing, please report");
572 {ws}+ want_id = wanted_id; /* Eat whitespace */
574 <INITIAL>. return yytext[0];
577 if(YY_START == pp_strips || YY_START == pp_stripe || YY_START == pp_stripp || YY_START == pp_stripp_final)
578 parser_error("Unexpected end of file during c-junk scanning (started at %d)", cjunk_tagline);
584 /* Catch all rule to find any unmatched text */
590 parser_warning("Unmatched text '%c' (0x%02x) YY_START=%d stripslevel=%d",
591 isprint(*yytext & 0xff) ? *yytext : '.', *yytext, YY_START,stripslevel);
597 int parser_wrap(void)
600 if(bufferstackidx > 0)
609 /* These dup functions copy the enclosed '\0' from
610 * the resource string.
612 static void addcchar(char c)
614 if(cbufidx >= cbufalloc)
617 cbuffer = xrealloc(cbuffer, cbufalloc * sizeof(cbuffer[0]));
618 if(cbufalloc > 65536)
619 parser_warning("Reallocating string buffer larger than 64kB");
621 cbuffer[cbufidx++] = c;
624 static void addwchar(WCHAR s)
626 if(wbufidx >= wbufalloc)
629 wbuffer = xrealloc(wbuffer, wbufalloc * sizeof(wbuffer[0]));
630 if(wbufalloc > 65536)
631 parser_warning("Reallocating wide string buffer larger than 64kB");
633 wbuffer[wbufidx++] = s;
636 static string_t *get_buffered_cstring(void)
638 string_t *str = new_string();
641 str->type = str_char;
642 str->str.cstr = xmalloc(cbufidx+1);
643 memcpy(str->str.cstr, cbuffer, cbufidx);
644 str->str.cstr[cbufidx] = '\0';
646 if (!current_codepage || current_codepage == -1 || !win32) /* store as ANSI string */
648 if (!current_codepage) parser_error("Codepage set to Unicode only, cannot use ASCII string here");
651 else /* convert to Unicode before storing */
653 string_t *str_w = convert_string( str, str_unicode, current_codepage );
654 if (!check_unicode_conversion( str, str_w, current_codepage ))
655 parser_error("String %s does not convert identically to Unicode and back in codepage %d. "
656 "Try using a Unicode string instead.", str->str.cstr, current_codepage );
662 static string_t *get_buffered_wstring(void)
664 string_t *str = new_string();
666 str->type = str_unicode;
667 str->str.wstr = xmalloc((wbufidx+1)*sizeof(WCHAR));
668 memcpy(str->str.wstr, wbuffer, wbufidx*sizeof(WCHAR));
669 str->str.wstr[wbufidx] = 0;
673 static string_t *make_string(char *s)
675 string_t *str = new_string();
676 str->size = strlen(s);
677 str->type = str_char;
678 str->str.cstr = xmalloc(str->size+1);
679 memcpy(str->str.cstr, s, str->size+1);