Analyze.cpp

   1 /* Cycript - The Truly Universal Scripting Language
   2  * Copyright (C) 2009-2016  Jay Freeman (saurik)
   3 */
   4
   5 /* GNU Affero General Public License, Version 3 {{{ */
   6 /*
   7  * This program is free software: you can redistribute it and/or modify
   8  * it under the terms of the GNU Affero General Public License as published by
   9  * the Free Software Foundation, either version 3 of the License, or
  10  * (at your option) any later version.
  11
  12  * This program 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 Affero General Public License for more details.
  16
  17  * You should have received a copy of the GNU Affero General Public License
  18  * along with this program.  If not, see <http://www.gnu.org/licenses/>.
  19 **/
  20 /* }}} */
  21
  22 #include <cmath>
  23 #include <cstring>
  24 #include <iostream>
  25 #include <map>
  26 #include <sstream>
  27 #include <string>
  28
  29 #include <clang-c/Index.h>
  30
  31 #include "Functor.hpp"
  32 #include "Replace.hpp"
  33 #include "Syntax.hpp"
  34
  35 static CXChildVisitResult CYVisit(CXCursor cursor, CXCursor parent, CXClientData arg) {
  36     (*reinterpret_cast<const Functor<void (CXCursor)> *>(arg))(cursor);
  37     return CXChildVisit_Continue;
  38 }
  39
  40 static unsigned CYForChild(CXCursor cursor, const Functor<void (CXCursor)> &visitor) {
  41     return clang_visitChildren(cursor, &CYVisit, const_cast<void *>(static_cast<const void *>(&visitor)));
  42 }
  43
  44 static bool CYOneChild(CXCursor cursor, const Functor<void (CXCursor)> &visitor) {
  45     bool visited(false);
  46     CYForChild(cursor, fun([&](CXCursor child) {
  47         _assert(!visited);
  48         visited = true;
  49         visitor(child);
  50     }));
  51     return visited;
  52 }
  53
  54 struct CYCXString {
  55     CXString value_;
  56
  57     CYCXString(CXString value) :
  58         value_(value)
  59     {
  60     }
  61
  62     CYCXString(CXCursor cursor) :
  63         value_(clang_getCursorSpelling(cursor))
  64     {
  65     }
  66
  67     CYCXString(CXCursorKind kind) :
  68         value_(clang_getCursorKindSpelling(kind))
  69     {
  70     }
  71
  72     CYCXString(CXFile file) :
  73         value_(clang_getFileName(file))
  74     {
  75     }
  76
  77     CYCXString(CXTranslationUnit unit, CXToken token) :
  78         value_(clang_getTokenSpelling(unit, token))
  79     {
  80     }
  81
  82     ~CYCXString() {
  83         clang_disposeString(value_);
  84     }
  85
  86     operator const char *() const {
  87         return clang_getCString(value_);
  88     }
  89
  90     const char *Pool(CYPool &pool) const {
  91         return pool.strdup(*this);
  92     }
  93
  94     bool operator ==(const char *rhs) const {
  95         const char *lhs(*this);
  96         return lhs == rhs || strcmp(lhs, rhs) == 0;
  97     }
  98 };
  99
 100 template <void (&clang_get_Location)(CXSourceLocation, CXFile *, unsigned *, unsigned *, unsigned *) = clang_getSpellingLocation>
 101 struct CYCXPosition {
 102     CXFile file_;
 103     unsigned line_;
 104     unsigned column_;
 105     unsigned offset_;
 106
 107     CYCXPosition(CXSourceLocation location) {
 108         clang_get_Location(location, &file_, &line_, &column_, &offset_);
 109     }
 110
 111     CYCXPosition(CXTranslationUnit unit, CXToken token) :
 112         CYCXPosition(clang_getTokenLocation(unit, token))
 113     {
 114     }
 115
 116     CXSourceLocation Get(CXTranslationUnit unit) const {
 117         return clang_getLocation(unit, file_, line_, column_);
 118     }
 119 };
 120
 121 template <void (&clang_get_Location)(CXSourceLocation, CXFile *, unsigned *, unsigned *, unsigned *)>
 122 std::ostream &operator <<(std::ostream &out, const CYCXPosition<clang_get_Location> &position) {
 123     if (position.file_ != NULL)
 124         out << "[" << CYCXString(position.file_) << "]:";
 125     out << position.line_ << ":" << position.column_ << "@" << position.offset_;
 126     return out;
 127 }
 128
 129 struct CYKey {
 130     unsigned priority_ = 0;
 131
 132     std::string code_;
 133     unsigned flags_;
 134 };
 135
 136 typedef std::map<std::string, CYKey> CYKeyMap;
 137
 138 struct CYChildBaton {
 139     CXTranslationUnit unit;
 140     CYKeyMap &keys;
 141
 142     CYChildBaton(CXTranslationUnit unit, CYKeyMap &keys) :
 143         unit(unit),
 144         keys(keys)
 145     {
 146     }
 147 };
 148
 149 struct CYTokens {
 150   private:
 151     CXTranslationUnit unit_;
 152     CXToken *tokens_;
 153     unsigned count_;
 154     unsigned valid_;
 155
 156   public:
 157     CYTokens(CXTranslationUnit unit, CXSourceRange range) :
 158         unit_(unit)
 159     {
 160         clang_tokenize(unit_, range, &tokens_, &count_);
 161
 162
 163         // libclang's tokenizer is horribly broken and returns "extra" tokens.
 164         // this code goes back through the tokens and filters for good ones :/
 165
 166         CYCXPosition<> end(clang_getRangeEnd(range));
 167         CYCXString file(end.file_);
 168
 169         for (valid_ = 0; valid_ != count_; ++valid_) {
 170             CYCXPosition<> position(unit, tokens_[valid_]);
 171             _assert(CYCXString(position.file_) == file);
 172             if (position.offset_ >= end.offset_)
 173                 break;
 174         }
 175     }
 176
 177     CYTokens(CXTranslationUnit unit, CXCursor cursor) :
 178         CYTokens(unit, clang_getCursorExtent(cursor))
 179     {
 180     }
 181
 182     ~CYTokens() {
 183         clang_disposeTokens(unit_, tokens_, count_);
 184     }
 185
 186     operator CXToken *() const {
 187         return tokens_;
 188     }
 189
 190     size_t size() const {
 191         return valid_;
 192     }
 193 };
 194
 195 static CYUTF8String CYCXPoolUTF8Range(CYPool &pool, CXSourceRange range) {
 196     CYCXPosition<> start(clang_getRangeStart(range));
 197     CYCXPosition<> end(clang_getRangeEnd(range));
 198     CYCXString file(start.file_);
 199     _assert(file == CYCXString(end.file_));
 200
 201     CYPool temp;
 202     size_t size;
 203     char *data(static_cast<char *>(CYPoolFile(temp, file, &size)));
 204     _assert(start.offset_ <= size && end.offset_ <= size && start.offset_ <= end.offset_);
 205
 206     CYUTF8String code;
 207     code.size = end.offset_ - start.offset_;
 208     code.data = pool.strndup(data + start.offset_, code.size);
 209     return code;
 210 }
 211
 212 static CYExpression *CYTranslateExpression(CXTranslationUnit unit, CXCursor cursor) {
 213     switch (CXCursorKind kind = clang_getCursorKind(cursor)) {
 214         case CXCursor_CallExpr: {
 215             CYExpression *function(NULL);
 216             CYList<CYArgument> arguments;
 217             CYForChild(cursor, fun([&](CXCursor child) {
 218                 CYExpression *expression(CYTranslateExpression(unit, child));
 219                 if (function == NULL)
 220                     function = expression;
 221                 else
 222                     arguments->*$C_(expression);
 223             }));
 224             return $C(function, arguments);
 225         } break;
 226
 227         case CXCursor_DeclRefExpr: {
 228             return $V(CYCXString(cursor).Pool($pool));
 229         } break;
 230
 231         case CXCursor_IntegerLiteral: {
 232             // libclang doesn't provide any reasonable way to do this
 233             // note: clang_tokenize doesn't work if this is a macro
 234             // the token range starts inside the macro but ends after it
 235             // the tokenizer freaks out and either fails with 0 tokens
 236             // or returns some massive number of tokens ending here :/
 237
 238             CYUTF8String token(CYCXPoolUTF8Range($pool, clang_getCursorExtent(cursor)));
 239             double value(CYCastDouble(token));
 240             if (std::isnan(value))
 241                 return $V(token.data);
 242             return $ CYNumber(value);
 243         } break;
 244
 245         case CXCursor_CStyleCastExpr:
 246             // XXX: most of the time, this is a "NoOp" integer cast; but we should check it
 247
 248         case CXCursor_UnexposedExpr:
 249             // there is a very high probability that this is actually an "ImplicitCastExpr"
 250             // "Douglas Gregor" <dgregor@apple.com> err'd on the incorrect side of this one
 251             // http://lists.llvm.org/pipermail/cfe-commits/Week-of-Mon-20110926/046998.html
 252
 253         case CXCursor_ParenExpr: {
 254             CYExpression *pass(NULL);
 255             CYOneChild(cursor, fun([&](CXCursor child) {
 256                 pass = CYTranslateExpression(unit, child);
 257             }));
 258             return pass;
 259         } break;
 260
 261         default:
 262             //std::cerr << "E:" << CYCXString(kind) << std::endl;
 263             _assert(false);
 264     }
 265 }
 266
 267 static CYStatement *CYTranslateStatement(CXTranslationUnit unit, CXCursor cursor) {
 268     switch (CXCursorKind kind = clang_getCursorKind(cursor)) {
 269         case CXCursor_ReturnStmt: {
 270             CYExpression *value(NULL);
 271             CYOneChild(cursor, fun([&](CXCursor child) {
 272                 value = CYTranslateExpression(unit, child);
 273             }));
 274             return $ CYReturn(value);
 275         } break;
 276
 277         default:
 278             //std::cerr << "S:" << CYCXString(kind) << std::endl;
 279             _assert(false);
 280     }
 281 }
 282
 283 static CYStatement *CYTranslateBlock(CXTranslationUnit unit, CXCursor cursor) {
 284     CYList<CYStatement> statements;
 285     CYForChild(cursor, fun([&](CXCursor child) {
 286         statements->*CYTranslateStatement(unit, child);
 287     }));
 288     return $ CYBlock(statements);
 289 }
 290
 291 static CXChildVisitResult CYChildVisit(CXCursor cursor, CXCursor parent, CXClientData arg) {
 292     CYChildBaton &baton(*static_cast<CYChildBaton *>(arg));
 293     CXTranslationUnit &unit(baton.unit);
 294
 295     CYCXString spelling(cursor);
 296     std::string name(spelling);
 297     std::ostringstream value;
 298     unsigned priority(2);
 299     unsigned flags(0);
 300
 301     /*CXSourceLocation location(clang_getCursorLocation(cursor));
 302     CYCXPosition<> position(location);
 303     std::cout << spelling << " " << position << std::endl;*/
 304
 305     switch (CXCursorKind kind = clang_getCursorKind(cursor)) {
 306         case CXCursor_EnumConstantDecl: {
 307             value << clang_getEnumConstantDeclValue(cursor);
 308         } break;
 309
 310         case CXCursor_MacroDefinition: try {
 311             CXSourceRange range(clang_getCursorExtent(cursor));
 312             CYTokens tokens(unit, range);
 313             _assert(tokens.size() != 0);
 314
 315             CXCursor cursors[tokens.size()];
 316             clang_annotateTokens(unit, tokens, tokens.size(), cursors);
 317
 318             CYLocalPool local;
 319             CYList<CYFunctionParameter> parameters;
 320             unsigned offset(1);
 321
 322             if (tokens.size() != 1) {
 323                 CYCXPosition<> start(clang_getRangeStart(range));
 324                 CYCXString first(unit, tokens[offset]);
 325                 if (first == "(") {
 326                     CYCXPosition<> paren(unit, tokens[offset]);
 327                     if (start.offset_ + strlen(spelling) == paren.offset_) {
 328                         for (;;) {
 329                             _assert(++offset != tokens.size());
 330                             CYCXString token(unit, tokens[offset]);
 331                             parameters->*$P($B($I(token.Pool($pool))));
 332                             _assert(++offset != tokens.size());
 333                             CYCXString comma(unit, tokens[offset]);
 334                             if (comma == ")")
 335                                 break;
 336                             _assert(comma == ",");
 337                         }
 338                         ++offset;
 339                     }
 340                 }
 341             }
 342
 343             std::ostringstream body;
 344             for (unsigned i(offset); i != tokens.size(); ++i) {
 345                 CYCXString token(unit, tokens[i]);
 346                 if (i != offset)
 347                     body << " ";
 348                 body << token;
 349             }
 350
 351             if (!parameters)
 352                 value << body.str();
 353             else {
 354                 CYOptions options;
 355                 CYOutput out(*value.rdbuf(), options);
 356                 out << '(' << "function" << '(';
 357                 out << parameters;
 358                 out << ')' << '{';
 359                 out << "return" << ' ';
 360                 value << body.str();
 361                 out << ';' << '}' << ')';
 362             }
 363         } catch (const CYException &error) {
 364             CYPool pool;
 365             //std::cerr << error.PoolCString(pool) << std::endl;
 366             goto skip;
 367         } break;
 368
 369         case CXCursor_StructDecl: {
 370             if (spelling[0] == '\0')
 371                 goto skip;
 372             if (!clang_isCursorDefinition(cursor))
 373                 priority = 1;
 374
 375             std::ostringstream types;
 376             std::ostringstream names;
 377
 378             CYForChild(cursor, fun([&](CXCursor child) {
 379                 if (clang_getCursorKind(child) == CXCursor_FieldDecl) {
 380                     CXType type(clang_getCursorType(child));
 381                     types << "(typedef " << CYCXString(clang_getTypeSpelling(type)) << "),";
 382                     names << "'" << CYCXString(child) << "',";
 383                 }
 384             }));
 385
 386             name += "$cy";
 387             value << "new Type([" << types.str() << "],[" << names.str() << "])";
 388         } break;
 389
 390         case CXCursor_TypedefDecl: {
 391             CXType type(clang_getTypedefDeclUnderlyingType(cursor));
 392             value << "(typedef " << CYCXString(clang_getTypeSpelling(type)) << ")";
 393         } break;
 394
 395         case CXCursor_FunctionDecl:
 396         case CXCursor_VarDecl: try {
 397             std::string label;
 398
 399             CYList<CYFunctionParameter> parameters;
 400             CYStatement *code(NULL);
 401
 402             CYLocalPool local;
 403
 404             CYForChild(cursor, fun([&](CXCursor child) {
 405                 switch (CXCursorKind kind = clang_getCursorKind(child)) {
 406                     case CXCursor_AsmLabelAttr:
 407                         label = CYCXString(child);
 408                         break;
 409
 410                     case CXCursor_CompoundStmt:
 411                         code = CYTranslateBlock(unit, child);
 412                         break;
 413
 414                     case CXCursor_ParmDecl:
 415                         parameters->*$P($B($I(CYCXString(child).Pool($pool))));
 416                         break;
 417
 418                     case CXCursor_IntegerLiteral:
 419                     case CXCursor_ObjCClassRef:
 420                     case CXCursor_TypeRef:
 421                     case CXCursor_UnexposedAttr:
 422                         break;
 423
 424                     default:
 425                         //std::cerr << "A:" << CYCXString(child) << std::endl;
 426                         break;
 427                 }
 428             }));
 429
 430             if (label.empty()) {
 431                 label = spelling;
 432                 label = '_' + label;
 433             } else if (label[0] != '_')
 434                 goto skip;
 435
 436             if (code == NULL) {
 437                 CXType type(clang_getCursorType(cursor));
 438                 value << "*(typedef " << CYCXString(clang_getTypeSpelling(type)) << ").pointerTo()(dlsym(RTLD_DEFAULT,'" << label.substr(1) << "'))";
 439             } else {
 440                 CYOptions options;
 441                 CYOutput out(*value.rdbuf(), options);
 442                 CYFunctionExpression *function($ CYFunctionExpression(NULL, parameters, code));
 443                 function->Output(out, CYNoBFC);
 444                 //std::cerr << value.str() << std::endl;
 445             }
 446         } catch (const CYException &error) {
 447             CYPool pool;
 448             //std::cerr << error.PoolCString(pool) << std::endl;
 449             goto skip;
 450         } break;
 451
 452         default: {
 453             return CXChildVisit_Recurse;
 454         } break;
 455     }
 456
 457     {
 458         CYKey &key(baton.keys[name]);
 459         if (key.priority_ < priority) {
 460             key.priority_ = priority;
 461             key.code_ = value.str();
 462             key.flags_ = flags;
 463         }
 464     }
 465
 466   skip:
 467     return CXChildVisit_Continue;
 468 }
 469
 470 int main(int argc, const char *argv[]) {
 471     CXIndex index(clang_createIndex(0, 0));
 472
 473     const char *file(argv[1]);
 474
 475     unsigned offset(3);
 476 #if CY_OBJECTIVEC
 477     argv[--offset] = "-ObjC++";
 478 #endif
 479
 480     CXTranslationUnit unit(clang_parseTranslationUnit(index, file, argv + offset, argc - offset, NULL, 0, CXTranslationUnit_DetailedPreprocessingRecord));
 481
 482     for (unsigned i(0), e(clang_getNumDiagnostics(unit)); i != e; ++i) {
 483         CXDiagnostic diagnostic(clang_getDiagnostic(unit, i));
 484         CYCXString spelling(clang_getDiagnosticSpelling(diagnostic));
 485         std::cerr << spelling << std::endl;
 486     }
 487
 488     CYKeyMap keys;
 489     CYChildBaton baton(unit, keys);
 490     clang_visitChildren(clang_getTranslationUnitCursor(unit), &CYChildVisit, &baton);
 491
 492     for (CYKeyMap::const_iterator key(keys.begin()); key != keys.end(); ++key) {
 493         std::string code(key->second.code_);
 494         for (size_t i(0), e(code.size()); i != e; ++i)
 495             if (code[i] <= 0 || code[i] >= 0x7f || code[i] == '\n')
 496                 goto skip;
 497         std::cout << key->first << "|" << key->second.flags_ << "\"" << code << "\"" << std::endl;
 498     skip:; }
 499
 500     clang_disposeTranslationUnit(unit);
 501     clang_disposeIndex(index);
 502
 503     return 0;
 504 }