binutils/rclex.l

   1 %{ /* rclex.l -- lexer for Windows rc files parser  */
   2 /* Copyright 1997, 1998, 1999, 2001, 2002, 2003, 2005, 2007
   3    Free Software Foundation, Inc.
   4    Written by Ian Lance Taylor, Cygnus Support.
   5
   6    This file is part of GNU Binutils.
   7
   8    This program is free software; you can redistribute it and/or modify
   9    it under the terms of the GNU General Public License as published by
  10    the Free Software Foundation; either version 2 of the License, or
  11    (at your option) any later version.
  12
  13    This program is distributed in the hope that it will be useful,
  14    but WITHOUT ANY WARRANTY; without even the implied warranty of
  15    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  16    GNU General Public License for more details.
  17
  18    You should have received a copy of the GNU General Public License
  19    along with this program; if not, write to the Free Software
  20    Foundation, Inc., 51 Franklin Street - Fifth Floor, Boston, MA
  21    02110-1301, USA.  */
  22
  23 /* This is a lex input file which generates a lexer used by the
  24    Windows rc file parser.  It basically just recognized a bunch of
  25    keywords.  */
  26
  27 #include "sysdep.h"
  28 #include "bfd.h"
  29 #include "libiberty.h"
  30 #include "safe-ctype.h"
  31 #include "windres.h"
  32 #include "rcparse.h"
  33
  34 #include <assert.h>
  35
  36 #define YY_NO_UNPUT
  37
  38 /* Whether we are in rcdata mode, in which we returns the lengths of
  39    strings.  */
  40
  41 static int rcdata_mode;
  42
  43 /* Whether we are supressing lines from cpp (including windows.h or
  44    headers from your C sources may bring in externs and typedefs).
  45    When active, we return IGNORED_TOKEN, which lets us ignore these
  46    outside of resource constructs.  Thus, it isn't required to protect
  47    all the non-preprocessor lines in your header files with #ifdef
  48    RC_INVOKED.  It also means your RC file can't include other RC
  49    files if they're named "*.h".  Sorry.  Name them *.rch or whatever.  */
  50
  51 static int suppress_cpp_data;
  52
  53 #define MAYBE_RETURN(x) return suppress_cpp_data ? IGNORED_TOKEN : (x)
  54
  55 /* The first filename we detect in the cpp output.  We use this to
  56    tell included files from the original file.  */
  57
  58 static char *initial_fn;
  59
  60 /* List of allocated strings.  */
  61
  62 struct alloc_string
  63 {
  64   struct alloc_string *next;
  65   char *s;
  66 };
  67
  68 static struct alloc_string *strings;
  69
  70 /* Local functions.  */
  71
  72 static void cpp_line (const char *);
  73 static char *handle_quotes (const char *, unsigned long *);
  74 static char *get_string (int);
  75
  76 %}
  77
  78 %%
  79
  80 "BEGIN"                 { MAYBE_RETURN (BEG); }
  81 "{"                     { MAYBE_RETURN (BEG); }
  82 "END"                   { MAYBE_RETURN (END); }
  83 "}"                     { MAYBE_RETURN (END); }
  84 "ACCELERATORS"          { MAYBE_RETURN (ACCELERATORS); }
  85 "VIRTKEY"               { MAYBE_RETURN (VIRTKEY); }
  86 "ASCII"                 { MAYBE_RETURN (ASCII); }
  87 "NOINVERT"              { MAYBE_RETURN (NOINVERT); }
  88 "SHIFT"                 { MAYBE_RETURN (SHIFT); }
  89 "CONTROL"               { MAYBE_RETURN (CONTROL); }
  90 "ALT"                   { MAYBE_RETURN (ALT); }
  91 "BITMAP"                { MAYBE_RETURN (BITMAP); }
  92 "CURSOR"                { MAYBE_RETURN (CURSOR); }
  93 "DIALOG"                { MAYBE_RETURN (DIALOG); }
  94 "DIALOGEX"              { MAYBE_RETURN (DIALOGEX); }
  95 "EXSTYLE"               { MAYBE_RETURN (EXSTYLE); }
  96 "CAPTION"               { MAYBE_RETURN (CAPTION); }
  97 "CLASS"                 { MAYBE_RETURN (CLASS); }
  98 "STYLE"                 { MAYBE_RETURN (STYLE); }
  99 "AUTO3STATE"            { MAYBE_RETURN (AUTO3STATE); }
 100 "AUTOCHECKBOX"          { MAYBE_RETURN (AUTOCHECKBOX); }
 101 "AUTORADIOBUTTON"       { MAYBE_RETURN (AUTORADIOBUTTON); }
 102 "CHECKBOX"              { MAYBE_RETURN (CHECKBOX); }
 103 "COMBOBOX"              { MAYBE_RETURN (COMBOBOX); }
 104 "CTEXT"                 { MAYBE_RETURN (CTEXT); }
 105 "DEFPUSHBUTTON"         { MAYBE_RETURN (DEFPUSHBUTTON); }
 106 "EDITTEXT"              { MAYBE_RETURN (EDITTEXT); }
 107 "GROUPBOX"              { MAYBE_RETURN (GROUPBOX); }
 108 "LISTBOX"               { MAYBE_RETURN (LISTBOX); }
 109 "LTEXT"                 { MAYBE_RETURN (LTEXT); }
 110 "PUSHBOX"               { MAYBE_RETURN (PUSHBOX); }
 111 "PUSHBUTTON"            { MAYBE_RETURN (PUSHBUTTON); }
 112 "RADIOBUTTON"           { MAYBE_RETURN (RADIOBUTTON); }
 113 "RTEXT"                 { MAYBE_RETURN (RTEXT); }
 114 "SCROLLBAR"             { MAYBE_RETURN (SCROLLBAR); }
 115 "STATE3"                { MAYBE_RETURN (STATE3); }
 116 "USERBUTTON"            { MAYBE_RETURN (USERBUTTON); }
 117 "BEDIT"                 { MAYBE_RETURN (BEDIT); }
 118 "HEDIT"                 { MAYBE_RETURN (HEDIT); }
 119 "IEDIT"                 { MAYBE_RETURN (IEDIT); }
 120 "FONT"                  { MAYBE_RETURN (FONT); }
 121 "ICON"                  { MAYBE_RETURN (ICON); }
 122 "LANGUAGE"              { MAYBE_RETURN (LANGUAGE); }
 123 "CHARACTERISTICS"       { MAYBE_RETURN (CHARACTERISTICS); }
 124 "VERSION"               { MAYBE_RETURN (VERSIONK); }
 125 "MENU"                  { MAYBE_RETURN (MENU); }
 126 "MENUEX"                { MAYBE_RETURN (MENUEX); }
 127 "MENUITEM"              { MAYBE_RETURN (MENUITEM); }
 128 "SEPARATOR"             { MAYBE_RETURN (SEPARATOR); }
 129 "POPUP"                 { MAYBE_RETURN (POPUP); }
 130 "CHECKED"               { MAYBE_RETURN (CHECKED); }
 131 "GRAYED"                { MAYBE_RETURN (GRAYED); }
 132 "HELP"                  { MAYBE_RETURN (HELP); }
 133 "INACTIVE"              { MAYBE_RETURN (INACTIVE); }
 134 "MENUBARBREAK"          { MAYBE_RETURN (MENUBARBREAK); }
 135 "MENUBREAK"             { MAYBE_RETURN (MENUBREAK); }
 136 "MESSAGETABLE"          { MAYBE_RETURN (MESSAGETABLE); }
 137 "RCDATA"                { MAYBE_RETURN (RCDATA); }
 138 "STRINGTABLE"           { MAYBE_RETURN (STRINGTABLE); }
 139 "VERSIONINFO"           { MAYBE_RETURN (VERSIONINFO); }
 140 "FILEVERSION"           { MAYBE_RETURN (FILEVERSION); }
 141 "PRODUCTVERSION"        { MAYBE_RETURN (PRODUCTVERSION); }
 142 "FILEFLAGSMASK"         { MAYBE_RETURN (FILEFLAGSMASK); }
 143 "FILEFLAGS"             { MAYBE_RETURN (FILEFLAGS); }
 144 "FILEOS"                { MAYBE_RETURN (FILEOS); }
 145 "FILETYPE"              { MAYBE_RETURN (FILETYPE); }
 146 "FILESUBTYPE"           { MAYBE_RETURN (FILESUBTYPE); }
 147 "VALUE"                 { MAYBE_RETURN (VALUE); }
 148 "MOVEABLE"              { MAYBE_RETURN (MOVEABLE); }
 149 "FIXED"                 { MAYBE_RETURN (FIXED); }
 150 "PURE"                  { MAYBE_RETURN (PURE); }
 151 "IMPURE"                { MAYBE_RETURN (IMPURE); }
 152 "PRELOAD"               { MAYBE_RETURN (PRELOAD); }
 153 "LOADONCALL"            { MAYBE_RETURN (LOADONCALL); }
 154 "DISCARDABLE"           { MAYBE_RETURN (DISCARDABLE); }
 155 "NOT"                   { MAYBE_RETURN (NOT); }
 156
 157 "BLOCK"[ \t\n]*"\""[^\#\n]*"\"" {
 158                           char *s, *send;
 159
 160                           /* This is a hack to let us parse version
 161                              information easily.  */
 162
 163                           s = strchr (yytext, '"');
 164                           ++s;
 165                           send = strchr (s, '"');
 166                           if (strncmp (s, "StringFileInfo",
 167                                        sizeof "StringFileInfo" - 1) == 0
 168                               && s + sizeof "StringFileInfo" - 1 == send)
 169                             MAYBE_RETURN (BLOCKSTRINGFILEINFO);
 170                           else if (strncmp (s, "VarFileInfo",
 171                                             sizeof "VarFileInfo" - 1) == 0
 172                                    && s + sizeof "VarFileInfo" - 1 == send)
 173                             MAYBE_RETURN (BLOCKVARFILEINFO);
 174                           else
 175                             {
 176                               char *r;
 177
 178                               r = get_string (send - s + 1);
 179                               strncpy (r, s, send - s);
 180                               r[send - s] = '\0';
 181                               yylval.s = r;
 182                               MAYBE_RETURN (BLOCK);
 183                             }
 184                         }
 185
 186 "#"[^\n]*               {
 187                           cpp_line (yytext);
 188                         }
 189
 190 [0-9][x0-9A-Fa-f]*L     {
 191                           yylval.i.val = strtoul (yytext, 0, 0);
 192                           yylval.i.dword = 1;
 193                           MAYBE_RETURN (NUMBER);
 194                         }
 195
 196 [0-9][x0-9A-Fa-f]*      {
 197                           yylval.i.val = strtoul (yytext, 0, 0);
 198                           yylval.i.dword = 0;
 199                           MAYBE_RETURN (NUMBER);
 200                         }
 201
 202 ("\""[^\"\n]*"\""[ \t\n]*)+ {
 203                           char *s;
 204                           unsigned long length;
 205
 206                           s = handle_quotes (yytext, &length);
 207                           if (! rcdata_mode)
 208                             {
 209                               yylval.s = s;
 210                               MAYBE_RETURN (QUOTEDSTRING);
 211                             }
 212                           else
 213                             {
 214                               yylval.ss.length = length;
 215                               yylval.ss.s = s;
 216                               MAYBE_RETURN (SIZEDSTRING);
 217                             }
 218                         }
 219
 220 [_A-Za-z][^ ,\t\r\n]*   {
 221                           char *s;
 222
 223                           /* I rejected comma in a string in order to
 224                              handle VIRTKEY, CONTROL in an accelerator
 225                              resource.  This means that an unquoted
 226                              file name can not contain a comma.  I
 227                              don't know what rc permits.  */
 228
 229                           s = get_string (strlen (yytext) + 1);
 230                           strcpy (s, yytext);
 231                           yylval.s = s;
 232                           MAYBE_RETURN (STRING);
 233                         }
 234
 235 [\n]                    { ++rc_lineno; }
 236 [ \t\r]+                { /* ignore whitespace */ }
 237 .                       { MAYBE_RETURN (*yytext); }
 238
 239 %%
 240 #ifndef yywrap
 241 /* This is needed for some versions of lex.  */
 242 int yywrap (void)
 243 {
 244   return 1;
 245 }
 246 #endif
 247
 248 /* Handle a C preprocessor line.  */
 249
 250 static void
 251 cpp_line (const char *s)
 252 {
 253   int line;
 254   char *send, *fn;
 255
 256   ++s;
 257   while (ISSPACE (*s))
 258     ++s;
 259
 260   line = strtol (s, &send, 0);
 261   if (*send != '\0' && ! ISSPACE (*send))
 262     return;
 263
 264   /* Subtract 1 because we are about to count the newline.  */
 265   rc_lineno = line - 1;
 266
 267   s = send;
 268   while (ISSPACE (*s))
 269     ++s;
 270
 271   if (*s != '"')
 272     return;
 273
 274   ++s;
 275   send = strchr (s, '"');
 276   if (send == NULL)
 277     return;
 278
 279   fn = (char *) xmalloc (send - s + 1);
 280   strncpy (fn, s, send - s);
 281   fn[send - s] = '\0';
 282
 283   free (rc_filename);
 284   rc_filename = fn;
 285
 286   if (!initial_fn)
 287     {
 288       initial_fn = xmalloc (strlen (fn) + 1);
 289       strcpy (initial_fn, fn);
 290     }
 291
 292   /* Allow the initial file, regardless of name.  Suppress all other
 293      files if they end in ".h" (this allows included "*.rc").  */
 294   if (strcmp (initial_fn, fn) == 0
 295       || strcmp (fn + strlen (fn) - 2, ".h") != 0)
 296     suppress_cpp_data = 0;
 297   else
 298     suppress_cpp_data = 1;
 299 }
 300
 301 /* Handle a quoted string.  The quotes are stripped.  A pair of quotes
 302    in a string are turned into a single quote.  Adjacent strings are
 303    merged separated by whitespace are merged, as in C.  */
 304
 305 static char *
 306 handle_quotes (const char *input, unsigned long *len)
 307 {
 308   char *ret, *s;
 309   const char *t;
 310   int ch;
 311   int num_xdigits;
 312
 313   ret = get_string (strlen (input) + 1);
 314
 315   s = ret;
 316   t = input;
 317   if (*t == '"')
 318     ++t;
 319   while (*t != '\0')
 320     {
 321       if (*t == '\\')
 322         {
 323           ++t;
 324           switch (*t)
 325             {
 326             case '\0':
 327               rcparse_warning ("backslash at end of string");
 328               break;
 329
 330             case '\"':
 331               rcparse_warning ("use \"\" to put \" in a string");
 332               break;
 333
 334             case 'a':
 335               *s++ = ESCAPE_B; /* Strange, but true...  */
 336               ++t;
 337               break;
 338
 339             case 'b':
 340               *s++ = ESCAPE_B;
 341               ++t;
 342               break;
 343
 344             case 'f':
 345               *s++ = ESCAPE_F;
 346               ++t;
 347               break;
 348
 349             case 'n':
 350               *s++ = ESCAPE_N;
 351               ++t;
 352               break;
 353
 354             case 'r':
 355               *s++ = ESCAPE_R;
 356               ++t;
 357               break;
 358
 359             case 't':
 360               *s++ = ESCAPE_T;
 361               ++t;
 362               break;
 363
 364             case 'v':
 365               *s++ = ESCAPE_V;
 366               ++t;
 367               break;
 368
 369             case '\\':
 370               *s++ = *t++;
 371               break;
 372
 373             case '0': case '1': case '2': case '3':
 374             case '4': case '5': case '6': case '7':
 375               ch = *t - '0';
 376               ++t;
 377               if (*t >= '0' && *t <= '7')
 378                 {
 379                   ch = (ch << 3) | (*t - '0');
 380                   ++t;
 381                   if (*t >= '0' && *t <= '7')
 382                     {
 383                       ch = (ch << 3) | (*t - '0');
 384                       ++t;
 385                     }
 386                 }
 387               *s++ = ch;
 388               break;
 389
 390             case 'x':
 391               ++t;
 392               ch = 0;
 393               /* We only handle single byte chars here.  Make sure
 394                  we finish an escape sequence like "/xB0ABC" after
 395                  the first two digits.  */
 396               num_xdigits = 2;
 397               while (num_xdigits--)
 398                 {
 399                   if (*t >= '0' && *t <= '9')
 400                     ch = (ch << 4) | (*t - '0');
 401                   else if (*t >= 'a' && *t <= 'f')
 402                     ch = (ch << 4) | (*t - 'a' + 10);
 403                   else if (*t >= 'A' && *t <= 'F')
 404                     ch = (ch << 4) | (*t - 'A' + 10);
 405                   else
 406                     break;
 407                   ++t;
 408                 }
 409               *s++ = ch;
 410               break;
 411
 412             default:
 413               rcparse_warning ("unrecognized escape sequence");
 414               *s++ = '\\';
 415               *s++ = *t++;
 416               break;
 417             }
 418         }
 419       else if (*t != '"')
 420         *s++ = *t++;
 421       else if (t[1] == '\0')
 422         break;
 423       else if (t[1] == '"')
 424         {
 425           *s++ = '"';
 426           t += 2;
 427         }
 428       else
 429         {
 430           ++t;
 431           assert (ISSPACE (*t));
 432           while (ISSPACE (*t))
 433             {
 434               if ((*t) == '\n')
 435                 ++rc_lineno;
 436               ++t;
 437             }
 438           if (*t == '\0')
 439             break;
 440           assert (*t == '"');
 441           ++t;
 442         }
 443     }
 444
 445   *s = '\0';
 446
 447   *len = s - ret;
 448
 449   return ret;
 450 }
 451
 452 /* Allocate a string of a given length.  */
 453
 454 static char *
 455 get_string (int len)
 456 {
 457   struct alloc_string *as;
 458
 459   as = (struct alloc_string *) xmalloc (sizeof *as);
 460   as->s = xmalloc (len);
 461
 462   as->next = strings;
 463   strings = as;
 464
 465   return as->s;
 466 }
 467
 468 /* Discard all the strings we have allocated.  The parser calls this
 469    when it no longer needs them.  */
 470
 471 void
 472 rcparse_discard_strings (void)
 473 {
 474   struct alloc_string *as;
 475
 476   as = strings;
 477   while (as != NULL)
 478     {
 479       struct alloc_string *n;
 480
 481       free (as->s);
 482       n = as->next;
 483       free (as);
 484       as = n;
 485     }
 486
 487   strings = NULL;
 488 }
 489
 490 /* Enter rcdata mode.  */
 491
 492 void
 493 rcparse_rcdata (void)
 494 {
 495   rcdata_mode = 1;
 496 }
 497
 498 /* Go back to normal mode from rcdata mode.  */
 499
 500 void
 501 rcparse_normal (void)
 502 {
 503   rcdata_mode = 0;
 504 }