Check in the long promised SourceLocation rewrite. This lays the
ground work for implementing #line, and fixes the "out of macro ID's"
problem.
There is nothing particularly tricky about the code, other than the
very performance sensitive SourceManager::getFileID() method.
git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@62978 91177308-0d34-0410-b5e6-96231b3b80d8
diff --git a/lib/Lex/Lexer.cpp b/lib/Lex/Lexer.cpp
index 3174a05..9e8d1aa 100644
--- a/lib/Lex/Lexer.cpp
+++ b/lib/Lex/Lexer.cpp
@@ -169,8 +169,8 @@
// Set the SourceLocation with the remapping information. This ensures that
// GetMappedTokenLoc will remap the tokens as they are lexed.
- L->FileLoc = SM.getInstantiationLoc(SM.getLocForStartOfFile(SpellingFID),
- InstantiationLoc);
+ L->FileLoc = SM.createInstantiationLoc(SM.getLocForStartOfFile(SpellingFID),
+ InstantiationLoc, TokLen);
// Ensure that the lexer thinks it is inside a directive, so that end \n will
// return an EOM token.
@@ -214,16 +214,15 @@
/// that are part of that.
unsigned Lexer::MeasureTokenLength(SourceLocation Loc,
const SourceManager &SM) {
- // If this comes from a macro expansion, we really do want the macro name, not
- // the token this macro expanded to.
- Loc = SM.getInstantiationLoc(Loc);
-
// TODO: this could be special cased for common tokens like identifiers, ')',
// etc to make this faster, if it mattered. Just look at StrData[0] to handle
// all obviously single-char tokens. This could use
// Lexer::isObviouslySimpleCharacter for example to handle identifiers or
// something.
- std::pair<FileID, unsigned> LocInfo = SM.getDecomposedFileLoc(Loc);
+
+ // If this comes from a macro expansion, we really do want the macro name, not
+ // the token this macro expanded to.
+ std::pair<FileID, unsigned> LocInfo = SM.getDecomposedInstantiationLoc(Loc);
std::pair<const char *,const char *> Buffer = SM.getBufferData(LocInfo.first);
const char *StrData = Buffer.first+LocInfo.second;
@@ -310,10 +309,11 @@
/// path of the hot getSourceLocation method. Do not allow it to be inlined.
static SourceLocation GetMappedTokenLoc(Preprocessor &PP,
SourceLocation FileLoc,
- unsigned CharNo) DISABLE_INLINE;
+ unsigned CharNo,
+ unsigned TokLen) DISABLE_INLINE;
static SourceLocation GetMappedTokenLoc(Preprocessor &PP,
SourceLocation FileLoc,
- unsigned CharNo) {
+ unsigned CharNo, unsigned TokLen) {
// Otherwise, we're lexing "mapped tokens". This is used for things like
// _Pragma handling. Combine the instantiation location of FileLoc with the
// spelling location.
@@ -324,12 +324,13 @@
SourceLocation InstLoc = SourceMgr.getInstantiationLoc(FileLoc);
SourceLocation SpellingLoc = SourceMgr.getSpellingLoc(FileLoc);
SpellingLoc = SpellingLoc.getFileLocWithOffset(CharNo);
- return SourceMgr.getInstantiationLoc(SpellingLoc, InstLoc);
+ return SourceMgr.createInstantiationLoc(SpellingLoc, InstLoc, TokLen);
}
/// getSourceLocation - Return a source location identifier for the specified
/// offset in the current file.
-SourceLocation Lexer::getSourceLocation(const char *Loc) const {
+SourceLocation Lexer::getSourceLocation(const char *Loc,
+ unsigned TokLen) const {
assert(Loc >= BufferStart && Loc <= BufferEnd &&
"Location out of range for this buffer!");
@@ -342,7 +343,7 @@
// Otherwise, this is the _Pragma lexer case, which pretends that all of the
// tokens are lexed from where the _Pragma was defined.
assert(PP && "This doesn't work on raw lexers");
- return GetMappedTokenLoc(*PP, FileLoc, CharNo);
+ return GetMappedTokenLoc(*PP, FileLoc, CharNo, TokLen);
}
/// Diag - Forwarding function for diagnostics. This translate a source