gcc/java/mangle_name.c

   1 /* Shared functions related to mangling names for the GNU compiler
   2    for the Java(TM) language.
   3    Copyright (C) 2001-2014 Free Software Foundation, Inc.
   4
   5 This file is part of GCC.
   6
   7 GCC is free software; you can redistribute it and/or modify
   8 it under the terms of the GNU General Public License as published by
   9 the Free Software Foundation; either version 3, or (at your option)
  10 any later version.
  11
  12 GCC is distributed in the hope that it will be useful,
  13 but WITHOUT ANY WARRANTY; without even the implied warranty of
  14 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  15 GNU General Public License for more details.
  16
  17 You should have received a copy of the GNU General Public License
  18 along with GCC; see the file COPYING3.  If not see
  19 <http://www.gnu.org/licenses/>.
  20
  21 Java and all Java-based marks are trademarks or registered trademarks
  22 of Sun Microsystems, Inc. in the United States and other countries.
  23 The Free Software Foundation is independent of Sun Microsystems, Inc.  */
  24
  25 /* Written by Alexandre Petit-Bianco <apbianco@cygnus.com> */
  26
  27 #include "config.h"
  28 #include "system.h"
  29 #include "coretypes.h"
  30 #include "jcf.h"
  31 #include "tree.h"
  32 #include "java-tree.h"
  33 #include "obstack.h"
  34 #include "diagnostic-core.h"
  35
  36 static void append_unicode_mangled_name (const char *, int);
  37 #ifndef HAVE_AS_UTF8
  38 static int  unicode_mangling_length (const char *, int);
  39 #endif
  40
  41 extern struct obstack *mangle_obstack;
  42
  43 static int
  44 utf8_cmp (const unsigned char *str, int length, const char *name)
  45 {
  46   const unsigned char *limit = str + length;
  47   int i;
  48
  49   for (i = 0; name[i]; ++i)
  50     {
  51       int ch = UTF8_GET (str, limit);
  52       if (ch != name[i])
  53         return ch - name[i];
  54     }
  55
  56   return str == limit ? 0 : 1;
  57 }
  58
  59 /* A sorted list of all C++ keywords.  If you change this, be sure
  60    also to change the list in
  61    libjava/classpath/tools/gnu/classpath/tools/javah/Keywords.java.  */
  62 static const char *const cxx_keywords[] =
  63 {
  64   "_Complex",
  65   "__alignof",
  66   "__alignof__",
  67   "__asm",
  68   "__asm__",
  69   "__attribute",
  70   "__attribute__",
  71   "__builtin_va_arg",
  72   "__complex",
  73   "__complex__",
  74   "__const",
  75   "__const__",
  76   "__extension__",
  77   "__imag",
  78   "__imag__",
  79   "__inline",
  80   "__inline__",
  81   "__label__",
  82   "__null",
  83   "__real",
  84   "__real__",
  85   "__restrict",
  86   "__restrict__",
  87   "__signed",
  88   "__signed__",
  89   "__typeof",
  90   "__typeof__",
  91   "__volatile",
  92   "__volatile__",
  93   "and",
  94   "and_eq",
  95   "asm",
  96   "auto",
  97   "bitand",
  98   "bitor",
  99   "bool",
 100   "break",
 101   "case",
 102   "catch",
 103   "char",
 104   "class",
 105   "compl",
 106   "const",
 107   "const_cast",
 108   "continue",
 109   "default",
 110   "delete",
 111   "do",
 112   "double",
 113   "dynamic_cast",
 114   "else",
 115   "enum",
 116   "explicit",
 117   "export",
 118   "extern",
 119   "false",
 120   "float",
 121   "for",
 122   "friend",
 123   "goto",
 124   "if",
 125   "inline",
 126   "int",
 127   "long",
 128   "mutable",
 129   "namespace",
 130   "new",
 131   "not",
 132   "not_eq",
 133   "operator",
 134   "or",
 135   "or_eq",
 136   "private",
 137   "protected",
 138   "public",
 139   "register",
 140   "reinterpret_cast",
 141   "return",
 142   "short",
 143   "signed",
 144   "sizeof",
 145   "static",
 146   "static_cast",
 147   "struct",
 148   "switch",
 149   "template",
 150   "this",
 151   "throw",
 152   "true",
 153   "try",
 154   "typedef",
 155   "typeid",
 156   "typename",
 157   "typeof",
 158   "union",
 159   "unsigned",
 160   "using",
 161   "virtual",
 162   "void",
 163   "volatile",
 164   "wchar_t",
 165   "while",
 166   "xor",
 167   "xor_eq"
 168 };
 169
 170 /* Return true if NAME is a C++ keyword.  */
 171 int
 172 cxx_keyword_p (const char *name, int length)
 173 {
 174   int last = ARRAY_SIZE (cxx_keywords);
 175   int first = 0;
 176   int mid = (last + first) / 2;
 177   int old = -1;
 178
 179   for (mid = (last + first) / 2;
 180        mid != old;
 181        old = mid, mid = (last + first) / 2)
 182     {
 183       int kwl = strlen (cxx_keywords[mid]);
 184       int min_length = kwl > length ? length : kwl;
 185       int r = utf8_cmp ((const unsigned char *) name, min_length, cxx_keywords[mid]);
 186
 187       if (r == 0)
 188         {
 189           int i;
 190           /* We've found a match if all the remaining characters are `$'.  */
 191           for (i = min_length; i < length && name[i] == '$'; ++i)
 192             ;
 193           if (i == length)
 194             return 1;
 195           r = 1;
 196         }
 197
 198       if (r < 0)
 199         last = mid;
 200       else
 201         first = mid;
 202     }
 203   return 0;
 204 }
 205
 206 /* If NAME happens to be a C++ keyword, add `$'.  */
 207 #define MANGLE_CXX_KEYWORDS(NAME, LEN)                  \
 208 do                                                      \
 209   {                                                     \
 210     if (cxx_keyword_p ((NAME), (LEN)))                  \
 211       {                                                 \
 212         char *tmp_buf = (char *)alloca ((LEN)+1);       \
 213         memcpy (tmp_buf, (NAME), (LEN));                \
 214         tmp_buf[LEN]= '$';                              \
 215         (NAME) = tmp_buf;                               \
 216         (LEN)++;                                        \
 217       }                                                 \
 218   }                                                     \
 219 while (0)
 220
 221
 222 /* If the assembler doesn't support UTF8 in symbol names, some
 223    characters might need to be escaped.  */
 224
 225 #ifndef HAVE_AS_UTF8
 226
 227 /* Assuming (NAME, LEN) is a Utf8-encoding string, emit the string
 228    appropriately mangled (with Unicode escapes if needed) to
 229    MANGLE_OBSTACK.  Note that `java', `lang' and `Object' are used so
 230    frequently that they could be cached.  */
 231
 232 void
 233 append_gpp_mangled_name (const char *name, int len)
 234 {
 235   int encoded_len, needs_escapes;
 236   char buf[6];
 237
 238   MANGLE_CXX_KEYWORDS (name, len);
 239
 240   encoded_len = unicode_mangling_length (name, len);
 241   needs_escapes = encoded_len > 0;
 242
 243   sprintf (buf, "%d", (needs_escapes ? encoded_len : len));
 244   obstack_grow (mangle_obstack, buf, strlen (buf));
 245
 246   if (needs_escapes)
 247     append_unicode_mangled_name (name, len);
 248   else
 249     obstack_grow (mangle_obstack, name, len);
 250 }
 251
 252 /* Assuming (NAME, LEN) is a Utf8-encoded string, emit the string
 253    appropriately mangled (with Unicode escapes) to MANGLE_OBSTACK.
 254    Characters needing an escape are encoded `__UNN_' to `__UNNNN_', in
 255    which case `__U' will be mangled `__U_'.  */
 256
 257 static void
 258 append_unicode_mangled_name (const char *name, int len)
 259 {
 260   const unsigned char *ptr;
 261   const unsigned char *limit = (const unsigned char *)name + len;
 262   int uuU = 0;
 263   for (ptr = (const unsigned char *) name;  ptr < limit;  )
 264     {
 265       int ch = UTF8_GET(ptr, limit);
 266
 267       if ((ISALNUM (ch) && ch != 'U') || ch == '$')
 268         {
 269           obstack_1grow (mangle_obstack, ch);
 270           uuU = 0;
 271         }
 272       /* Everything else needs encoding */
 273       else
 274         {
 275           char buf [9];
 276           if (ch == '_' || ch == 'U')
 277             {
 278               /* Prepare to recognize __U */
 279               if (ch == '_' && (uuU < 3))
 280                 {
 281                   uuU++;
 282                   obstack_1grow (mangle_obstack, ch);
 283                 }
 284               /* We recognize __U that we wish to encode
 285                  __U_. Finish the encoding. */
 286               else if (ch == 'U' && (uuU == 2))
 287                 {
 288                   uuU = 0;
 289                   obstack_grow (mangle_obstack, "U_", 2);
 290                 }
 291               /* Otherwise, just reset uuU and emit the character we
 292                  have. */
 293               else
 294                 {
 295                   uuU = 0;
 296                   obstack_1grow (mangle_obstack, ch);
 297                 }
 298               continue;
 299             }
 300           sprintf (buf, "__U%x_", ch);
 301           obstack_grow (mangle_obstack, buf, strlen (buf));
 302           uuU = 0;
 303         }
 304     }
 305 }
 306
 307 /* Assuming (NAME, LEN) is a Utf8-encoding string, calculate the
 308    length of the string as mangled (a la g++) including Unicode
 309    escapes.  If no escapes are needed, return 0.  */
 310
 311 static int
 312 unicode_mangling_length (const char *name, int len)
 313 {
 314   const unsigned char *ptr;
 315   const unsigned char *limit = (const unsigned char *)name + len;
 316   int need_escapes = 0;         /* Whether we need an escape or not */
 317   int num_chars = 0;            /* Number of characters in the mangled name */
 318   int uuU = 0;                  /* Help us to find __U. 0: '_', 1: '__' */
 319   for (ptr = (const unsigned char *) name;  ptr < limit;  )
 320     {
 321       int ch = UTF8_GET(ptr, limit);
 322
 323       if (ch < 0)
 324         error ("internal error - invalid Utf8 name");
 325       if ((ISALNUM (ch) && ch != 'U') || ch == '$')
 326         {
 327           num_chars++;
 328           uuU = 0;
 329         }
 330       /* Everything else needs encoding */
 331       else
 332         {
 333           int encoding_length = 2;
 334
 335           if (ch == '_' || ch == 'U')
 336             {
 337               /* It's always at least one character. */
 338               num_chars++;
 339
 340               /* Prepare to recognize __U */
 341               if (ch == '_' && (uuU < 3))
 342                 uuU++;
 343
 344               /* We recognize __U that we wish to encode __U_, we
 345                  count one more character. */
 346               else if (ch == 'U' && (uuU == 2))
 347                 {
 348                   num_chars++;
 349                   need_escapes = 1;
 350                   uuU = 0;
 351                 }
 352               /* Otherwise, just reset uuU */
 353               else
 354                 uuU = 0;
 355
 356               continue;
 357             }
 358
 359           if (ch > 0xff)
 360             encoding_length++;
 361           if (ch > 0xfff)
 362             encoding_length++;
 363
 364           num_chars += (4 + encoding_length);
 365           need_escapes = 1;
 366           uuU = 0;
 367         }
 368     }
 369   if (need_escapes)
 370     return num_chars;
 371   else
 372     return 0;
 373 }
 374
 375 #else
 376
 377 /* The assembler supports UTF8, we don't use escapes. Mangling is
 378    simply <N>NAME. <N> is the number of UTF8 encoded characters that
 379    are found in NAME. Note that `java', `lang' and `Object' are used
 380    so frequently that they could be cached.  */
 381
 382 void
 383 append_gpp_mangled_name (const char *name, int len)
 384 {
 385   const unsigned char *ptr;
 386   const unsigned char *limit;
 387   int encoded_len;
 388   char buf [6];
 389
 390   MANGLE_CXX_KEYWORDS (name, len);
 391
 392   limit = (const unsigned char *)name + len;
 393
 394   /* Compute the length of the string we wish to mangle. */
 395   for (encoded_len =  0, ptr = (const unsigned char *) name;
 396        ptr < limit; encoded_len++)
 397     {
 398       int ch = UTF8_GET(ptr, limit);
 399
 400       if (ch < 0)
 401         error ("internal error - invalid Utf8 name");
 402     }
 403
 404   sprintf (buf, "%d", encoded_len);
 405   obstack_grow (mangle_obstack, buf, strlen (buf));
 406   obstack_grow (mangle_obstack, name, len);
 407 }
 408
 409 #endif /* HAVE_AS_UTF8 */