1 /* Cycript - The Truly Universal Scripting Language
2 * Copyright (C) 2009-2016 Jay Freeman (saurik)
5 /* GNU Affero General Public License, Version 3 {{{ */
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.
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.
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/>.
29 #include <clang-c/Index.h>
32 #include "Functor.hpp"
33 #include "Replace.hpp"
36 static CXChildVisitResult
CYVisit(CXCursor cursor
, CXCursor parent
, CXClientData arg
) {
37 (*reinterpret_cast<const Functor
<void (CXCursor
)> *>(arg
))(cursor
);
38 return CXChildVisit_Continue
;
41 static unsigned CYForChild(CXCursor cursor
, const Functor
<void (CXCursor
)> &visitor
) {
42 return clang_visitChildren(cursor
, &CYVisit
, const_cast<void *>(static_cast<const void *>(&visitor
)));
45 static bool CYOneChild(CXCursor cursor
, const Functor
<void (CXCursor
)> &visitor
) {
47 CYForChild(cursor
, fun([&](CXCursor child
) {
58 CYCXString(CXString value
) :
63 CYCXString(CXCursor cursor
) :
64 value_(clang_getCursorSpelling(cursor
))
68 CYCXString(CXCursorKind kind
) :
69 value_(clang_getCursorKindSpelling(kind
))
73 CYCXString(CXFile file
) :
74 value_(clang_getFileName(file
))
78 CYCXString(CXTranslationUnit unit
, CXToken token
) :
79 value_(clang_getTokenSpelling(unit
, token
))
84 clang_disposeString(value_
);
87 operator const char *() const {
88 return clang_getCString(value_
);
91 const char *Pool(CYPool
&pool
) const {
92 return pool
.strdup(*this);
95 bool operator ==(const char *rhs
) const {
96 const char *lhs(*this);
97 return lhs
== rhs
|| strcmp(lhs
, rhs
) == 0;
101 template <void (&clang_get_Location
)(CXSourceLocation
, CXFile
*, unsigned *, unsigned *, unsigned *) = clang_getSpellingLocation
>
102 struct CYCXPosition
{
108 CYCXPosition(CXSourceLocation location
) {
109 clang_get_Location(location
, &file_
, &line_
, &column_
, &offset_
);
112 CYCXPosition(CXTranslationUnit unit
, CXToken token
) :
113 CYCXPosition(clang_getTokenLocation(unit
, token
))
117 CXSourceLocation
Get(CXTranslationUnit unit
) const {
118 return clang_getLocation(unit
, file_
, line_
, column_
);
122 template <void (&clang_get_Location
)(CXSourceLocation
, CXFile
*, unsigned *, unsigned *, unsigned *)>
123 std::ostream
&operator <<(std::ostream
&out
, const CYCXPosition
<clang_get_Location
> &position
) {
124 if (position
.file_
!= NULL
)
125 out
<< "[" << CYCXString(position
.file_
) << "]:";
126 out
<< position
.line_
<< ":" << position
.column_
<< "@" << position
.offset_
;
131 unsigned priority_
= 0;
137 typedef std::map
<std::string
, CYKey
> CYKeyMap
;
139 struct CYChildBaton
{
140 CXTranslationUnit unit
;
143 CYChildBaton(CXTranslationUnit unit
, CYKeyMap
&keys
) :
152 CXTranslationUnit unit_
;
158 CYTokens(CXTranslationUnit unit
, CXSourceRange range
) :
161 clang_tokenize(unit_
, range
, &tokens_
, &count_
);
164 // libclang's tokenizer is horribly broken and returns "extra" tokens.
165 // this code goes back through the tokens and filters for good ones :/
167 CYCXPosition
<> end(clang_getRangeEnd(range
));
168 CYCXString
file(end
.file_
);
170 for (valid_
= 0; valid_
!= count_
; ++valid_
) {
171 CYCXPosition
<> position(unit
, tokens_
[valid_
]);
172 _assert(CYCXString(position
.file_
) == file
);
173 if (position
.offset_
>= end
.offset_
)
178 CYTokens(CXTranslationUnit unit
, CXCursor cursor
) :
179 CYTokens(unit
, clang_getCursorExtent(cursor
))
184 clang_disposeTokens(unit_
, tokens_
, count_
);
187 operator CXToken
*() const {
191 size_t size() const {
196 static CYUTF8String
CYCXPoolUTF8Range(CYPool
&pool
, CXSourceRange range
) {
197 CYCXPosition
<> start(clang_getRangeStart(range
));
198 CYCXPosition
<> end(clang_getRangeEnd(range
));
199 CYCXString
file(start
.file_
);
200 _assert(file
== CYCXString(end
.file_
));
204 char *data(static_cast<char *>(CYPoolFile(temp
, file
, &size
)));
205 _assert(start
.offset_
<= size
&& end
.offset_
<= size
&& start
.offset_
<= end
.offset_
);
208 code
.size
= end
.offset_
- start
.offset_
;
209 code
.data
= pool
.strndup(data
+ start
.offset_
, code
.size
);
213 static CYExpression
*CYTranslateExpression(CXTranslationUnit unit
, CXCursor cursor
) {
214 switch (CXCursorKind kind
= clang_getCursorKind(cursor
)) {
215 case CXCursor_CallExpr
: {
216 CYExpression
*function(NULL
);
217 CYList
<CYArgument
> arguments
;
218 CYForChild(cursor
, fun([&](CXCursor child
) {
219 CYExpression
*expression(CYTranslateExpression(unit
, child
));
220 if (function
== NULL
)
221 function
= expression
;
223 arguments
->*$
C_(expression
);
225 return $
C(function
, arguments
);
228 case CXCursor_DeclRefExpr
: {
229 return $
V(CYCXString(cursor
).Pool($pool
));
232 case CXCursor_IntegerLiteral
: {
233 // libclang doesn't provide any reasonable way to do this
234 // note: clang_tokenize doesn't work if this is a macro
235 // the token range starts inside the macro but ends after it
236 // the tokenizer freaks out and either fails with 0 tokens
237 // or returns some massive number of tokens ending here :/
239 CYUTF8String
token(CYCXPoolUTF8Range($pool
, clang_getCursorExtent(cursor
)));
240 double value(CYCastDouble(token
));
241 if (std::isnan(value
))
242 return $
V(token
.data
);
243 return $
CYNumber(value
);
246 case CXCursor_CStyleCastExpr
:
247 // XXX: most of the time, this is a "NoOp" integer cast; but we should check it
249 case CXCursor_UnexposedExpr
:
250 // there is a very high probability that this is actually an "ImplicitCastExpr"
251 // "Douglas Gregor" <dgregor@apple.com> err'd on the incorrect side of this one
252 // http://lists.llvm.org/pipermail/cfe-commits/Week-of-Mon-20110926/046998.html
254 case CXCursor_ParenExpr
: {
255 CYExpression
*pass(NULL
);
256 CYOneChild(cursor
, fun([&](CXCursor child
) {
257 pass
= CYTranslateExpression(unit
, child
);
263 //std::cerr << "E:" << CYCXString(kind) << std::endl;
268 static CYStatement
*CYTranslateStatement(CXTranslationUnit unit
, CXCursor cursor
) {
269 switch (CXCursorKind kind
= clang_getCursorKind(cursor
)) {
270 case CXCursor_ReturnStmt
: {
271 CYExpression
*value(NULL
);
272 CYOneChild(cursor
, fun([&](CXCursor child
) {
273 value
= CYTranslateExpression(unit
, child
);
275 return $
CYReturn(value
);
279 //std::cerr << "S:" << CYCXString(kind) << std::endl;
284 static CYStatement
*CYTranslateBlock(CXTranslationUnit unit
, CXCursor cursor
) {
285 CYList
<CYStatement
> statements
;
286 CYForChild(cursor
, fun([&](CXCursor child
) {
287 statements
->*CYTranslateStatement(unit
, child
);
289 return $
CYBlock(statements
);
292 static CYType
*CYDecodeType(CXType type
);
293 static void CYParseType(CXType type
, CYType
*typed
);
295 static void CYParseEnumeration(CXCursor cursor
, CYType
*typed
) {
296 CYList
<CYEnumConstant
> constants
;
298 CYForChild(cursor
, fun([&](CXCursor child
) {
299 if (clang_getCursorKind(child
) == CXCursor_EnumConstantDecl
)
300 constants
->*$
CYEnumConstant($
I($pool
.strdup(CYCXString(child
))), $
D(clang_getEnumConstantDeclValue(child
)));
303 CYType
*integer(CYDecodeType(clang_getEnumDeclIntegerType(cursor
)));
304 typed
->specifier_
= $
CYTypeEnum(NULL
, integer
->specifier_
, constants
);
307 static void CYParseStructure(CXCursor cursor
, CYType
*typed
) {
308 CYList
<CYTypeStructField
> fields
;
309 CYForChild(cursor
, fun([&](CXCursor child
) {
310 if (clang_getCursorKind(child
) == CXCursor_FieldDecl
)
311 fields
->*$
CYTypeStructField(CYDecodeType(clang_getCursorType(child
)), $
I(CYCXString(child
).Pool($pool
)));
314 typed
->specifier_
= $
CYTypeStruct(NULL
, $
CYStructTail(fields
));
317 static void CYParseCursor(CXType type
, CXCursor cursor
, CYType
*typed
) {
318 CYCXString
spelling(cursor
);
320 switch (CXCursorKind kind
= clang_getCursorKind(cursor
)) {
321 case CXCursor_EnumDecl
:
322 if (spelling
[0] != '\0')
323 typed
->specifier_
= $
CYTypeReference(CYTypeReferenceEnum
, $
I(spelling
.Pool($pool
)));
325 CYParseEnumeration(cursor
, typed
);
328 case CXCursor_StructDecl
: {
329 if (spelling
[0] != '\0')
330 typed
->specifier_
= $
CYTypeReference(CYTypeReferenceStruct
, $
I(spelling
.Pool($pool
)));
332 CYParseStructure(cursor
, typed
);
335 case CXCursor_UnionDecl
: {
340 std::cerr
<< "C:" << CYCXString(kind
) << std::endl
;
346 static CYTypedParameter
*CYParseSignature(CXType type
, CYType
*typed
) {
347 CYParseType(clang_getResultType(type
), typed
);
348 CYList
<CYTypedParameter
> parameters
;
349 for (int i(0), e(clang_getNumArgTypes(type
)); i
!= e
; ++i
)
350 parameters
->*$
CYTypedParameter(CYDecodeType(clang_getArgType(type
, i
)), NULL
);
354 static void CYParseFunction(CXType type
, CYType
*typed
) {
355 typed
= typed
->Modify($
CYTypeFunctionWith(clang_isFunctionTypeVariadic(type
), CYParseSignature(type
, typed
)));
358 static void CYParseType(CXType type
, CYType
*typed
) {
359 switch (CXTypeKind kind
= type
.kind
) {
360 case CXType_Unexposed
: {
361 CXType
result(clang_getResultType(type
));
362 if (result
.kind
== CXType_Invalid
)
363 CYParseCursor(type
, clang_getTypeDeclaration(type
), typed
);
365 // clang marks function pointers as Unexposed but still supports them
366 CYParseFunction(type
, typed
);
369 case CXType_Bool
: typed
->specifier_
= $
CYTypeVariable("bool"); break;
370 case CXType_WChar
: typed
->specifier_
= $
CYTypeVariable("wchar_t"); break;
371 case CXType_Float
: typed
->specifier_
= $
CYTypeFloating(0); break;
372 case CXType_Double
: typed
->specifier_
= $
CYTypeFloating(1); break;
373 case CXType_LongDouble
: typed
->specifier_
= $
CYTypeFloating(2); break;
375 case CXType_Char_U
: typed
->specifier_
= $
CYTypeCharacter(CYTypeNeutral
); break;
376 case CXType_Char_S
: typed
->specifier_
= $
CYTypeCharacter(CYTypeNeutral
); break;
377 case CXType_SChar
: typed
->specifier_
= $
CYTypeCharacter(CYTypeSigned
); break;
378 case CXType_UChar
: typed
->specifier_
= $
CYTypeCharacter(CYTypeUnsigned
); break;
380 case CXType_Short
: typed
->specifier_
= $
CYTypeIntegral(CYTypeSigned
, 0); break;
381 case CXType_UShort
: typed
->specifier_
= $
CYTypeIntegral(CYTypeUnsigned
, 0); break;
383 case CXType_Int
: typed
->specifier_
= $
CYTypeIntegral(CYTypeSigned
, 1); break;
384 case CXType_UInt
: typed
->specifier_
= $
CYTypeIntegral(CYTypeUnsigned
, 1); break;
386 case CXType_Long
: typed
->specifier_
= $
CYTypeIntegral(CYTypeSigned
, 2); break;
387 case CXType_ULong
: typed
->specifier_
= $
CYTypeIntegral(CYTypeUnsigned
, 2); break;
389 case CXType_LongLong
: typed
->specifier_
= $
CYTypeIntegral(CYTypeSigned
, 3); break;
390 case CXType_ULongLong
: typed
->specifier_
= $
CYTypeIntegral(CYTypeUnsigned
, 3); break;
392 case CXType_Int128
: typed
->specifier_
= $
CYTypeInt128(CYTypeSigned
); break;
393 case CXType_UInt128
: typed
->specifier_
= $
CYTypeInt128(CYTypeUnsigned
); break;
395 case CXType_BlockPointer
: {
396 CXType
pointee(clang_getPointeeType(type
));
397 _assert(!clang_isFunctionTypeVariadic(pointee
));
398 typed
= typed
->Modify($
CYTypeBlockWith(CYParseSignature(pointee
, typed
)));
401 case CXType_ConstantArray
:
402 CYParseType(clang_getArrayElementType(type
), typed
);
403 typed
= typed
->Modify($
CYTypeArrayOf($
D(clang_getArraySize(type
))));
407 typed
->specifier_
= $
CYTypeVariable($pool
.strdup(CYCXString(clang_getTypeSpelling(type
))));
410 case CXType_FunctionProto
:
411 CYParseFunction(type
, typed
);
414 case CXType_IncompleteArray
:
415 // XXX: I probably should not decay to Pointer
416 CYParseType(clang_getArrayElementType(type
), typed
);
417 typed
= typed
->Modify($
CYTypePointerTo());
420 case CXType_ObjCClass
:
421 typed
->specifier_
= $
CYTypeVariable("Class");
425 typed
->specifier_
= $
CYTypeVariable("id");
428 case CXType_ObjCInterface
:
429 typed
->specifier_
= $
CYTypeVariable($pool
.strdup(CYCXString(clang_getTypeSpelling(type
))));
432 case CXType_ObjCObjectPointer
: {
433 CXType
pointee(clang_getPointeeType(type
));
434 if (pointee
.kind
!= CXType_Unexposed
) {
435 CYParseType(pointee
, typed
);
436 typed
= typed
->Modify($
CYTypePointerTo());
438 // Clang seems to have internal typedefs for id and Class that are awkward
443 typed
->specifier_
= $
CYTypeVariable("SEL");
447 CYParseType(clang_getPointeeType(type
), typed
);
448 typed
= typed
->Modify($
CYTypePointerTo());
452 typed
->specifier_
= $
CYTypeReference(CYTypeReferenceStruct
, $
I($pool
.strdup(CYCXString(clang_getTypeSpelling(type
)))));
456 // use the declaration in order to isolate the name of the typedef itself
457 typed
->specifier_
= $
CYTypeVariable($pool
.strdup(CYCXString(clang_getTypeDeclaration(type
))));
465 typed
->specifier_
= $
CYTypeVoid();
469 std::cerr
<< "T:" << CYCXString(clang_getTypeKindSpelling(kind
)) << std::endl
;
470 std::cerr
<< "_: " << CYCXString(clang_getTypeSpelling(type
)) << std::endl
;
474 if (clang_isConstQualifiedType(type
))
475 typed
= typed
->Modify($
CYTypeConstant());
478 static CYType
*CYDecodeType(CXType type
) {
479 CYType
*typed($
CYType(NULL
));
480 CYParseType(type
, typed
);
484 static CXChildVisitResult
CYChildVisit(CXCursor cursor
, CXCursor parent
, CXClientData arg
) {
485 CYChildBaton
&baton(*static_cast<CYChildBaton
*>(arg
));
486 CXTranslationUnit
&unit(baton
.unit
);
488 CXChildVisitResult
result(CXChildVisit_Continue
);
489 CYCXString
spelling(cursor
);
490 std::string
name(spelling
);
491 std::ostringstream value
;
492 unsigned priority(2);
493 unsigned flags(CYBridgeHold
);
495 /*CXSourceLocation location(clang_getCursorLocation(cursor));
496 CYCXPosition<> position(location);
497 std::cerr << spelling << " " << position << std::endl;*/
499 try { switch (CXCursorKind kind
= clang_getCursorKind(cursor
)) {
500 case CXCursor_EnumConstantDecl
: {
501 value
<< clang_getEnumConstantDeclValue(cursor
);
504 case CXCursor_EnumDecl
: {
505 // the enum constants are implemented separately *also*
506 // XXX: maybe move output logic to function we can call
507 result
= CXChildVisit_Recurse
;
509 if (spelling
[0] == '\0')
511 // XXX: this was blindly copied from StructDecl
512 if (!clang_isCursorDefinition(cursor
))
518 CYParseEnumeration(cursor
, &typed
);
521 CYOutput
out(*value
.rdbuf(), options
);
522 CYTypeExpression(&typed
).Output(out
, CYNoBFC
);
524 value
<< ".withName(\"" << name
<< "\")";
525 name
= "$cye" + name
;
526 flags
= CYBridgeType
;
529 case CXCursor_MacroDefinition
: {
530 CXSourceRange
range(clang_getCursorExtent(cursor
));
531 CYTokens
tokens(unit
, range
);
532 _assert(tokens
.size() != 0);
534 CXCursor cursors
[tokens
.size()];
535 clang_annotateTokens(unit
, tokens
, tokens
.size(), cursors
);
538 CYList
<CYFunctionParameter
> parameters
;
541 if (tokens
.size() != 1) {
542 CYCXPosition
<> start(clang_getRangeStart(range
));
543 CYCXString
first(unit
, tokens
[offset
]);
545 CYCXPosition
<> paren(unit
, tokens
[offset
]);
546 if (start
.offset_
+ strlen(spelling
) == paren
.offset_
) {
548 _assert(++offset
!= tokens
.size());
549 CYCXString
token(unit
, tokens
[offset
]);
550 parameters
->*$
P($
B($
I(token
.Pool($pool
))));
551 _assert(++offset
!= tokens
.size());
552 CYCXString
comma(unit
, tokens
[offset
]);
555 _assert(comma
== ",");
562 std::ostringstream body
;
563 for (unsigned i(offset
); i
!= tokens
.size(); ++i
) {
564 CYCXString
token(unit
, tokens
[i
]);
574 CYOutput
out(*value
.rdbuf(), options
);
575 out
<< '(' << "function" << '(';
578 out
<< "return" << ' ';
580 out
<< ';' << '}' << ')';
584 case CXCursor_StructDecl
: {
585 if (spelling
[0] == '\0')
587 if (!clang_isCursorDefinition(cursor
))
593 CYParseStructure(cursor
, &typed
);
596 CYOutput
out(*value
.rdbuf(), options
);
597 CYTypeExpression(&typed
).Output(out
, CYNoBFC
);
599 value
<< ".withName(\"" << name
<< "\")";
600 name
= "$cys" + name
;
601 flags
= CYBridgeType
;
604 case CXCursor_TypedefDecl
: {
607 CYType
*typed(CYDecodeType(clang_getTypedefDeclUnderlyingType(cursor
)));
608 if (typed
->specifier_
== NULL
)
609 value
<< "(typedef " << CYCXString(clang_getTypeSpelling(clang_getTypedefDeclUnderlyingType(cursor
))) << ")";
612 CYOutput
out(*value
.rdbuf(), options
);
613 CYTypeExpression(typed
).Output(out
, CYNoBFC
);
617 case CXCursor_FunctionDecl
:
618 case CXCursor_VarDecl
: {
621 CYList
<CYFunctionParameter
> parameters
;
622 CYStatement
*code(NULL
);
626 CYForChild(cursor
, fun([&](CXCursor child
) {
627 switch (CXCursorKind kind
= clang_getCursorKind(child
)) {
628 case CXCursor_AsmLabelAttr
:
629 label
= CYCXString(child
);
632 case CXCursor_CompoundStmt
:
633 code
= CYTranslateBlock(unit
, child
);
636 case CXCursor_ParmDecl
:
637 parameters
->*$
P($
B($
I(CYCXString(child
).Pool($pool
))));
640 case CXCursor_IntegerLiteral
:
641 case CXCursor_ObjCClassRef
:
642 case CXCursor_TypeRef
:
643 case CXCursor_UnexposedAttr
:
647 //std::cerr << "A:" << CYCXString(child) << std::endl;
655 } else if (label
[0] != '_')
660 CXType
type(clang_getCursorType(cursor
));
661 CYType
*typed(CYDecodeType(type
));
663 CYOutput
out(*value
.rdbuf(), options
);
664 CYTypeExpression(typed
).Output(out
, CYNoBFC
);
665 value
<< ".pointerTo()(dlsym(RTLD_DEFAULT,'" << label
.substr(1) << "'))";
668 CYOutput
out(*value
.rdbuf(), options
);
669 CYFunctionExpression
*function($
CYFunctionExpression(NULL
, parameters
, code
));
670 function
->Output(out
, CYNoBFC
);
671 //std::cerr << value.str() << std::endl;
676 result
= CXChildVisit_Recurse
;
680 CYKey
&key(baton
.keys
[name
]);
681 if (key
.priority_
<= priority
) {
682 key
.priority_
= priority
;
683 key
.code_
= value
.str();
686 } } catch (const CYException
&error
) {
688 //std::cerr << error.PoolCString(pool) << std::endl;
695 int main(int argc
, const char *argv
[]) {
696 CXIndex
index(clang_createIndex(0, 0));
698 const char *file(argv
[1]);
702 argv
[--offset
] = "-ObjC++";
705 CXTranslationUnit
unit(clang_parseTranslationUnit(index
, file
, argv
+ offset
, argc
- offset
, NULL
, 0, CXTranslationUnit_DetailedPreprocessingRecord
));
707 for (unsigned i(0), e(clang_getNumDiagnostics(unit
)); i
!= e
; ++i
) {
708 CXDiagnostic
diagnostic(clang_getDiagnostic(unit
, i
));
709 CYCXString
spelling(clang_getDiagnosticSpelling(diagnostic
));
710 std::cerr
<< spelling
<< std::endl
;
714 CYChildBaton
baton(unit
, keys
);
715 clang_visitChildren(clang_getTranslationUnitCursor(unit
), &CYChildVisit
, &baton
);
717 for (CYKeyMap::const_iterator
key(keys
.begin()); key
!= keys
.end(); ++key
) {
718 std::string
code(key
->second
.code_
);
719 for (size_t i(0), e(code
.size()); i
!= e
; ++i
)
720 if (code
[i
] <= 0 || code
[i
] >= 0x7f || code
[i
] == '\n')
722 std::cout
<< key
->first
<< "|" << key
->second
.flags_
<< "\"" << code
<< "\"" << std::endl
;
725 clang_disposeTranslationUnit(unit
);
726 clang_disposeIndex(index
);