/// \brief Obtain the original source code text from a SourceRange.
static StringRef getStringFromRange(SourceManager &SourceMgr,
const LangOptions &LangOpts,
SourceRange Range) {
if (SourceMgr.getFileID(Range.getBegin()) !=
SourceMgr.getFileID(Range.getEnd()))
return NULL;
CharSourceRange SourceChars(Range, true);
return Lexer::getSourceText(SourceChars, SourceMgr, LangOpts);
}
/// Find the suitable set of lines to show to include a set of ranges.
static llvm::Optional<std::pair<unsigned, unsigned>>
findLinesForRange(const CharSourceRange &R, FileID FID,
const SourceManager &SM) {
if (!R.isValid()) return None;
SourceLocation Begin = R.getBegin();
SourceLocation End = R.getEnd();
if (SM.getFileID(Begin) != FID || SM.getFileID(End) != FID)
return None;
return std::make_pair(SM.getExpansionLineNumber(Begin),
SM.getExpansionLineNumber(End));
}
/// Determine whether two source locations come from the same file.
static bool IsFromSameFile(SourceManager &SM, SourceLocation DirectiveLoc,
SourceLocation DiagnosticLoc) {
while (DiagnosticLoc.isMacroID())
DiagnosticLoc = SM.getImmediateMacroCallerLoc(DiagnosticLoc);
if (SM.isWrittenInSameFile(DirectiveLoc, DiagnosticLoc))
return true;
const FileEntry *DiagFile = SM.getFileEntryForID(SM.getFileID(DiagnosticLoc));
if (!DiagFile && SM.isWrittenInMainFile(DirectiveLoc))
return true;
return (DiagFile == SM.getFileEntryForID(SM.getFileID(DirectiveLoc)));
}
static void PrintSourceForLocation(const SourceLocation &Loc,
SourceManager &SM) {
const char *LocData = SM.getCharacterData(Loc, /*Invalid=*/nullptr);
unsigned LocColumn =
SM.getSpellingColumnNumber(Loc, /*Invalid=*/nullptr) - 1;
FileID FID = SM.getFileID(Loc);
llvm::MemoryBuffer *Buffer = SM.getBuffer(FID, Loc, /*Invalid=*/nullptr);
assert(LocData >= Buffer->getBufferStart() &&
LocData < Buffer->getBufferEnd());
const char *LineBegin = LocData - LocColumn;
assert(LineBegin >= Buffer->getBufferStart());
const char *LineEnd = nullptr;
for (LineEnd = LineBegin; *LineEnd != '\n' && *LineEnd != '\r' &&
LineEnd < Buffer->getBufferEnd();
++LineEnd)
;
llvm::StringRef LineString(LineBegin, LineEnd - LineBegin);
llvm::errs() << LineString << '\n';
std::string Space(LocColumn, ' ');
llvm::errs() << Space.c_str() << '\n';
}
static unsigned GetFID(const FIDMap& FIDs, const SourceManager &SM,
SourceLocation L) {
FileID FID = SM.getFileID(SM.getExpansionLoc(L));
FIDMap::const_iterator I = FIDs.find(FID);
assert(I != FIDs.end());
return I->second;
}
/// \brief Retrieve the name of the immediate macro expansion.
///
/// This routine starts from a source location, and finds the name of the macro
/// responsible for its immediate expansion. It looks through any intervening
/// macro argument expansions to compute this. It returns a StringRef which
/// refers to the SourceManager-owned buffer of the source where that macro
/// name is spelled. Thus, the result shouldn't out-live that SourceManager.
///
/// This differs from Lexer::getImmediateMacroName in that any macro argument
/// location will result in the topmost function macro that accepted it.
/// e.g.
/// \code
/// MAC1( MAC2(foo) )
/// \endcode
/// for location of 'foo' token, this function will return "MAC1" while
/// Lexer::getImmediateMacroName will return "MAC2".
static StringRef getImmediateMacroName(SourceLocation Loc,
const SourceManager &SM,
const LangOptions &LangOpts) {
assert(Loc.isMacroID() && "Only reasonble to call this on macros");
// Walk past macro argument expanions.
while (SM.isMacroArgExpansion(Loc))
Loc = SM.getImmediateExpansionRange(Loc).first;
// If the macro's spelling has no FileID, then it's actually a token paste
// or stringization (or similar) and not a macro at all.
if (!SM.getFileEntryForID(SM.getFileID(SM.getSpellingLoc(Loc))))
return StringRef();
// Find the spelling location of the start of the non-argument expansion
// range. This is where the macro name was spelled in order to begin
// expanding this macro.
Loc = SM.getSpellingLoc(SM.getImmediateExpansionRange(Loc).first);
// Dig out the buffer where the macro name was spelled and the extents of the
// name so that we can render it into the expansion note.
std::pair<FileID, unsigned> ExpansionInfo = SM.getDecomposedLoc(Loc);
unsigned MacroTokenLength = Lexer::MeasureTokenLength(Loc, SM, LangOpts);
StringRef ExpansionBuffer = SM.getBufferData(ExpansionInfo.first);
return ExpansionBuffer.substr(ExpansionInfo.second, MacroTokenLength);
}
static void Scan(IvarUsageMap &M, const DeclContext *C, const FileID FID,
SourceManager &SM) {
for (const auto *I : C->decls())
if (const auto *FD = dyn_cast<FunctionDecl>(I)) {
SourceLocation L = FD->getLocStart();
if (SM.getFileID(L) == FID)
Scan(M, FD->getBody());
}
}
static void AddFID(FIDMap &FIDs, SmallVectorImpl<FileID> &V,
const SourceManager &SM, SourceLocation L) {
FileID FID = SM.getFileID(SM.getExpansionLoc(L));
FIDMap::iterator I = FIDs.find(FID);
if (I != FIDs.end()) return;
FIDs[FID] = V.size();
V.push_back(FID);
}
static void Scan(IvarUsageMap &M, const DeclContext *C, const FileID FID,
SourceManager &SM) {
for (DeclContext::decl_iterator I=C->decls_begin(), E=C->decls_end();
I!=E; ++I)
if (const FunctionDecl *FD = dyn_cast<FunctionDecl>(*I)) {
SourceLocation L = FD->getLocStart();
if (SM.getFileID(L) == FID)
Scan(M, FD->getBody());
}
}
bool Transform::isFileModifiable(const SourceManager &SM,
const SourceLocation &Loc) const {
if (SM.isWrittenInMainFile(Loc))
return true;
const FileEntry *FE = SM.getFileEntryForID(SM.getFileID(Loc));
if (!FE)
return false;
return GlobalOptions.ModifiableFiles.isFileIncluded(FE->getName());
}
void MakeSmartPtrCheck::checkReset(SourceManager &SM,
const CXXMemberCallExpr *Reset,
const CXXNewExpr *New) {
const auto *Expr = cast<MemberExpr>(Reset->getCallee());
SourceLocation OperatorLoc = Expr->getOperatorLoc();
SourceLocation ResetCallStart = Reset->getExprLoc();
SourceLocation ExprStart = Expr->getLocStart();
SourceLocation ExprEnd =
Lexer::getLocForEndOfToken(Expr->getLocEnd(), 0, SM, getLangOpts());
bool InMacro = ExprStart.isMacroID();
if (InMacro && IgnoreMacros) {
return;
}
// There are some cases where we don't have operator ("." or "->") of the
// "reset" expression, e.g. call "reset()" method directly in the subclass of
// "std::unique_ptr<>". We skip these cases.
if (OperatorLoc.isInvalid()) {
return;
}
auto Diag = diag(ResetCallStart, "use %0 instead")
<< MakeSmartPtrFunctionName;
// Disable the fix in macros.
if (InMacro) {
return;
}
if (!replaceNew(Diag, New, SM)) {
return;
}
Diag << FixItHint::CreateReplacement(
CharSourceRange::getCharRange(OperatorLoc, ExprEnd),
(llvm::Twine(" = ") + MakeSmartPtrFunctionName + "<" +
GetNewExprName(New, SM, getLangOpts()) + ">")
.str());
if (Expr->isArrow())
Diag << FixItHint::CreateInsertion(ExprStart, "*");
insertHeader(Diag, SM.getFileID(OperatorLoc));
}
/// \brief Find the end of the end of the directive, either the beginning of a
/// newline or the end of file.
//
// \return The offset into the file where the directive ends along with a
// boolean value indicating whether the directive ends because the end of file
// was reached or not.
static std::pair<unsigned, bool> findDirectiveEnd(SourceLocation HashLoc,
SourceManager &SM,
const LangOptions &LangOpts) {
FileID FID = SM.getFileID(HashLoc);
unsigned Offset = SM.getFileOffset(HashLoc);
StringRef Content = SM.getBufferData(FID);
Lexer Lex(SM.getLocForStartOfFile(FID), LangOpts, Content.begin(),
Content.begin() + Offset, Content.end());
Lex.SetCommentRetentionState(true);
Token Tok;
// This loop look for the newline after our directive but avoids the ones part
// of a multi-line comments:
//
// #include <foo> /* long \n comment */\n
// ~~ no ~~ yes
for (;;) {
// find the beginning of the end-of-line sequence
StringRef::size_type EOLOffset = Content.find_first_of("\r\n", Offset);
// ends because EOF was reached
if (EOLOffset == StringRef::npos)
return std::make_pair(Content.size(), true);
// find the token that contains our end-of-line
unsigned TokEnd = 0;
do {
Lex.LexFromRawLexer(Tok);
TokEnd = SM.getFileOffset(Tok.getLocation()) + Tok.getLength();
// happens when the whitespaces are eaten after a multiline comment
if (Tok.is(tok::eof))
return std::make_pair(EOLOffset, false);
} while (TokEnd < EOLOffset);
// the end-of-line is not part of a multi-line comment, return its location
if (Tok.isNot(tok::comment))
return std::make_pair(EOLOffset, false);
// for the next search to start after the end of this token
Offset = TokEnd;
}
}
// visit a class/struct/enum node
bool VisitCXXRecordDecl(CXXRecordDecl *D) {
if(!D) return true;
//llvm::errs() << "begin to visit record: " << D->getNameAsString() << "\n";
SourceLocation Loc = D->getLocation();
Loc = SM->getSpellingLoc(Loc);
FileID ID = SM->getFileID(Loc);
//llvm::errs() << "end to visit record: " << D->getNameAsString() << "\n";
// we do not handle the declaration if it's in the system header file
if (!SM->isInSystemHeader(Loc) && !SM->isInExternCSystemHeader(Loc)) {
//llvm::errs() << ">>>>> visit record: " << D->getNameAsString() << "\n";
Record *r = HandleRecordDecl(D);
if (r) {
records.insert(std::make_pair(r->qualifiedname, r));
}
} //else llvm::errs() << "uninteresting record: " << D->getNameAsString() << "\n";
return true;
}
// Checks if 'typedef' keyword can be removed - we do it only if
// it is the only declaration in a declaration chain.
static bool CheckRemoval(SourceManager &SM, const SourceLocation &LocStart,
const SourceLocation &LocEnd, ASTContext &Context,
SourceRange &ResultRange) {
FileID FID = SM.getFileID(LocEnd);
llvm::MemoryBuffer *Buffer = SM.getBuffer(FID, LocEnd);
Lexer DeclLexer(SM.getLocForStartOfFile(FID), Context.getLangOpts(),
Buffer->getBufferStart(), SM.getCharacterData(LocStart),
Buffer->getBufferEnd());
Token DeclToken;
bool result = false;
int parenthesisLevel = 0;
while (!DeclLexer.LexFromRawLexer(DeclToken)) {
if (DeclToken.getKind() == tok::TokenKind::l_paren)
parenthesisLevel++;
if (DeclToken.getKind() == tok::TokenKind::r_paren)
parenthesisLevel--;
if (DeclToken.getKind() == tok::TokenKind::semi)
break;
// if there is comma and we are not between open parenthesis then it is
// two or more declatarions in this chain
if (parenthesisLevel == 0 && DeclToken.getKind() == tok::TokenKind::comma)
return false;
if (DeclToken.isOneOf(tok::TokenKind::identifier,
tok::TokenKind::raw_identifier)) {
auto TokenStr = DeclToken.getRawIdentifier().str();
if (TokenStr == "typedef") {
ResultRange =
SourceRange(DeclToken.getLocation(), DeclToken.getEndLoc());
result = true;
}
}
}
// assert if there was keyword 'typedef' in declaration
assert(result && "No typedef found");
return result;
}
// Looks for the token matching the predicate and returns the range of the found
// token including trailing whitespace.
static SourceRange FindToken(const SourceManager &Sources, LangOptions LangOpts,
SourceLocation StartLoc, SourceLocation EndLoc,
bool (*Pred)(const Token &)) {
if (StartLoc.isMacroID() || EndLoc.isMacroID())
return SourceRange();
FileID File = Sources.getFileID(Sources.getSpellingLoc(StartLoc));
StringRef Buf = Sources.getBufferData(File);
const char *StartChar = Sources.getCharacterData(StartLoc);
Lexer Lex(StartLoc, LangOpts, StartChar, StartChar, Buf.end());
Lex.SetCommentRetentionState(true);
Token Tok;
do {
Lex.LexFromRawLexer(Tok);
if (Pred(Tok)) {
Token NextTok;
Lex.LexFromRawLexer(NextTok);
return SourceRange(Tok.getLocation(), NextTok.getLocation());
}
} while (Tok.isNot(tok::eof) && Tok.getLocation() < EndLoc);
return SourceRange();
}
/// \brief Highlight a SourceRange (with ~'s) for any characters on LineNo.
void TextDiagnostic::highlightRange(const CharSourceRange &R,
unsigned LineNo, FileID FID,
const SourceColumnMap &map,
std::string &CaretLine,
const SourceManager &SM) {
if (!R.isValid()) return;
SourceLocation Begin = SM.getExpansionLoc(R.getBegin());
SourceLocation End = SM.getExpansionLoc(R.getEnd());
// If the End location and the start location are the same and are a macro
// location, then the range was something that came from a macro expansion
// or _Pragma. If this is an object-like macro, the best we can do is to
// highlight the range. If this is a function-like macro, we'd also like to
// highlight the arguments.
if (Begin == End && R.getEnd().isMacroID())
End = SM.getExpansionRange(R.getEnd()).second;
unsigned StartLineNo = SM.getExpansionLineNumber(Begin);
if (StartLineNo > LineNo || SM.getFileID(Begin) != FID)
return; // No intersection.
unsigned EndLineNo = SM.getExpansionLineNumber(End);
if (EndLineNo < LineNo || SM.getFileID(End) != FID)
return; // No intersection.
// Compute the column number of the start.
unsigned StartColNo = 0;
if (StartLineNo == LineNo) {
StartColNo = SM.getExpansionColumnNumber(Begin);
if (StartColNo) --StartColNo; // Zero base the col #.
}
// Compute the column number of the end.
unsigned EndColNo = map.getSourceLine().size();
if (EndLineNo == LineNo) {
EndColNo = SM.getExpansionColumnNumber(End);
if (EndColNo) {
--EndColNo; // Zero base the col #.
// Add in the length of the token, so that we cover multi-char tokens if
// this is a token range.
if (R.isTokenRange())
EndColNo += Lexer::MeasureTokenLength(End, SM, LangOpts);
} else {
EndColNo = CaretLine.size();
}
}
assert(StartColNo <= EndColNo && "Invalid range!");
// Check that a token range does not highlight only whitespace.
if (R.isTokenRange()) {
// Pick the first non-whitespace column.
while (StartColNo < map.getSourceLine().size() &&
(map.getSourceLine()[StartColNo] == ' ' ||
map.getSourceLine()[StartColNo] == '\t'))
++StartColNo;
// Pick the last non-whitespace column.
if (EndColNo > map.getSourceLine().size())
EndColNo = map.getSourceLine().size();
while (EndColNo-1 &&
(map.getSourceLine()[EndColNo-1] == ' ' ||
map.getSourceLine()[EndColNo-1] == '\t'))
--EndColNo;
// If the start/end passed each other, then we are trying to highlight a
// range that just exists in whitespace, which must be some sort of other
// bug.
assert(StartColNo <= EndColNo && "Trying to highlight whitespace??");
}
assert(StartColNo <= map.getSourceLine().size() && "Invalid range!");
assert(EndColNo <= map.getSourceLine().size() && "Invalid range!");
// Fill the range with ~'s.
StartColNo = map.byteToColumn(StartColNo);
EndColNo = map.byteToColumn(EndColNo);
assert(StartColNo <= EndColNo && "Invalid range!");
if (CaretLine.size() < EndColNo)
CaretLine.resize(EndColNo,' ');
std::fill(CaretLine.begin()+StartColNo,CaretLine.begin()+EndColNo,'~');
}
static bool locationsInSameFile(const SourceManager &Sources,
SourceLocation Loc1, SourceLocation Loc2) {
return Loc1.isFileID() && Loc2.isFileID() &&
Sources.getFileID(Loc1) == Sources.getFileID(Loc2);
}
void MakeSmartPtrCheck::checkConstruct(SourceManager &SM,
const CXXConstructExpr *Construct,
const QualType *Type,
const CXXNewExpr *New) {
SourceLocation ConstructCallStart = Construct->getExprLoc();
bool InMacro = ConstructCallStart.isMacroID();
if (InMacro && IgnoreMacros) {
return;
}
bool Invalid = false;
StringRef ExprStr = Lexer::getSourceText(
CharSourceRange::getCharRange(
ConstructCallStart, Construct->getParenOrBraceRange().getBegin()),
SM, getLangOpts(), &Invalid);
if (Invalid)
return;
auto Diag = diag(ConstructCallStart, "use %0 instead")
<< MakeSmartPtrFunctionName;
// Disable the fix in macros.
if (InMacro) {
return;
}
if (!replaceNew(Diag, New, SM)) {
return;
}
// Find the location of the template's left angle.
size_t LAngle = ExprStr.find("<");
SourceLocation ConstructCallEnd;
if (LAngle == StringRef::npos) {
// If the template argument is missing (because it is part of the alias)
// we have to add it back.
ConstructCallEnd = ConstructCallStart.getLocWithOffset(ExprStr.size());
Diag << FixItHint::CreateInsertion(
ConstructCallEnd,
"<" + GetNewExprName(New, SM, getLangOpts()) + ">");
} else {
ConstructCallEnd = ConstructCallStart.getLocWithOffset(LAngle);
}
Diag << FixItHint::CreateReplacement(
CharSourceRange::getCharRange(ConstructCallStart, ConstructCallEnd),
MakeSmartPtrFunctionName);
// If the smart_ptr is built with brace enclosed direct initialization, use
// parenthesis instead.
if (Construct->isListInitialization()) {
SourceRange BraceRange = Construct->getParenOrBraceRange();
Diag << FixItHint::CreateReplacement(
CharSourceRange::getCharRange(
BraceRange.getBegin(), BraceRange.getBegin().getLocWithOffset(1)),
"(");
Diag << FixItHint::CreateReplacement(
CharSourceRange::getCharRange(BraceRange.getEnd(),
BraceRange.getEnd().getLocWithOffset(1)),
")");
}
insertHeader(Diag, SM.getFileID(ConstructCallStart));
}
/// ParseDirective - Go through the comment and see if it indicates expected
/// diagnostics. If so, then put them in the appropriate directive list.
///
/// Returns true if any valid directives were found.
static bool ParseDirective(StringRef S, ExpectedData *ED, SourceManager &SM,
SourceLocation Pos, DiagnosticsEngine &Diags,
VerifyDiagnosticConsumer::DirectiveStatus &Status) {
// A single comment may contain multiple directives.
bool FoundDirective = false;
for (ParseHelper PH(S); !PH.Done();) {
// Search for token: expected
if (!PH.Search("expected", true))
break;
PH.Advance();
// Next token: -
if (!PH.Next("-"))
continue;
PH.Advance();
// Next token: { error | warning | note }
DirectiveList* DL = NULL;
if (PH.Next("error"))
DL = ED ? &ED->Errors : NULL;
else if (PH.Next("warning"))
DL = ED ? &ED->Warnings : NULL;
else if (PH.Next("note"))
DL = ED ? &ED->Notes : NULL;
else if (PH.Next("no-diagnostics")) {
if (Status == VerifyDiagnosticConsumer::HasOtherExpectedDirectives)
Diags.Report(Pos, diag::err_verify_invalid_no_diags)
<< /*IsExpectedNoDiagnostics=*/true;
else
Status = VerifyDiagnosticConsumer::HasExpectedNoDiagnostics;
continue;
} else
continue;
PH.Advance();
if (Status == VerifyDiagnosticConsumer::HasExpectedNoDiagnostics) {
Diags.Report(Pos, diag::err_verify_invalid_no_diags)
<< /*IsExpectedNoDiagnostics=*/false;
continue;
}
Status = VerifyDiagnosticConsumer::HasOtherExpectedDirectives;
// If a directive has been found but we're not interested
// in storing the directive information, return now.
if (!DL)
return true;
// Default directive kind.
bool RegexKind = false;
const char* KindStr = "string";
// Next optional token: -
if (PH.Next("-re")) {
PH.Advance();
RegexKind = true;
KindStr = "regex";
}
// Next optional token: @
SourceLocation ExpectedLoc;
if (!PH.Next("@")) {
ExpectedLoc = Pos;
} else {
PH.Advance();
unsigned Line = 0;
bool FoundPlus = PH.Next("+");
if (FoundPlus || PH.Next("-")) {
// Relative to current line.
PH.Advance();
bool Invalid = false;
unsigned ExpectedLine = SM.getSpellingLineNumber(Pos, &Invalid);
if (!Invalid && PH.Next(Line) && (FoundPlus || Line < ExpectedLine)) {
if (FoundPlus) ExpectedLine += Line;
else ExpectedLine -= Line;
ExpectedLoc = SM.translateLineCol(SM.getFileID(Pos), ExpectedLine, 1);
}
} else {
// Absolute line number.
if (PH.Next(Line) && Line > 0)
ExpectedLoc = SM.translateLineCol(SM.getFileID(Pos), Line, 1);
}
if (ExpectedLoc.isInvalid()) {
Diags.Report(Pos.getLocWithOffset(PH.C-PH.Begin),
diag::err_verify_missing_line) << KindStr;
continue;
}
PH.Advance();
}
// Skip optional whitespace.
PH.SkipWhitespace();
// Next optional token: positive integer or a '+'.
unsigned Min = 1;
unsigned Max = 1;
if (PH.Next(Min)) {
PH.Advance();
// A positive integer can be followed by a '+' meaning min
// or more, or by a '-' meaning a range from min to max.
if (PH.Next("+")) {
Max = Directive::MaxCount;
PH.Advance();
} else if (PH.Next("-")) {
PH.Advance();
if (!PH.Next(Max) || Max < Min) {
Diags.Report(Pos.getLocWithOffset(PH.C-PH.Begin),
diag::err_verify_invalid_range) << KindStr;
continue;
}
PH.Advance();
} else {
Max = Min;
}
} else if (PH.Next("+")) {
// '+' on its own means "1 or more".
Max = Directive::MaxCount;
PH.Advance();
}
// Skip optional whitespace.
PH.SkipWhitespace();
// Next token: {{
if (!PH.Next("{{")) {
Diags.Report(Pos.getLocWithOffset(PH.C-PH.Begin),
diag::err_verify_missing_start) << KindStr;
continue;
}
PH.Advance();
const char* const ContentBegin = PH.C; // mark content begin
// Search for token: }}
if (!PH.Search("}}")) {
Diags.Report(Pos.getLocWithOffset(PH.C-PH.Begin),
diag::err_verify_missing_end) << KindStr;
continue;
}
const char* const ContentEnd = PH.P; // mark content end
PH.Advance();
// Build directive text; convert \n to newlines.
std::string Text;
StringRef NewlineStr = "\\n";
StringRef Content(ContentBegin, ContentEnd-ContentBegin);
size_t CPos = 0;
size_t FPos;
while ((FPos = Content.find(NewlineStr, CPos)) != StringRef::npos) {
Text += Content.substr(CPos, FPos-CPos);
Text += '\n';
CPos = FPos + NewlineStr.size();
}
if (Text.empty())
Text.assign(ContentBegin, ContentEnd);
// Construct new directive.
Directive *D = Directive::create(RegexKind, Pos, ExpectedLoc, Text,
Min, Max);
std::string Error;
if (D->isValid(Error)) {
DL->push_back(D);
FoundDirective = true;
} else {
Diags.Report(Pos.getLocWithOffset(ContentBegin-PH.Begin),
diag::err_verify_invalid_content)
<< KindStr << Error;
}
}
return FoundDirective;
}
/// HighlightRange - Given a SourceRange and a line number, highlight (with ~'s)
/// any characters in LineNo that intersect the SourceRange.
void TextDiagnosticPrinter::HighlightRange(const SourceRange &R,
const SourceManager &SM,
unsigned LineNo, FileID FID,
std::string &CaretLine,
const std::string &SourceLine) {
assert(CaretLine.size() == SourceLine.size() &&
"Expect a correspondence between source and caret line!");
if (!R.isValid()) return;
SourceLocation Begin = SM.getInstantiationLoc(R.getBegin());
SourceLocation End = SM.getInstantiationLoc(R.getEnd());
// If the End location and the start location are the same and are a macro
// location, then the range was something that came from a macro expansion
// or _Pragma. If this is an object-like macro, the best we can do is to
// highlight the range. If this is a function-like macro, we'd also like to
// highlight the arguments.
if (Begin == End && R.getEnd().isMacroID())
End = SM.getInstantiationRange(R.getEnd()).second;
unsigned StartLineNo = SM.getInstantiationLineNumber(Begin);
if (StartLineNo > LineNo || SM.getFileID(Begin) != FID)
return; // No intersection.
unsigned EndLineNo = SM.getInstantiationLineNumber(End);
if (EndLineNo < LineNo || SM.getFileID(End) != FID)
return; // No intersection.
// Compute the column number of the start.
unsigned StartColNo = 0;
if (StartLineNo == LineNo) {
StartColNo = SM.getInstantiationColumnNumber(Begin);
if (StartColNo) --StartColNo; // Zero base the col #.
}
// Compute the column number of the end.
unsigned EndColNo = CaretLine.size();
if (EndLineNo == LineNo) {
EndColNo = SM.getInstantiationColumnNumber(End);
if (EndColNo) {
--EndColNo; // Zero base the col #.
// Add in the length of the token, so that we cover multi-char tokens.
EndColNo += Lexer::MeasureTokenLength(End, SM, *LangOpts);
} else {
EndColNo = CaretLine.size();
}
}
assert(StartColNo <= EndColNo && "Invalid range!");
// Pick the first non-whitespace column.
while (StartColNo < SourceLine.size() &&
(SourceLine[StartColNo] == ' ' || SourceLine[StartColNo] == '\t'))
++StartColNo;
// Pick the last non-whitespace column.
if (EndColNo > SourceLine.size())
EndColNo = SourceLine.size();
while (EndColNo-1 &&
(SourceLine[EndColNo-1] == ' ' || SourceLine[EndColNo-1] == '\t'))
--EndColNo;
// If the start/end passed each other, then we are trying to highlight a range
// that just exists in whitespace, which must be some sort of other bug.
assert(StartColNo <= EndColNo && "Trying to highlight whitespace??");
// Fill the range with ~'s.
for (unsigned i = StartColNo; i < EndColNo; ++i)
CaretLine[i] = '~';
}
/// ParseDirective - Go through the comment and see if it indicates expected
/// diagnostics. If so, then put them in the appropriate directive list.
///
/// Returns true if any valid directives were found.
static bool ParseDirective(StringRef S, ExpectedData *ED, SourceManager &SM,
Preprocessor *PP, SourceLocation Pos,
VerifyDiagnosticConsumer::DirectiveStatus &Status,
VerifyDiagnosticConsumer::MarkerTracker &Markers) {
DiagnosticsEngine &Diags = PP ? PP->getDiagnostics() : SM.getDiagnostics();
// First, scan the comment looking for markers.
for (ParseHelper PH(S); !PH.Done();) {
if (!PH.Search("#", true))
break;
PH.C = PH.P;
if (!PH.NextMarker()) {
PH.Next("#");
PH.Advance();
continue;
}
PH.Advance();
Markers.addMarker(PH.Match(), Pos);
}
// A single comment may contain multiple directives.
bool FoundDirective = false;
for (ParseHelper PH(S); !PH.Done();) {
// Search for the initial directive token.
// If one prefix, save time by searching only for its directives.
// Otherwise, search for any potential directive token and check it later.
const auto &Prefixes = Diags.getDiagnosticOptions().VerifyPrefixes;
if (!(Prefixes.size() == 1 ? PH.Search(*Prefixes.begin(), true, true)
: PH.Search("", true, true)))
break;
StringRef DToken = PH.Match();
PH.Advance();
// Default directive kind.
UnattachedDirective D;
const char *KindStr = "string";
// Parse the initial directive token in reverse so we can easily determine
// its exact actual prefix. If we were to parse it from the front instead,
// it would be harder to determine where the prefix ends because there
// might be multiple matching -verify prefixes because some might prefix
// others.
// Regex in initial directive token: -re
if (DToken.endswith("-re")) {
D.RegexKind = true;
KindStr = "regex";
DToken = DToken.substr(0, DToken.size()-3);
}
// Type in initial directive token: -{error|warning|note|no-diagnostics}
bool NoDiag = false;
StringRef DType;
if (DToken.endswith(DType="-error"))
D.DL = ED ? &ED->Errors : nullptr;
else if (DToken.endswith(DType="-warning"))
D.DL = ED ? &ED->Warnings : nullptr;
else if (DToken.endswith(DType="-remark"))
D.DL = ED ? &ED->Remarks : nullptr;
else if (DToken.endswith(DType="-note"))
D.DL = ED ? &ED->Notes : nullptr;
else if (DToken.endswith(DType="-no-diagnostics")) {
NoDiag = true;
if (D.RegexKind)
continue;
}
else
continue;
DToken = DToken.substr(0, DToken.size()-DType.size());
// What's left in DToken is the actual prefix. That might not be a -verify
// prefix even if there is only one -verify prefix (for example, the full
// DToken is foo-bar-warning, but foo is the only -verify prefix).
if (!std::binary_search(Prefixes.begin(), Prefixes.end(), DToken))
continue;
if (NoDiag) {
if (Status == VerifyDiagnosticConsumer::HasOtherExpectedDirectives)
Diags.Report(Pos, diag::err_verify_invalid_no_diags)
<< /*IsExpectedNoDiagnostics=*/true;
else
Status = VerifyDiagnosticConsumer::HasExpectedNoDiagnostics;
continue;
}
if (Status == VerifyDiagnosticConsumer::HasExpectedNoDiagnostics) {
Diags.Report(Pos, diag::err_verify_invalid_no_diags)
<< /*IsExpectedNoDiagnostics=*/false;
continue;
}
Status = VerifyDiagnosticConsumer::HasOtherExpectedDirectives;
// If a directive has been found but we're not interested
// in storing the directive information, return now.
if (!D.DL)
return true;
// Next optional token: @
SourceLocation ExpectedLoc;
StringRef Marker;
bool MatchAnyLine = false;
if (!PH.Next("@")) {
ExpectedLoc = Pos;
} else {
PH.Advance();
unsigned Line = 0;
bool FoundPlus = PH.Next("+");
if (FoundPlus || PH.Next("-")) {
// Relative to current line.
PH.Advance();
bool Invalid = false;
unsigned ExpectedLine = SM.getSpellingLineNumber(Pos, &Invalid);
if (!Invalid && PH.Next(Line) && (FoundPlus || Line < ExpectedLine)) {
if (FoundPlus) ExpectedLine += Line;
else ExpectedLine -= Line;
ExpectedLoc = SM.translateLineCol(SM.getFileID(Pos), ExpectedLine, 1);
}
} else if (PH.Next(Line)) {
// Absolute line number.
if (Line > 0)
ExpectedLoc = SM.translateLineCol(SM.getFileID(Pos), Line, 1);
} else if (PH.NextMarker()) {
Marker = PH.Match();
} else if (PP && PH.Search(":")) {
// Specific source file.
StringRef Filename(PH.C, PH.P-PH.C);
PH.Advance();
// Lookup file via Preprocessor, like a #include.
const DirectoryLookup *CurDir;
const FileEntry *FE =
PP->LookupFile(Pos, Filename, false, nullptr, nullptr, CurDir,
nullptr, nullptr, nullptr, nullptr, nullptr);
if (!FE) {
Diags.Report(Pos.getLocWithOffset(PH.C-PH.Begin),
diag::err_verify_missing_file) << Filename << KindStr;
continue;
}
if (SM.translateFile(FE).isInvalid())
SM.createFileID(FE, Pos, SrcMgr::C_User);
if (PH.Next(Line) && Line > 0)
ExpectedLoc = SM.translateFileLineCol(FE, Line, 1);
else if (PH.Next("*")) {
MatchAnyLine = true;
ExpectedLoc = SM.translateFileLineCol(FE, 1, 1);
}
} else if (PH.Next("*")) {
MatchAnyLine = true;
ExpectedLoc = SourceLocation();
}
if (ExpectedLoc.isInvalid() && !MatchAnyLine && Marker.empty()) {
Diags.Report(Pos.getLocWithOffset(PH.C-PH.Begin),
diag::err_verify_missing_line) << KindStr;
continue;
}
PH.Advance();
}
// Skip optional whitespace.
PH.SkipWhitespace();
// Next optional token: positive integer or a '+'.
if (PH.Next(D.Min)) {
PH.Advance();
// A positive integer can be followed by a '+' meaning min
// or more, or by a '-' meaning a range from min to max.
if (PH.Next("+")) {
D.Max = Directive::MaxCount;
PH.Advance();
} else if (PH.Next("-")) {
PH.Advance();
if (!PH.Next(D.Max) || D.Max < D.Min) {
Diags.Report(Pos.getLocWithOffset(PH.C-PH.Begin),
diag::err_verify_invalid_range) << KindStr;
continue;
}
PH.Advance();
} else {
D.Max = D.Min;
}
} else if (PH.Next("+")) {
// '+' on its own means "1 or more".
D.Max = Directive::MaxCount;
PH.Advance();
}
// Skip optional whitespace.
PH.SkipWhitespace();
// Next token: {{
if (!PH.Next("{{")) {
Diags.Report(Pos.getLocWithOffset(PH.C-PH.Begin),
diag::err_verify_missing_start) << KindStr;
continue;
}
PH.Advance();
const char* const ContentBegin = PH.C; // mark content begin
// Search for token: }}
if (!PH.SearchClosingBrace("{{", "}}")) {
Diags.Report(Pos.getLocWithOffset(PH.C-PH.Begin),
diag::err_verify_missing_end) << KindStr;
continue;
}
const char* const ContentEnd = PH.P; // mark content end
PH.Advance();
D.DirectivePos = Pos;
D.ContentBegin = Pos.getLocWithOffset(ContentBegin - PH.Begin);
// Build directive text; convert \n to newlines.
StringRef NewlineStr = "\\n";
StringRef Content(ContentBegin, ContentEnd-ContentBegin);
size_t CPos = 0;
size_t FPos;
while ((FPos = Content.find(NewlineStr, CPos)) != StringRef::npos) {
D.Text += Content.substr(CPos, FPos-CPos);
D.Text += '\n';
CPos = FPos + NewlineStr.size();
}
if (D.Text.empty())
D.Text.assign(ContentBegin, ContentEnd);
// Check that regex directives contain at least one regex.
if (D.RegexKind && D.Text.find("{{") == StringRef::npos) {
Diags.Report(D.ContentBegin, diag::err_verify_missing_regex) << D.Text;
return false;
}
if (Marker.empty())
attachDirective(Diags, D, ExpectedLoc, MatchAnyLine);
else
Markers.addDirective(Marker, D);
FoundDirective = true;
}
return FoundDirective;
}
/// \brief Highlight a SourceRange (with ~'s) for any characters on LineNo.
static void highlightRange(const CharSourceRange &R,
unsigned LineNo, FileID FID,
const SourceColumnMap &map,
std::string &CaretLine,
const SourceManager &SM,
const LangOptions &LangOpts) {
if (!R.isValid()) return;
SourceLocation Begin = R.getBegin();
SourceLocation End = R.getEnd();
unsigned StartLineNo = SM.getExpansionLineNumber(Begin);
if (StartLineNo > LineNo || SM.getFileID(Begin) != FID)
return; // No intersection.
unsigned EndLineNo = SM.getExpansionLineNumber(End);
if (EndLineNo < LineNo || SM.getFileID(End) != FID)
return; // No intersection.
// Compute the column number of the start.
unsigned StartColNo = 0;
if (StartLineNo == LineNo) {
StartColNo = SM.getExpansionColumnNumber(Begin);
if (StartColNo) --StartColNo; // Zero base the col #.
}
// Compute the column number of the end.
unsigned EndColNo = map.getSourceLine().size();
if (EndLineNo == LineNo) {
EndColNo = SM.getExpansionColumnNumber(End);
if (EndColNo) {
--EndColNo; // Zero base the col #.
// Add in the length of the token, so that we cover multi-char tokens if
// this is a token range.
if (R.isTokenRange())
EndColNo += Lexer::MeasureTokenLength(End, SM, LangOpts);
} else {
EndColNo = CaretLine.size();
}
}
assert(StartColNo <= EndColNo && "Invalid range!");
// Check that a token range does not highlight only whitespace.
if (R.isTokenRange()) {
// Pick the first non-whitespace column.
while (StartColNo < map.getSourceLine().size() &&
(map.getSourceLine()[StartColNo] == ' ' ||
map.getSourceLine()[StartColNo] == '\t'))
StartColNo = map.startOfNextColumn(StartColNo);
// Pick the last non-whitespace column.
if (EndColNo > map.getSourceLine().size())
EndColNo = map.getSourceLine().size();
while (EndColNo-1 &&
(map.getSourceLine()[EndColNo-1] == ' ' ||
map.getSourceLine()[EndColNo-1] == '\t'))
EndColNo = map.startOfPreviousColumn(EndColNo);
// If the start/end passed each other, then we are trying to highlight a
// range that just exists in whitespace, which must be some sort of other
// bug.
assert(StartColNo <= EndColNo && "Trying to highlight whitespace??");
}
assert(StartColNo <= map.getSourceLine().size() && "Invalid range!");
assert(EndColNo <= map.getSourceLine().size() && "Invalid range!");
// Fill the range with ~'s.
StartColNo = map.byteToContainingColumn(StartColNo);
EndColNo = map.byteToContainingColumn(EndColNo);
assert(StartColNo <= EndColNo && "Invalid range!");
if (CaretLine.size() < EndColNo)
CaretLine.resize(EndColNo,' ');
std::fill(CaretLine.begin()+StartColNo,CaretLine.begin()+EndColNo,'~');
}
/// \brief Print out the file/line/column information and include trace.
///
/// This method handlen the emission of the diagnostic location information.
/// This includes extracting as much location information as is present for
/// the diagnostic and printing it, as well as any include stack or source
/// ranges necessary.
void TextDiagnostic::emitDiagnosticLoc(SourceLocation Loc, PresumedLoc PLoc,
DiagnosticsEngine::Level Level,
ArrayRef<CharSourceRange> Ranges,
const SourceManager &SM) {
if (PLoc.isInvalid()) {
// At least print the file name if available:
FileID FID = SM.getFileID(Loc);
if (!FID.isInvalid()) {
const FileEntry* FE = SM.getFileEntryForID(FID);
if (FE && FE->getName()) {
OS << FE->getName();
if (FE->getDevice() == 0 && FE->getInode() == 0
&& FE->getFileMode() == 0) {
// in PCH is a guess, but a good one:
OS << " (in PCH)";
}
OS << ": ";
}
}
return;
}
unsigned LineNo = PLoc.getLine();
if (!DiagOpts->ShowLocation)
return;
if (DiagOpts->ShowColors)
OS.changeColor(savedColor, true);
OS << PLoc.getFilename();
switch (DiagOpts->getFormat()) {
case DiagnosticOptions::LFort: OS << ':' << LineNo; break;
case DiagnosticOptions::Msvc: OS << '(' << LineNo; break;
case DiagnosticOptions::Vi: OS << " +" << LineNo; break;
}
if (DiagOpts->ShowColumn)
// Compute the column number.
if (unsigned ColNo = PLoc.getColumn()) {
if (DiagOpts->getFormat() == DiagnosticOptions::Msvc) {
OS << ',';
ColNo--;
} else
OS << ':';
OS << ColNo;
}
switch (DiagOpts->getFormat()) {
case DiagnosticOptions::LFort:
case DiagnosticOptions::Vi: OS << ':'; break;
case DiagnosticOptions::Msvc: OS << ") : "; break;
}
if (DiagOpts->ShowSourceRanges && !Ranges.empty()) {
FileID CaretFileID =
SM.getFileID(SM.getExpansionLoc(Loc));
bool PrintedRange = false;
for (ArrayRef<CharSourceRange>::const_iterator RI = Ranges.begin(),
RE = Ranges.end();
RI != RE; ++RI) {
// Ignore invalid ranges.
if (!RI->isValid()) continue;
SourceLocation B = SM.getExpansionLoc(RI->getBegin());
SourceLocation E = SM.getExpansionLoc(RI->getEnd());
// If the End location and the start location are the same and are a
// macro location, then the range was something that came from a
// macro expansion or _Pragma. If this is an object-like macro, the
// best we can do is to highlight the range. If this is a
// function-like macro, we'd also like to highlight the arguments.
if (B == E && RI->getEnd().isMacroID())
E = SM.getExpansionRange(RI->getEnd()).second;
std::pair<FileID, unsigned> BInfo = SM.getDecomposedLoc(B);
std::pair<FileID, unsigned> EInfo = SM.getDecomposedLoc(E);
// If the start or end of the range is in another file, just discard
// it.
if (BInfo.first != CaretFileID || EInfo.first != CaretFileID)
continue;
// Add in the length of the token, so that we cover multi-char
// tokens.
unsigned TokSize = 0;
if (RI->isTokenRange())
TokSize = Lexer::MeasureTokenLength(E, SM, LangOpts);
OS << '{' << SM.getLineNumber(BInfo.first, BInfo.second) << ':'
<< SM.getColumnNumber(BInfo.first, BInfo.second) << '-'
<< SM.getLineNumber(EInfo.first, EInfo.second) << ':'
<< (SM.getColumnNumber(EInfo.first, EInfo.second)+TokSize)
<< '}';
PrintedRange = true;
}
if (PrintedRange)
OS << ':';
}
OS << ' ';
}
