X-Git-Url: https://git.saurik.com/cycript.git/blobdiff_plain/b7854baa232c47ae81470471e96d63fff09706e5..436a877be73ebe14fecd3ef0e9b7dd6b854d2e3b:/Analyze.cpp

diff --git a/Analyze.cpp b/Analyze.cpp
index c6e5777..4f84e03 100644
--- a/Analyze.cpp
+++ b/Analyze.cpp
@@ -1,3 +1,25 @@
+/* Cycript - The Truly Universal Scripting Language
+ * Copyright (C) 2009-2016  Jay Freeman (saurik)
+*/
+
+/* GNU Affero General Public License, Version 3 {{{ */
+/*
+ * This program is free software: you can redistribute it and/or modify
+ * it under the terms of the GNU Affero General Public License as published by
+ * the Free Software Foundation, either version 3 of the License, or
+ * (at your option) any later version.
+
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU Affero General Public License for more details.
+
+ * You should have received a copy of the GNU Affero General Public License
+ * along with this program.  If not, see <http://www.gnu.org/licenses/>.
+**/
+/* }}} */
+
+#include <cmath>
 #include <cstring>
 #include <iostream>
 #include <map>
@@ -6,6 +28,29 @@
 
 #include <clang-c/Index.h>
 
+#include "Functor.hpp"
+#include "Replace.hpp"
+#include "Syntax.hpp"
+
+static CXChildVisitResult CYVisit(CXCursor cursor, CXCursor parent, CXClientData arg) {
+    (*reinterpret_cast<const Functor<void (CXCursor)> *>(arg))(cursor);
+    return CXChildVisit_Continue;
+}
+
+static unsigned CYForChild(CXCursor cursor, const Functor<void (CXCursor)> &visitor) {
+    return clang_visitChildren(cursor, &CYVisit, const_cast<void *>(static_cast<const void *>(&visitor)));
+}
+
+static bool CYOneChild(CXCursor cursor, const Functor<void (CXCursor)> &visitor) {
+    bool visited(false);
+    CYForChild(cursor, fun([&](CXCursor child) {
+        _assert(!visited);
+        visited = true;
+        visitor(child);
+    }));
+    return visited;
+}
+
 struct CYCXString {
     CXString value_;
 
@@ -14,6 +59,26 @@ struct CYCXString {
     {
     }
 
+    CYCXString(CXCursor cursor) :
+        value_(clang_getCursorSpelling(cursor))
+    {
+    }
+
+    CYCXString(CXCursorKind kind) :
+        value_(clang_getCursorKindSpelling(kind))
+    {
+    }
+
+    CYCXString(CXFile file) :
+        value_(clang_getFileName(file))
+    {
+    }
+
+    CYCXString(CXTranslationUnit unit, CXToken token) :
+        value_(clang_getTokenSpelling(unit, token))
+    {
+    }
+
     ~CYCXString() {
         clang_disposeString(value_);
     }
@@ -21,39 +86,54 @@ struct CYCXString {
     operator const char *() const {
         return clang_getCString(value_);
     }
-};
 
-struct CYFieldBaton {
-    std::ostringstream types;
-    std::ostringstream names;
+    const char *Pool(CYPool &pool) const {
+        return pool.strdup(*this);
+    }
+
+    bool operator ==(const char *rhs) const {
+        const char *lhs(*this);
+        return lhs == rhs || strcmp(lhs, rhs) == 0;
+    }
 };
 
-static CXChildVisitResult CYFieldVisit(CXCursor cursor, CXCursor parent, CXClientData arg) {
-    CYFieldBaton &baton(*static_cast<CYFieldBaton *>(arg));
+template <void (&clang_get_Location)(CXSourceLocation, CXFile *, unsigned *, unsigned *, unsigned *) = clang_getSpellingLocation>
+struct CYCXPosition {
+    CXFile file_;
+    unsigned line_;
+    unsigned column_;
+    unsigned offset_;
 
-    if (clang_getCursorKind(cursor) == CXCursor_FieldDecl) {
-        CXType type(clang_getCursorType(cursor));
-        baton.types << "(typedef " << CYCXString(clang_getTypeSpelling(type)) << "),";
-        baton.names << "'" << CYCXString(clang_getCursorSpelling(cursor)) << "',";
+    CYCXPosition(CXSourceLocation location) {
+        clang_get_Location(location, &file_, &line_, &column_, &offset_);
     }
 
-    return CXChildVisit_Continue;
-}
+    CYCXPosition(CXTranslationUnit unit, CXToken token) :
+        CYCXPosition(clang_getTokenLocation(unit, token))
+    {
+    }
 
-struct CYAttributeBaton {
-    std::string label;
+    CXSourceLocation Get(CXTranslationUnit unit) const {
+        return clang_getLocation(unit, file_, line_, column_);
+    }
 };
 
-static CXChildVisitResult CYAttributeVisit(CXCursor cursor, CXCursor parent, CXClientData arg) {
-    CYAttributeBaton &baton(*static_cast<CYAttributeBaton *>(arg));
+template <void (&clang_get_Location)(CXSourceLocation, CXFile *, unsigned *, unsigned *, unsigned *)>
+std::ostream &operator <<(std::ostream &out, const CYCXPosition<clang_get_Location> &position) {
+    if (position.file_ != NULL)
+        out << "[" << CYCXString(position.file_) << "]:";
+    out << position.line_ << ":" << position.column_ << "@" << position.offset_;
+    return out;
+}
 
-    if (clang_getCursorKind(cursor) == CXCursor_AsmLabelAttr)
-        baton.label = CYCXString(clang_getCursorSpelling(cursor));
+struct CYKey {
+    unsigned priority_ = 0;
 
-    return CXChildVisit_Continue;
-}
+    std::string code_;
+    unsigned flags_;
+};
 
-typedef std::map<std::string, std::string> CYKeyMap;
+typedef std::map<std::string, CYKey> CYKeyMap;
 
 struct CYChildBaton {
     CXTranslationUnit unit;
@@ -67,86 +147,244 @@ struct CYChildBaton {
 };
 
 struct CYTokens {
-    CXTranslationUnit unit;
-    CXToken *tokens;
-    unsigned count;
+  private:
+    CXTranslationUnit unit_;
+    CXToken *tokens_;
+    unsigned count_;
+    unsigned valid_;
+
+  public:
+    CYTokens(CXTranslationUnit unit, CXSourceRange range) :
+        unit_(unit)
+    {
+        clang_tokenize(unit_, range, &tokens_, &count_);
+
+
+        // libclang's tokenizer is horribly broken and returns "extra" tokens.
+        // this code goes back through the tokens and filters for good ones :/
+
+        CYCXPosition<> end(clang_getRangeEnd(range));
+        CYCXString file(end.file_);
+
+        for (valid_ = 0; valid_ != count_; ++valid_) {
+            CYCXPosition<> position(unit, tokens_[valid_]);
+            _assert(CYCXString(position.file_) == file);
+            if (position.offset_ >= end.offset_)
+                break;
+        }
+    }
 
     CYTokens(CXTranslationUnit unit, CXCursor cursor) :
-        unit(unit)
+        CYTokens(unit, clang_getCursorExtent(cursor))
     {
-        CXSourceRange range(clang_getCursorExtent(cursor));
-        clang_tokenize(unit, range, &tokens, &count);
     }
 
     ~CYTokens() {
-        clang_disposeTokens(unit, tokens, count);
+        clang_disposeTokens(unit_, tokens_, count_);
     }
 
     operator CXToken *() const {
-        return tokens;
+        return tokens_;
+    }
+
+    size_t size() const {
+        return valid_;
     }
 };
 
+static CYUTF8String CYCXPoolUTF8Range(CYPool &pool, CXSourceRange range) {
+    CYCXPosition<> start(clang_getRangeStart(range));
+    CYCXPosition<> end(clang_getRangeEnd(range));
+    CYCXString file(start.file_);
+    _assert(file == CYCXString(end.file_));
+
+    CYPool temp;
+    size_t size;
+    char *data(static_cast<char *>(CYPoolFile(temp, file, &size)));
+    _assert(start.offset_ <= size && end.offset_ <= size && start.offset_ <= end.offset_);
+
+    CYUTF8String code;
+    code.size = end.offset_ - start.offset_;
+    code.data = pool.strndup(data + start.offset_, code.size);
+    return code;
+}
+
+static CYExpression *CYTranslateExpression(CXTranslationUnit unit, CXCursor cursor) {
+    switch (CXCursorKind kind = clang_getCursorKind(cursor)) {
+        case CXCursor_CallExpr: {
+            CYExpression *function(NULL);
+            CYList<CYArgument> arguments;
+            CYForChild(cursor, fun([&](CXCursor child) {
+                CYExpression *expression(CYTranslateExpression(unit, child));
+                if (function == NULL)
+                    function = expression;
+                else
+                    arguments->*$C_(expression);
+            }));
+            return $C(function, arguments);
+        } break;
+
+        case CXCursor_DeclRefExpr: {
+            return $V(CYCXString(cursor).Pool($pool));
+        } break;
+
+        case CXCursor_IntegerLiteral: {
+            // libclang doesn't provide any reasonable way to do this
+            // note: clang_tokenize doesn't work if this is a macro
+            // the token range starts inside the macro but ends after it
+            // the tokenizer freaks out and either fails with 0 tokens
+            // or returns some massive number of tokens ending here :/
+
+            CYUTF8String token(CYCXPoolUTF8Range($pool, clang_getCursorExtent(cursor)));
+            double value(CYCastDouble(token));
+            if (std::isnan(value))
+                return $V(token.data);
+            return $ CYNumber(value);
+        } break;
+
+        case CXCursor_CStyleCastExpr:
+            // XXX: most of the time, this is a "NoOp" integer cast; but we should check it
+
+        case CXCursor_UnexposedExpr:
+            // there is a very high probability that this is actually an "ImplicitCastExpr"
+            // "Douglas Gregor" <dgregor@apple.com> err'd on the incorrect side of this one
+            // http://lists.llvm.org/pipermail/cfe-commits/Week-of-Mon-20110926/046998.html
+
+        case CXCursor_ParenExpr: {
+            CYExpression *pass(NULL);
+            CYOneChild(cursor, fun([&](CXCursor child) {
+                pass = CYTranslateExpression(unit, child);
+            }));
+            return pass;
+        } break;
+
+        default:
+            //std::cerr << "E:" << CYCXString(kind) << std::endl;
+            _assert(false);
+    }
+}
+
+static CYStatement *CYTranslateStatement(CXTranslationUnit unit, CXCursor cursor) {
+    switch (CXCursorKind kind = clang_getCursorKind(cursor)) {
+        case CXCursor_ReturnStmt: {
+            CYExpression *value(NULL);
+            CYOneChild(cursor, fun([&](CXCursor child) {
+                value = CYTranslateExpression(unit, child);
+            }));
+            return $ CYReturn(value);
+        } break;
+
+        default:
+            //std::cerr << "S:" << CYCXString(kind) << std::endl;
+            _assert(false);
+    }
+}
+
+static CYStatement *CYTranslateBlock(CXTranslationUnit unit, CXCursor cursor) {
+    CYList<CYStatement> statements;
+    CYForChild(cursor, fun([&](CXCursor child) {
+        statements->*CYTranslateStatement(unit, child);
+    }));
+    return $ CYBlock(statements);
+}
+
 static CXChildVisitResult CYChildVisit(CXCursor cursor, CXCursor parent, CXClientData arg) {
     CYChildBaton &baton(*static_cast<CYChildBaton *>(arg));
     CXTranslationUnit &unit(baton.unit);
 
-    CYCXString spelling(clang_getCursorSpelling(cursor));
+    CYCXString spelling(cursor);
     std::string name(spelling);
     std::ostringstream value;
+    unsigned priority(2);
+    unsigned flags(0);
 
     /*CXSourceLocation location(clang_getCursorLocation(cursor));
+    CYCXPosition<> position(location);
+    std::cout << spelling << " " << position << std::endl;*/
 
-    CXFile file;
-    unsigned line;
-    unsigned column;
-    unsigned offset;
-    clang_getSpellingLocation(location, &file, &line, &column, &offset);
-
-    if (file != NULL) {
-        CYCXString path(clang_getFileName(file));
-        std::cout << spelling << " " << path << ":" << line << std::endl;
-    }*/
-
-    switch (clang_getCursorKind(cursor)) {
+    switch (CXCursorKind kind = clang_getCursorKind(cursor)) {
         case CXCursor_EnumConstantDecl: {
             value << clang_getEnumConstantDeclValue(cursor);
         } break;
 
-        case CXCursor_MacroDefinition: {
-            CYTokens tokens(unit, cursor);
-            if (tokens.count <= 2)
-                goto skip;
+        case CXCursor_MacroDefinition: try {
+            CXSourceRange range(clang_getCursorExtent(cursor));
+            CYTokens tokens(unit, range);
+            _assert(tokens.size() != 0);
+
+            CXCursor cursors[tokens.size()];
+            clang_annotateTokens(unit, tokens, tokens.size(), cursors);
+
+            CYLocalPool local;
+            CYList<CYFunctionParameter> parameters;
+            unsigned offset(1);
+
+            if (tokens.size() != 1) {
+                CYCXPosition<> start(clang_getRangeStart(range));
+                CYCXString first(unit, tokens[offset]);
+                if (first == "(") {
+                    CYCXPosition<> paren(unit, tokens[offset]);
+                    if (start.offset_ + strlen(spelling) == paren.offset_) {
+                        for (;;) {
+                            _assert(++offset != tokens.size());
+                            CYCXString token(unit, tokens[offset]);
+                            parameters->*$P($B($I(token.Pool($pool))));
+                            _assert(++offset != tokens.size());
+                            CYCXString comma(unit, tokens[offset]);
+                            if (comma == ")")
+                                break;
+                            _assert(comma == ",");
+                        }
+                        ++offset;
+                    }
+                }
+            }
 
-            CXCursor cursors[tokens.count];
-            clang_annotateTokens(unit, tokens, tokens.count, cursors);
+            std::ostringstream body;
+            for (unsigned i(offset); i != tokens.size(); ++i) {
+                CYCXString token(unit, tokens[i]);
+                if (i != offset)
+                    body << " ";
+                body << token;
+            }
 
-            for (unsigned i(1); i != tokens.count - 1; ++i) {
-                CYCXString token(clang_getTokenSpelling(unit, tokens[i]));
-                if (i != 1)
-                    value << " ";
-                else if (strcmp(token, "(") == 0)
-                    goto skip;
-                value << token;
+            if (!parameters)
+                value << body.str();
+            else {
+                CYOptions options;
+                CYOutput out(*value.rdbuf(), options);
+                out << '(' << "function" << '(';
+                out << parameters;
+                out << ')' << '{';
+                out << "return" << ' ';
+                value << body.str();
+                out << ';' << '}' << ')';
             }
+        } catch (const CYException &error) {
+            CYPool pool;
+            //std::cerr << error.PoolCString(pool) << std::endl;
+            goto skip;
         } break;
 
         case CXCursor_StructDecl: {
-            if (!clang_isCursorDefinition(cursor))
-                goto skip;
             if (spelling[0] == '\0')
                 goto skip;
+            if (!clang_isCursorDefinition(cursor))
+                priority = 1;
 
-            CYFieldBaton baton;
+            std::ostringstream types;
+            std::ostringstream names;
 
-            baton.types << "[";
-            baton.names << "[";
-            clang_visitChildren(cursor, &CYFieldVisit, &baton);
-            baton.types << "]";
-            baton.names << "]";
+            CYForChild(cursor, fun([&](CXCursor child) {
+                if (clang_getCursorKind(child) == CXCursor_FieldDecl) {
+                    CXType type(clang_getCursorType(child));
+                    types << "(typedef " << CYCXString(clang_getTypeSpelling(type)) << "),";
+                    names << "'" << CYCXString(child) << "',";
+                }
+            }));
 
+            value << "new Type([" << types.str() << "],[" << names.str() << "]).withName(\"" << name << "\")";
             name += "$cy";
-            value << "new Type(" << baton.types.str() << "," << baton.names.str() << ")";
         } break;
 
         case CXCursor_TypedefDecl: {
@@ -155,18 +393,60 @@ static CXChildVisitResult CYChildVisit(CXCursor cursor, CXCursor parent, CXClien
         } break;
 
         case CXCursor_FunctionDecl:
-        case CXCursor_VarDecl: {
-            CYAttributeBaton baton;
-            clang_visitChildren(cursor, &CYAttributeVisit, &baton);
-
-            if (baton.label.empty()) {
-                baton.label = spelling;
-                baton.label = '_' + baton.label;
-            } else if (baton.label[0] != '_')
+        case CXCursor_VarDecl: try {
+            std::string label;
+
+            CYList<CYFunctionParameter> parameters;
+            CYStatement *code(NULL);
+
+            CYLocalPool local;
+
+            CYForChild(cursor, fun([&](CXCursor child) {
+                switch (CXCursorKind kind = clang_getCursorKind(child)) {
+                    case CXCursor_AsmLabelAttr:
+                        label = CYCXString(child);
+                        break;
+
+                    case CXCursor_CompoundStmt:
+                        code = CYTranslateBlock(unit, child);
+                        break;
+
+                    case CXCursor_ParmDecl:
+                        parameters->*$P($B($I(CYCXString(child).Pool($pool))));
+                        break;
+
+                    case CXCursor_IntegerLiteral:
+                    case CXCursor_ObjCClassRef:
+                    case CXCursor_TypeRef:
+                    case CXCursor_UnexposedAttr:
+                        break;
+
+                    default:
+                        //std::cerr << "A:" << CYCXString(child) << std::endl;
+                        break;
+                }
+            }));
+
+            if (label.empty()) {
+                label = spelling;
+                label = '_' + label;
+            } else if (label[0] != '_')
                 goto skip;
 
-            CXType type(clang_getCursorType(cursor));
-            value << "*(typedef " << CYCXString(clang_getTypeSpelling(type)) << ").pointerTo()(dlsym(RTLD_DEFAULT,'" << baton.label.substr(1) << "'))";
+            if (code == NULL) {
+                CXType type(clang_getCursorType(cursor));
+                value << "*(typedef " << CYCXString(clang_getTypeSpelling(type)) << ").pointerTo()(dlsym(RTLD_DEFAULT,'" << label.substr(1) << "'))";
+            } else {
+                CYOptions options;
+                CYOutput out(*value.rdbuf(), options);
+                CYFunctionExpression *function($ CYFunctionExpression(NULL, parameters, code));
+                function->Output(out, CYNoBFC);
+                //std::cerr << value.str() << std::endl;
+            }
+        } catch (const CYException &error) {
+            CYPool pool;
+            //std::cerr << error.PoolCString(pool) << std::endl;
+            goto skip;
         } break;
 
         default: {
@@ -174,7 +454,14 @@ static CXChildVisitResult CYChildVisit(CXCursor cursor, CXCursor parent, CXClien
         } break;
     }
 
-    baton.keys[name] = value.str();
+    {
+        CYKey &key(baton.keys[name]);
+        if (key.priority_ < priority) {
+            key.priority_ = priority;
+            key.code_ = value.str();
+            key.flags_ = flags;
+        }
+    }
 
   skip:
     return CXChildVisit_Continue;
@@ -190,7 +477,7 @@ int main(int argc, const char *argv[]) {
     argv[--offset] = "-ObjC++";
 #endif
 
-    CXTranslationUnit unit(clang_parseTranslationUnit(index, file, argv + offset, argc - offset, NULL, 0, CXTranslationUnit_DetailedPreprocessingRecord | CXTranslationUnit_SkipFunctionBodies));
+    CXTranslationUnit unit(clang_parseTranslationUnit(index, file, argv + offset, argc - offset, NULL, 0, CXTranslationUnit_DetailedPreprocessingRecord));
 
     for (unsigned i(0), e(clang_getNumDiagnostics(unit)); i != e; ++i) {
         CXDiagnostic diagnostic(clang_getDiagnostic(unit, i));
@@ -203,11 +490,11 @@ int main(int argc, const char *argv[]) {
     clang_visitChildren(clang_getTranslationUnitCursor(unit), &CYChildVisit, &baton);
 
     for (CYKeyMap::const_iterator key(keys.begin()); key != keys.end(); ++key) {
-        std::string value(key->second);
-        for (size_t i(0), e(value.size()); i != e; ++i)
-            if (value[i] <= 0 || value[i] >= 0x7f || value[i] == '\n')
+        std::string code(key->second.code_);
+        for (size_t i(0), e(code.size()); i != e; ++i)
+            if (code[i] <= 0 || code[i] >= 0x7f || code[i] == '\n')
                 goto skip;
-        std::cout << key->first << "|\"" << value << "\"" << std::endl;
+        std::cout << key->first << "|" << key->second.flags_ << "\"" << code << "\"" << std::endl;
     skip:; }
 
     clang_disposeTranslationUnit(unit);