1 | //===- TokenRewriter.cpp - Token-based code rewriting interface -----------===// |
2 | // |
3 | // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions. |
4 | // See https://llvm.org/LICENSE.txt for license information. |
5 | // SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception |
6 | // |
7 | //===----------------------------------------------------------------------===// |
8 | // |
9 | // This file implements the TokenRewriter class, which is used for code |
10 | // transformations. |
11 | // |
12 | //===----------------------------------------------------------------------===// |
13 | |
14 | #include "clang/Rewrite/Core/TokenRewriter.h" |
15 | #include "clang/Basic/SourceManager.h" |
16 | #include "clang/Lex/Lexer.h" |
17 | #include "clang/Lex/ScratchBuffer.h" |
18 | #include "clang/Lex/Token.h" |
19 | #include <cassert> |
20 | #include <cstring> |
21 | #include <map> |
22 | #include <utility> |
23 | |
24 | using namespace clang; |
25 | |
26 | TokenRewriter::TokenRewriter(FileID FID, SourceManager &SM, |
27 | const LangOptions &LangOpts) { |
28 | ScratchBuf.reset(p: new ScratchBuffer(SM)); |
29 | |
30 | // Create a lexer to lex all the tokens of the main file in raw mode. |
31 | llvm::MemoryBufferRef FromFile = SM.getBufferOrFake(FID); |
32 | Lexer RawLex(FID, FromFile, SM, LangOpts); |
33 | |
34 | // Return all comments and whitespace as tokens. |
35 | RawLex.SetKeepWhitespaceMode(true); |
36 | |
37 | // Lex the file, populating our datastructures. |
38 | Token RawTok; |
39 | RawLex.LexFromRawLexer(Result&: RawTok); |
40 | while (RawTok.isNot(K: tok::eof)) { |
41 | #if 0 |
42 | if (Tok.is(tok::raw_identifier)) { |
43 | // Look up the identifier info for the token. This should use |
44 | // IdentifierTable directly instead of PP. |
45 | PP.LookUpIdentifierInfo(Tok); |
46 | } |
47 | #endif |
48 | |
49 | AddToken(T: RawTok, Where: TokenList.end()); |
50 | RawLex.LexFromRawLexer(Result&: RawTok); |
51 | } |
52 | } |
53 | |
54 | TokenRewriter::~TokenRewriter() = default; |
55 | |
56 | /// RemapIterator - Convert from token_iterator (a const iterator) to |
57 | /// TokenRefTy (a non-const iterator). |
58 | TokenRewriter::TokenRefTy TokenRewriter::RemapIterator(token_iterator I) { |
59 | if (I == token_end()) return TokenList.end(); |
60 | |
61 | // FIXME: This is horrible, we should use our own list or something to avoid |
62 | // this. |
63 | std::map<SourceLocation, TokenRefTy>::iterator MapIt = |
64 | TokenAtLoc.find(x: I->getLocation()); |
65 | assert(MapIt != TokenAtLoc.end() && "iterator not in rewriter?" ); |
66 | return MapIt->second; |
67 | } |
68 | |
69 | /// AddToken - Add the specified token into the Rewriter before the other |
70 | /// position. |
71 | TokenRewriter::TokenRefTy |
72 | TokenRewriter::AddToken(const Token &T, TokenRefTy Where) { |
73 | Where = TokenList.insert(position: Where, x: T); |
74 | |
75 | bool InsertSuccess = TokenAtLoc.insert(x: std::make_pair(x: T.getLocation(), |
76 | y&: Where)).second; |
77 | assert(InsertSuccess && "Token location already in rewriter!" ); |
78 | (void)InsertSuccess; |
79 | return Where; |
80 | } |
81 | |
82 | TokenRewriter::token_iterator |
83 | TokenRewriter::AddTokenBefore(token_iterator I, const char *Val) { |
84 | unsigned Len = strlen(s: Val); |
85 | |
86 | // Plop the string into the scratch buffer, then create a token for this |
87 | // string. |
88 | Token Tok; |
89 | Tok.startToken(); |
90 | const char *Spelling; |
91 | Tok.setLocation(ScratchBuf->getToken(Buf: Val, Len, DestPtr&: Spelling)); |
92 | Tok.setLength(Len); |
93 | |
94 | // TODO: Form a whole lexer around this and relex the token! For now, just |
95 | // set kind to tok::unknown. |
96 | Tok.setKind(tok::unknown); |
97 | |
98 | return AddToken(T: Tok, Where: RemapIterator(I)); |
99 | } |
100 | |