Analyze.cpp

   1 /* Cycript - Optimizing JavaScript Compiler/Runtime
   2  * Copyright (C) 2009-2015  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 typedef std::map<std::string, std::string> CYKeyMap;
 130
 131 struct CYChildBaton {
 132     CXTranslationUnit unit;
 133     CYKeyMap &keys;
 134
 135     CYChildBaton(CXTranslationUnit unit, CYKeyMap &keys) :
 136         unit(unit),
 137         keys(keys)
 138     {
 139     }
 140 };
 141
 142 struct CYTokens {
 143   private:
 144     CXTranslationUnit unit_;
 145     CXToken *tokens_;
 146     unsigned count_;
 147     unsigned valid_;
 148
 149   public:
 150     CYTokens(CXTranslationUnit unit, CXSourceRange range) :
 151         unit_(unit)
 152     {
 153         clang_tokenize(unit_, range, &tokens_, &count_);
 154
 155
 156         // libclang's tokenizer is horribly broken and returns "extra" tokens.
 157         // this code goes back through the tokens and filters for good ones :/
 158
 159         CYCXPosition<> end(clang_getRangeEnd(range));
 160         CYCXString file(end.file_);
 161
 162         for (valid_ = 0; valid_ != count_; ++valid_) {
 163             CYCXPosition<> position(unit, tokens_[valid_]);
 164             _assert(CYCXString(position.file_) == file);
 165             if (position.offset_ >= end.offset_)
 166                 break;
 167         }
 168     }
 169
 170     CYTokens(CXTranslationUnit unit, CXCursor cursor) :
 171         CYTokens(unit, clang_getCursorExtent(cursor))
 172     {
 173     }
 174
 175     ~CYTokens() {
 176         clang_disposeTokens(unit_, tokens_, count_);
 177     }
 178
 179     operator CXToken *() const {
 180         return tokens_;
 181     }
 182
 183     size_t size() const {
 184         return valid_;
 185     }
 186 };
 187
 188 static CYExpression *CYTranslateExpression(CXTranslationUnit unit, CXCursor cursor) {
 189     switch (CXCursorKind kind = clang_getCursorKind(cursor)) {
 190         case CXCursor_CallExpr: {
 191             CYExpression *function(NULL);
 192             CYList<CYArgument> arguments;
 193             CYForChild(cursor, fun([&](CXCursor child) {
 194                 CYExpression *expression(CYTranslateExpression(unit, child));
 195                 if (function == NULL)
 196                     function = expression;
 197                 else
 198                     arguments->*$C_(expression);
 199             }));
 200             return $C(function, arguments);
 201         } break;
 202
 203         case CXCursor_DeclRefExpr: {
 204             return $V(CYCXString(cursor).Pool($pool));
 205         } break;
 206
 207         case CXCursor_IntegerLiteral: {
 208             // libclang doesn't provide any reasonable way to do this
 209             // note: clang_tokenize doesn't work if this is a macro
 210             // the token range starts inside the macro but ends after it
 211             // the tokenizer freaks out and either fails with 0 tokens
 212             // or returns some massive number of tokens ending here :/
 213
 214             CXSourceRange range(clang_getCursorExtent(cursor));
 215             CYCXPosition<> start(clang_getRangeStart(range));
 216             CYCXPosition<> end(clang_getRangeEnd(range));
 217             CYCXString file(start.file_);
 218             _assert(file == CYCXString(end.file_));
 219
 220             CYPool pool;
 221             size_t size;
 222             char *data(static_cast<char *>(CYPoolFile(pool, file, &size)));
 223             _assert(start.offset_ <= size && end.offset_ <= size && start.offset_ <= end.offset_);
 224
 225             const char *token($pool.strndup(data + start.offset_, end.offset_ - start.offset_));
 226             double value(CYCastDouble(token));
 227             if (!std::isnan(value))
 228                 return $ CYNumber(value);
 229
 230             return $V(token);
 231         } break;
 232
 233         case CXCursor_CStyleCastExpr:
 234             // XXX: most of the time, this is a "NoOp" integer cast; but we should check it
 235
 236         case CXCursor_UnexposedExpr:
 237             // there is a very high probability that this is actually an "ImplicitCastExpr"
 238             // "Douglas Gregor" <dgregor@apple.com> err'd on the incorrect side of this one
 239             // http://lists.llvm.org/pipermail/cfe-commits/Week-of-Mon-20110926/046998.html
 240
 241         case CXCursor_ParenExpr: {
 242             CYExpression *pass(NULL);
 243             CYOneChild(cursor, fun([&](CXCursor child) {
 244                 pass = CYTranslateExpression(unit, child);
 245             }));
 246             return pass;
 247         } break;
 248
 249         default:
 250             //std::cerr << "E:" << CYCXString(kind) << std::endl;
 251             _assert(false);
 252     }
 253 }
 254
 255 static CYStatement *CYTranslateStatement(CXTranslationUnit unit, CXCursor cursor) {
 256     switch (CXCursorKind kind = clang_getCursorKind(cursor)) {
 257         case CXCursor_ReturnStmt: {
 258             CYExpression *value(NULL);
 259             CYOneChild(cursor, fun([&](CXCursor child) {
 260                 value = CYTranslateExpression(unit, child);
 261             }));
 262             return $ CYReturn(value);
 263         } break;
 264
 265         default:
 266             //std::cerr << "S:" << CYCXString(kind) << std::endl;
 267             _assert(false);
 268     }
 269 }
 270
 271 static CYStatement *CYTranslateBlock(CXTranslationUnit unit, CXCursor cursor) {
 272     CYList<CYStatement> statements;
 273     CYForChild(cursor, fun([&](CXCursor child) {
 274         statements->*CYTranslateStatement(unit, child);
 275     }));
 276     return $ CYBlock(statements);
 277 }
 278
 279 static CXChildVisitResult CYChildVisit(CXCursor cursor, CXCursor parent, CXClientData arg) {
 280     CYChildBaton &baton(*static_cast<CYChildBaton *>(arg));
 281     CXTranslationUnit &unit(baton.unit);
 282
 283     CYCXString spelling(cursor);
 284     std::string name(spelling);
 285     std::ostringstream value;
 286
 287     /*CXSourceLocation location(clang_getCursorLocation(cursor));
 288     CYCXPosition<> position(location);
 289     std::cout << spelling << " " << position << std::endl;*/
 290
 291     switch (CXCursorKind kind = clang_getCursorKind(cursor)) {
 292         case CXCursor_EnumConstantDecl: {
 293             value << clang_getEnumConstantDeclValue(cursor);
 294         } break;
 295
 296         case CXCursor_MacroDefinition: try {
 297             CXSourceRange range(clang_getCursorExtent(cursor));
 298             CYTokens tokens(unit, range);
 299             _assert(tokens.size() != 0);
 300
 301             CXCursor cursors[tokens.size()];
 302             clang_annotateTokens(unit, tokens, tokens.size(), cursors);
 303
 304             CYCXPosition<> start(clang_getRangeStart(range));
 305             CYCXString first(unit, tokens[1]);
 306             if (first == "(") {
 307                 CYCXPosition<> paren(unit, tokens[1]);
 308                 if (start.offset_ + strlen(spelling) == paren.offset_)
 309                     _assert(false); // XXX: support parameterized macros
 310             }
 311
 312             for (unsigned i(1); i != tokens.size(); ++i) {
 313                 CYCXString token(unit, tokens[i]);
 314                 if (i != 1)
 315                     value << " ";
 316                 value << token;
 317             }
 318         } catch (const CYException &error) {
 319             CYPool pool;
 320             //std::cerr << error.PoolCString(pool) << std::endl;
 321             goto skip;
 322         } break;
 323
 324         case CXCursor_StructDecl: {
 325             if (!clang_isCursorDefinition(cursor))
 326                 goto skip;
 327             if (spelling[0] == '\0')
 328                 goto skip;
 329
 330             std::ostringstream types;
 331             std::ostringstream names;
 332
 333             CYForChild(cursor, fun([&](CXCursor child) {
 334                 if (clang_getCursorKind(child) == CXCursor_FieldDecl) {
 335                     CXType type(clang_getCursorType(child));
 336                     types << "(typedef " << CYCXString(clang_getTypeSpelling(type)) << "),";
 337                     names << "'" << CYCXString(child) << "',";
 338                 }
 339             }));
 340
 341             name += "$cy";
 342             value << "new Type([" << types.str() << "],[" << names.str() << "])";
 343         } break;
 344
 345         case CXCursor_TypedefDecl: {
 346             CXType type(clang_getTypedefDeclUnderlyingType(cursor));
 347             value << "(typedef " << CYCXString(clang_getTypeSpelling(type)) << ")";
 348         } break;
 349
 350         case CXCursor_FunctionDecl:
 351         case CXCursor_VarDecl: try {
 352             std::string label;
 353
 354             CYList<CYFunctionParameter> parameters;
 355             CYStatement *code(NULL);
 356
 357             CYLocalPool local;
 358
 359             CYForChild(cursor, fun([&](CXCursor child) {
 360                 switch (CXCursorKind kind = clang_getCursorKind(child)) {
 361                     case CXCursor_AsmLabelAttr:
 362                         label = CYCXString(child);
 363                         break;
 364
 365                     case CXCursor_CompoundStmt:
 366                         code = CYTranslateBlock(unit, child);
 367                         break;
 368
 369                     case CXCursor_ParmDecl:
 370                         parameters->*$P($B($I(CYCXString(child).Pool($pool))));
 371                         break;
 372
 373                     case CXCursor_IntegerLiteral:
 374                     case CXCursor_ObjCClassRef:
 375                     case CXCursor_TypeRef:
 376                     case CXCursor_UnexposedAttr:
 377                         break;
 378
 379                     default:
 380                         //std::cerr << "A:" << CYCXString(child) << std::endl;
 381                         break;
 382                 }
 383             }));
 384
 385             if (label.empty()) {
 386                 label = spelling;
 387                 label = '_' + label;
 388             } else if (label[0] != '_')
 389                 goto skip;
 390
 391             if (code == NULL) {
 392                 CXType type(clang_getCursorType(cursor));
 393                 value << "*(typedef " << CYCXString(clang_getTypeSpelling(type)) << ").pointerTo()(dlsym(RTLD_DEFAULT,'" << label.substr(1) << "'))";
 394             } else {
 395                 CYOptions options;
 396                 CYOutput out(*value.rdbuf(), options);
 397                 CYFunctionExpression *function($ CYFunctionExpression(NULL, parameters, code));
 398                 function->Output(out, CYNoBFC);
 399                 //std::cerr << value.str() << std::endl;
 400             }
 401         } catch (const CYException &error) {
 402             CYPool pool;
 403             //std::cerr << error.PoolCString(pool) << std::endl;
 404             goto skip;
 405         } break;
 406
 407         default: {
 408             return CXChildVisit_Recurse;
 409         } break;
 410     }
 411
 412     baton.keys[name] = value.str();
 413
 414   skip:
 415     return CXChildVisit_Continue;
 416 }
 417
 418 int main(int argc, const char *argv[]) {
 419     CXIndex index(clang_createIndex(0, 0));
 420
 421     const char *file(argv[1]);
 422
 423     unsigned offset(3);
 424 #if CY_OBJECTIVEC
 425     argv[--offset] = "-ObjC++";
 426 #endif
 427
 428     CXTranslationUnit unit(clang_parseTranslationUnit(index, file, argv + offset, argc - offset, NULL, 0, CXTranslationUnit_DetailedPreprocessingRecord));
 429
 430     for (unsigned i(0), e(clang_getNumDiagnostics(unit)); i != e; ++i) {
 431         CXDiagnostic diagnostic(clang_getDiagnostic(unit, i));
 432         CYCXString spelling(clang_getDiagnosticSpelling(diagnostic));
 433         std::cerr << spelling << std::endl;
 434     }
 435
 436     CYKeyMap keys;
 437     CYChildBaton baton(unit, keys);
 438     clang_visitChildren(clang_getTranslationUnitCursor(unit), &CYChildVisit, &baton);
 439
 440     for (CYKeyMap::const_iterator key(keys.begin()); key != keys.end(); ++key) {
 441         std::string value(key->second);
 442         for (size_t i(0), e(value.size()); i != e; ++i)
 443             if (value[i] <= 0 || value[i] >= 0x7f || value[i] == '\n')
 444                 goto skip;
 445         std::cout << key->first << "|\"" << value << "\"" << std::endl;
 446     skip:; }
 447
 448     clang_disposeTranslationUnit(unit);
 449     clang_disposeIndex(index);
 450
 451     return 0;
 452 }