static void escapeAndPrintString(llvm::raw_ostream &os, StringRef Str) {
if (Str.empty()) {
// Special-case the empty string.
os << "\"\"";
return;
}
bool NeedsEscape = Str.find_first_of(" \"\\$") != StringRef::npos;
if (!NeedsEscape) {
// This string doesn't have anything we need to escape, so print it directly
os << Str;
return;
}
// Quote and escape. This isn't really complete, but is good enough, and
// matches how Clang's Command handles escaping arguments.
os << '"';
for (const char c : Str) {
switch (c) {
case '"':
case '\\':
case '$':
// These characters need to be escaped.
os << '\\';
// Fall-through to the default case, since we still need to print the
// character.
SWIFT_FALLTHROUGH;
default:
os << c;
}
}
os << '"';
}
std::string get_line_from_offset(StringRef buffer, std::size_t offset) {
assert(buffer.size() > offset);
auto line_start = buffer.find_last_of("\r\n", offset) + 1;
auto line_end = buffer.find_first_of("\r\n", offset);
return std::string(buffer.begin() + line_start, buffer.begin() + line_end);
}
/// Print the filename, with escaping or quoting that accommodates the three
/// most likely tools that use dependency files: GNU Make, BSD Make, and
/// NMake/Jom.
///
/// BSD Make is the simplest case: It does no escaping at all. This means
/// characters that are normally delimiters, i.e. space and # (the comment
/// character) simply aren't supported in filenames.
///
/// GNU Make does allow space and # in filenames, but to avoid being treated
/// as a delimiter or comment, these must be escaped with a backslash. Because
/// backslash is itself the escape character, if a backslash appears in a
/// filename, it should be escaped as well. (As a special case, $ is escaped
/// as $$, which is the normal Make way to handle the $ character.)
/// For compatibility with BSD Make and historical practice, if GNU Make
/// un-escapes characters in a filename but doesn't find a match, it will
/// retry with the unmodified original string.
///
/// GCC tries to accommodate both Make formats by escaping any space or #
/// characters in the original filename, but not escaping backslashes. The
/// apparent intent is so that filenames with backslashes will be handled
/// correctly by BSD Make, and by GNU Make in its fallback mode of using the
/// unmodified original string; filenames with # or space characters aren't
/// supported by BSD Make at all, but will be handled correctly by GNU Make
/// due to the escaping.
///
/// A corner case that GCC gets only partly right is when the original filename
/// has a backslash immediately followed by space or #. GNU Make would expect
/// this backslash to be escaped; however GCC escapes the original backslash
/// only when followed by space, not #. It will therefore take a dependency
/// from a directive such as
/// #include "a\ b\#c.h"
/// and emit it as
/// a\\\ b\\#c.h
/// which GNU Make will interpret as
/// a\ b\
/// followed by a comment. Failing to find this file, it will fall back to the
/// original string, which probably doesn't exist either; in any case it won't
/// find
/// a\ b\#c.h
/// which is the actual filename specified by the include directive.
///
/// Clang does what GCC does, rather than what GNU Make expects.
///
/// NMake/Jom has a different set of scary characters, but wraps filespecs in
/// double-quotes to avoid misinterpreting them; see
/// https://msdn.microsoft.com/en-us/library/dd9y37ha.aspx for NMake info,
/// https://msdn.microsoft.com/en-us/library/windows/desktop/aa365247(v=vs.85).aspx
/// for Windows file-naming info.
static void PrintFilename(raw_ostream &OS, StringRef Filename,
DependencyOutputFormat OutputFormat) {
if (OutputFormat == DependencyOutputFormat::NMake) {
// Add quotes if needed. These are the characters listed as "special" to
// NMake, that are legal in a Windows filespec, and that could cause
// misinterpretation of the dependency string.
if (Filename.find_first_of(" #${}^!") != StringRef::npos)
OS << '\"' << Filename << '\"';
else
OS << Filename;
return;
}
assert(OutputFormat == DependencyOutputFormat::Make);
for (unsigned i = 0, e = Filename.size(); i != e; ++i) {
if (Filename[i] == '#') // Handle '#' the broken gcc way.
OS << '\\';
else if (Filename[i] == ' ') { // Handle space correctly.
OS << '\\';
unsigned j = i;
while (j > 0 && Filename[--j] == '\\')
OS << '\\';
} else if (Filename[i] == '$') // $ is escaped by $$.
OS << '$';
OS << Filename[i];
}
}
static void outputReplacementXML(StringRef Text) {
// FIXME: When we sort includes, we need to make sure the stream is correct
// utf-8.
size_t From = 0;
size_t Index;
while ((Index = Text.find_first_of("\n\r<&", From)) != StringRef::npos) {
outs() << Text.substr(From, Index - From);
switch (Text[Index]) {
case '\n':
outs() << " ";
break;
case '\r':
outs() << " ";
break;
case '<':
outs() << "<";
break;
case '&':
outs() << "&";
break;
default:
llvm_unreachable("Unexpected character encountered!");
}
From = Index + 1;
}
outs() << Text.substr(From);
}
static bool ByteArrayFromString(ByteArrayTy &ByteArray,
StringRef &Str,
SourceMgr &SM) {
while (SkipToToken(Str)) {
// Handled by higher level
if (Str[0] == '[' || Str[0] == ']')
return false;
// Get the current token.
size_t Next = Str.find_first_of(" \t\n\r,#[]");
StringRef Value = Str.substr(0, Next);
// Convert to a byte and add to the byte vector.
unsigned ByteVal;
if (Value.getAsInteger(0, ByteVal) || ByteVal > 255) {
// If we have an error, print it and skip to the end of line.
SM.PrintMessage(SMLoc::getFromPointer(Value.data()), SourceMgr::DK_Error,
"invalid input token");
Str = Str.substr(Str.find('\n'));
ByteArray.first.clear();
ByteArray.second.clear();
continue;
}
ByteArray.first.push_back(ByteVal);
ByteArray.second.push_back(Value.data());
Str = Str.substr(Next);
}
return false;
}
static bool ByteArrayFromString(ByteArrayTy &ByteArray,
StringRef &Str,
SourceMgr &SM) {
while (!Str.empty()) {
// Strip horizontal whitespace.
if (size_t Pos = Str.find_first_not_of(" \t\r")) {
Str = Str.substr(Pos);
continue;
}
// If this is the end of a line or start of a comment, remove the rest of
// the line.
if (Str[0] == '\n' || Str[0] == '#') {
// Strip to the end of line if we already processed any bytes on this
// line. This strips the comment and/or the \n.
if (Str[0] == '\n') {
Str = Str.substr(1);
} else {
Str = Str.substr(Str.find_first_of('\n'));
if (!Str.empty())
Str = Str.substr(1);
}
continue;
}
// Get the current token.
size_t Next = Str.find_first_of(" \t\n\r#");
StringRef Value = Str.substr(0, Next);
// Convert to a byte and add to the byte vector.
unsigned ByteVal;
if (Value.getAsInteger(0, ByteVal) || ByteVal > 255) {
// If we have an error, print it and skip to the end of line.
SM.PrintMessage(SMLoc::getFromPointer(Value.data()), SourceMgr::DK_Error,
"invalid input token");
Str = Str.substr(Str.find('\n'));
ByteArray.clear();
continue;
}
ByteArray.push_back(std::make_pair((unsigned char)ByteVal, Value.data()));
Str = Str.substr(Next);
}
return false;
}
// Returns a whole line containing the current token.
StringRef ScriptParserBase::getLine() {
StringRef S = getCurrentMB().getBuffer();
StringRef Tok = Tokens[Pos - 1];
size_t Pos = S.rfind('\n', Tok.data() - S.data());
if (Pos != StringRef::npos)
S = S.substr(Pos + 1);
return S.substr(0, S.find_first_of("\r\n"));
}
VecString split( const char* chars, StringRef str )
{
VecString vs;
String::size_type n0 = 0;
String::size_type n1 = str.find_first_of( chars, 0 );
vs.push_back( str.substr( n0, n1 ) );
while( n1 != String::npos )
{
n0 = n1+1;
n1 = str.find_first_of( chars, n0 );
if( n1 != String::npos ) vs.push_back( str.substr( n0, n1-n0 ) );
else vs.push_back( str.substr( n0 ) );
}
return vs;
}
// Primary pass pipeline description parsing routine.
// FIXME: Should this routine accept a TargetMachine or require the caller to
// pre-populate the analysis managers with target-specific stuff?
bool PassBuilder::parsePassPipeline(ModulePassManager &MPM,
StringRef PipelineText, bool VerifyEachPass,
bool DebugLogging) {
// By default, try to parse the pipeline as-if it were within an implicit
// 'module(...)' pass pipeline. If this will parse at all, it needs to
// consume the entire string.
if (parseModulePassPipeline(MPM, PipelineText, VerifyEachPass, DebugLogging))
return PipelineText.empty();
// This isn't parsable as a module pipeline, look for the end of a pass name
// and directly drop down to that layer.
StringRef FirstName =
PipelineText.substr(0, PipelineText.find_first_of(",)"));
assert(!isModulePassName(FirstName) &&
"Already handled all module pipeline options.");
// If this looks like a CGSCC pass, parse the whole thing as a CGSCC
// pipeline.
if (PipelineText.startswith("cgscc(") || isCGSCCPassName(FirstName)) {
CGSCCPassManager CGPM(DebugLogging);
if (!parseCGSCCPassPipeline(CGPM, PipelineText, VerifyEachPass,
DebugLogging) ||
!PipelineText.empty())
return false;
MPM.addPass(createModuleToPostOrderCGSCCPassAdaptor(std::move(CGPM)));
return true;
}
// Similarly, if this looks like a Function pass, parse the whole thing as
// a Function pipelien.
if (PipelineText.startswith("function(") || isFunctionPassName(FirstName)) {
FunctionPassManager FPM(DebugLogging);
if (!parseFunctionPassPipeline(FPM, PipelineText, VerifyEachPass,
DebugLogging) ||
!PipelineText.empty())
return false;
MPM.addPass(createModuleToFunctionPassAdaptor(std::move(FPM)));
return true;
}
// If this looks like a Loop pass, parse the whole thing as a Loop pipeline.
if (PipelineText.startswith("loop(") || isLoopPassName(FirstName)) {
LoopPassManager LPM(DebugLogging);
if (!parseLoopPassPipeline(LPM, PipelineText, VerifyEachPass,
DebugLogging) ||
!PipelineText.empty())
return false;
FunctionPassManager FPM(DebugLogging);
FPM.addPass(createFunctionToLoopPassAdaptor(std::move(LPM)));
MPM.addPass(createModuleToFunctionPassAdaptor(std::move(FPM)));
return true;
}
return false;
}
Optional<std::vector<PassBuilder::PipelineElement>>
PassBuilder::parsePipelineText(StringRef Text) {
std::vector<PipelineElement> ResultPipeline;
SmallVector<std::vector<PipelineElement> *, 4> PipelineStack = {
&ResultPipeline};
for (;;) {
std::vector<PipelineElement> &Pipeline = *PipelineStack.back();
size_t Pos = Text.find_first_of(",()");
Pipeline.push_back({Text.substr(0, Pos), {}});
// If we have a single terminating name, we're done.
if (Pos == Text.npos)
break;
char Sep = Text[Pos];
Text = Text.substr(Pos + 1);
if (Sep == ',')
// Just a name ending in a comma, continue.
continue;
if (Sep == '(') {
// Push the inner pipeline onto the stack to continue processing.
PipelineStack.push_back(&Pipeline.back().InnerPipeline);
continue;
}
assert(Sep == ')' && "Bogus separator!");
// When handling the close parenthesis, we greedily consume them to avoid
// empty strings in the pipeline.
do {
// If we try to pop the outer pipeline we have unbalanced parentheses.
if (PipelineStack.size() == 1)
return None;
PipelineStack.pop_back();
} while (Text.consume_front(")"));
// Check if we've finished parsing.
if (Text.empty())
break;
// Otherwise, the end of an inner pipeline always has to be followed by
// a comma, and then we can continue.
if (!Text.consume_front(","))
return None;
}
if (PipelineStack.size() > 1)
// Unbalanced paretheses.
return None;
assert(PipelineStack.back() == &ResultPipeline &&
"Wrong pipeline at the bottom of the stack!");
return {std::move(ResultPipeline)};
}
void Value::setNameImpl(const Twine &NewName) {
// Fast-path: LLVMContext can be set to strip out non-GlobalValue names
if (getContext().shouldDiscardValueNames() && !isa<GlobalValue>(this))
return;
// Fast path for common IRBuilder case of setName("") when there is no name.
if (NewName.isTriviallyEmpty() && !hasName())
return;
SmallString<256> NameData;
StringRef NameRef = NewName.toStringRef(NameData);
assert(NameRef.find_first_of(0) == StringRef::npos &&
"Null bytes are not allowed in names");
// Name isn't changing?
if (getName() == NameRef)
return;
assert(!getType()->isVoidTy() && "Cannot assign a name to void values!");
// Get the symbol table to update for this object.
ValueSymbolTable *ST;
if (getSymTab(this, ST))
return; // Cannot set a name on this value (e.g. constant).
if (!ST) { // No symbol table to update? Just do the change.
if (NameRef.empty()) {
// Free the name for this value.
destroyValueName();
return;
}
// NOTE: Could optimize for the case the name is shrinking to not deallocate
// then reallocated.
destroyValueName();
// Create the new name.
setValueName(ValueName::Create(NameRef));
getValueName()->setValue(this);
return;
}
// NOTE: Could optimize for the case the name is shrinking to not deallocate
// then reallocated.
if (hasName()) {
// Remove old name.
ST->removeValueName(getValueName());
destroyValueName();
if (NameRef.empty())
return;
}
// Name is changing to something new.
setValueName(ST->createValueName(NameRef, this));
}
static SourceLoc findEndOfLine(SourceManager &SM, SourceLoc loc,
unsigned bufferID) {
CharSourceRange entireBuffer = SM.getRangeForBuffer(bufferID);
CharSourceRange rangeFromLoc{SM, loc, entireBuffer.getEnd()};
StringRef textFromLoc = SM.extractText(rangeFromLoc);
size_t newlineOffset = textFromLoc.find_first_of({"\r\n\0", 3});
if (newlineOffset == StringRef::npos)
return entireBuffer.getEnd();
return loc.getAdvancedLoc(newlineOffset);
}
/// \brief Guess the end-of-line sequence used in the given FileID. If the
/// sequence can't be guessed return an Unix-style newline.
static StringRef guessEOL(SourceManager &SM, FileID ID) {
StringRef Content = SM.getBufferData(ID);
StringRef Buffer = Content.substr(Content.find_first_of("\r\n"));
return llvm::StringSwitch<StringRef>(Buffer)
.StartsWith("\r\n", "\r\n")
.StartsWith("\n\r", "\n\r")
.StartsWith("\r", "\r")
.Default("\n");
}
/// getToken - This function extracts one token from source, ignoring any
/// leading characters that appear in the Delimiters string, and ending the
/// token at any of the characters that appear in the Delimiters string. If
/// there are no tokens in the source string, an empty string is returned.
/// The function returns a pair containing the extracted token and the
/// remaining tail string.
std::pair<StringRef, StringRef> llvm::getToken(StringRef Source,
StringRef Delimiters) {
// Figure out where the token starts.
StringRef::size_type Start = Source.find_first_not_of(Delimiters);
// Find the next occurrence of the delimiter.
StringRef::size_type End = Source.find_first_of(Delimiters, Start);
return std::make_pair(Source.slice(Start, End), Source.substr(End));
}
static std::string QuoteProgramPathIfNeeded(StringRef Command) {
if (Command.find_first_of(' ') == StringRef::npos)
return Command;
else {
std::string ret;
ret.reserve(Command.size() + 3);
ret.push_back('"');
ret.append(Command.begin(), Command.end());
ret.push_back('"');
return ret;
}
}
bool PassBuilder::parseFunctionPassPipeline(FunctionPassManager &FPM,
StringRef &PipelineText,
bool VerifyEachPass,
bool DebugLogging) {
for (;;) {
// Parse nested pass managers by recursing.
if (PipelineText.startswith("function(")) {
FunctionPassManager NestedFPM(DebugLogging);
// Parse the inner pipeline inte the nested manager.
PipelineText = PipelineText.substr(strlen("function("));
if (!parseFunctionPassPipeline(NestedFPM, PipelineText, VerifyEachPass,
DebugLogging) ||
PipelineText.empty())
return false;
assert(PipelineText[0] == ')');
PipelineText = PipelineText.substr(1);
// Add the nested pass manager with the appropriate adaptor.
FPM.addPass(std::move(NestedFPM));
} else if (PipelineText.startswith("loop(")) {
LoopPassManager NestedLPM(DebugLogging);
// Parse the inner pipeline inte the nested manager.
PipelineText = PipelineText.substr(strlen("loop("));
if (!parseLoopPassPipeline(NestedLPM, PipelineText, VerifyEachPass,
DebugLogging) ||
PipelineText.empty())
return false;
assert(PipelineText[0] == ')');
PipelineText = PipelineText.substr(1);
// Add the nested pass manager with the appropriate adaptor.
FPM.addPass(createFunctionToLoopPassAdaptor(std::move(NestedLPM)));
} else {
// Otherwise try to parse a pass name.
size_t End = PipelineText.find_first_of(",)");
if (!parseFunctionPassName(FPM, PipelineText.substr(0, End)))
return false;
if (VerifyEachPass)
FPM.addPass(VerifierPass());
PipelineText = PipelineText.substr(End);
}
if (PipelineText.empty() || PipelineText[0] == ')')
return true;
assert(PipelineText[0] == ',');
PipelineText = PipelineText.substr(1);
}
}
static bool parseCGSCCPassPipeline(CGSCCPassManager &CGPM,
StringRef &PipelineText, bool VerifyEachPass,
bool DebugLogging) {
for (;;) {
// Parse nested pass managers by recursing.
if (PipelineText.startswith("cgscc(")) {
CGSCCPassManager NestedCGPM(DebugLogging);
// Parse the inner pipeline into the nested manager.
PipelineText = PipelineText.substr(strlen("cgscc("));
if (!parseCGSCCPassPipeline(NestedCGPM, PipelineText, VerifyEachPass,
DebugLogging) ||
PipelineText.empty())
return false;
assert(PipelineText[0] == ')');
PipelineText = PipelineText.substr(1);
// Add the nested pass manager with the appropriate adaptor.
CGPM.addPass(std::move(NestedCGPM));
} else if (PipelineText.startswith("function(")) {
FunctionPassManager NestedFPM(DebugLogging);
// Parse the inner pipeline inte the nested manager.
PipelineText = PipelineText.substr(strlen("function("));
if (!parseFunctionPassPipeline(NestedFPM, PipelineText, VerifyEachPass,
DebugLogging) ||
PipelineText.empty())
return false;
assert(PipelineText[0] == ')');
PipelineText = PipelineText.substr(1);
// Add the nested pass manager with the appropriate adaptor.
CGPM.addPass(createCGSCCToFunctionPassAdaptor(std::move(NestedFPM)));
} else {
// Otherwise try to parse a pass name.
size_t End = PipelineText.find_first_of(",)");
if (!parseCGSCCPassName(CGPM, PipelineText.substr(0, End)))
return false;
// FIXME: No verifier support for CGSCC passes!
PipelineText = PipelineText.substr(End);
}
if (PipelineText.empty() || PipelineText[0] == ')')
return true;
assert(PipelineText[0] == ',');
PipelineText = PipelineText.substr(1);
}
}
static bool parseModulePassPipeline(ModulePassManager &MPM,
StringRef &PipelineText,
bool VerifyEachPass) {
for (;;) {
// Parse nested pass managers by recursing.
if (PipelineText.startswith("module(")) {
ModulePassManager NestedMPM;
// Parse the inner pipeline into the nested manager.
PipelineText = PipelineText.substr(strlen("module("));
if (!parseModulePassPipeline(NestedMPM, PipelineText, VerifyEachPass) ||
PipelineText.empty())
return false;
assert(PipelineText[0] == ')');
PipelineText = PipelineText.substr(1);
// Now add the nested manager as a module pass.
MPM.addPass(std::move(NestedMPM));
} else if (PipelineText.startswith("function(")) {
FunctionPassManager NestedFPM;
// Parse the inner pipeline inte the nested manager.
PipelineText = PipelineText.substr(strlen("function("));
if (!parseFunctionPassPipeline(NestedFPM, PipelineText, VerifyEachPass) ||
PipelineText.empty())
return false;
assert(PipelineText[0] == ')');
PipelineText = PipelineText.substr(1);
// Add the nested pass manager with the appropriate adaptor.
MPM.addPass(createModuleToFunctionPassAdaptor(std::move(NestedFPM)));
} else {
// Otherwise try to parse a pass name.
size_t End = PipelineText.find_first_of(",)");
if (!parseModulePassName(MPM, PipelineText.substr(0, End)))
return false;
if (VerifyEachPass)
MPM.addPass(VerifierPass());
PipelineText = PipelineText.substr(End);
}
if (PipelineText.empty() || PipelineText[0] == ')')
return true;
assert(PipelineText[0] == ',');
PipelineText = PipelineText.substr(1);
}
}
/// CountNumNewlinesBetween - Count the number of newlines in the specified
/// range.
static unsigned CountNumNewlinesBetween(StringRef Range) {
unsigned NumNewLines = 0;
while (1) {
// Scan for newline.
Range = Range.substr(Range.find_first_of("\n\r"));
if (Range.empty()) return NumNewLines;
++NumNewLines;
// Handle \n\r and \r\n as a single newline.
if (Range.size() > 1 &&
(Range[1] == '\n' || Range[1] == '\r') &&
(Range[0] != Range[1]))
Range = Range.substr(1);
Range = Range.substr(1);
}
}
// Append a #define line to Buf for Macro. Macro should be of the form XXX,
// in which case we emit "#define XXX 1" or "XXX=Y z W" in which case we emit
// "#define XXX Y z W". To get a #define with no value, use "XXX=".
static void DefineBuiltinMacro(MacroBuilder &Builder, StringRef Macro,
DiagnosticsEngine &Diags) {
std::pair<StringRef, StringRef> MacroPair = Macro.split('=');
StringRef MacroName = MacroPair.first;
StringRef MacroBody = MacroPair.second;
if (MacroName.size() != Macro.size()) {
// Per GCC -D semantics, the macro ends at \n if it exists.
StringRef::size_type End = MacroBody.find_first_of("\n\r");
if (End != StringRef::npos)
Diags.Report(diag::warn_fe_macro_contains_embedded_newline)
<< MacroName;
Builder.defineMacro(MacroName, MacroBody.substr(0, End));
} else {
// Push "macroname 1".
Builder.defineMacro(Macro);
}
}
// Find file from search paths. You can omit ".obj", this function takes
// care of that. Note that the returned path is not guaranteed to exist.
StringRef LinkerDriver::doFindFile(StringRef Filename) {
bool hasPathSep = (Filename.find_first_of("/\\") != StringRef::npos);
if (hasPathSep)
return Filename;
bool hasExt = (Filename.find('.') != StringRef::npos);
for (StringRef Dir : SearchPaths) {
SmallString<128> Path = Dir;
llvm::sys::path::append(Path, Filename);
if (llvm::sys::fs::exists(Path.str()))
return Alloc.save(Path.str());
if (!hasExt) {
Path.append(".obj");
if (llvm::sys::fs::exists(Path.str()))
return Alloc.save(Path.str());
}
}
return Filename;
}
static bool SkipToToken(StringRef &Str) {
for (;;) {
if (Str.empty())
return false;
// Strip horizontal whitespace and commas.
if (size_t Pos = Str.find_first_not_of(" \t\r\n,")) {
Str = Str.substr(Pos);
continue;
}
// If this is the start of a comment, remove the rest of the line.
if (Str[0] == '#') {
Str = Str.substr(Str.find_first_of('\n'));
continue;
}
return true;
}
}
static void outputReplacementXML(StringRef Text) {
size_t From = 0;
size_t Index;
while ((Index = Text.find_first_of("\n\r", From)) != StringRef::npos) {
llvm::outs() << Text.substr(From, Index - From);
switch (Text[Index]) {
case '\n':
llvm::outs() << " ";
break;
case '\r':
llvm::outs() << " ";
break;
default:
llvm_unreachable("Unexpected character encountered!");
}
From = Index + 1;
}
llvm::outs() << Text.substr(From);
}
/// \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;
}
}
size_t swift::ide::getOffsetOfLine(unsigned LineIndex, StringRef Text) {
// SourceLoc start = SourceLoc(llvm::SMLoc::getFromPointer(Text.begin()));
// FIXME: We should have a cached line map in EditableTextBuffer, for now
// we just do the slow naive thing here.
size_t LineOffset = 0;
unsigned CurrentLine = 0;
while (LineOffset < Text.size() && ++CurrentLine < LineIndex) {
LineOffset = Text.find_first_of("\r\n", LineOffset);
if (LineOffset != std::string::npos) {
++LineOffset;
if (LineOffset < Text.size() &&
Text[LineOffset - 1] == '\r' && Text[LineOffset] == '\n')
++LineOffset;
}
}
if (LineOffset == std::string::npos)
LineOffset = 0;
return LineOffset;
}
void MCMachObjectSymbolizer::
tryAddingPcLoadReferenceComment(raw_ostream &cStream, int64_t Value,
uint64_t Address) {
if (const RelocationRef *R = findRelocationAt(Address)) {
const MCExpr *RelExpr = RelInfo->createExprForRelocation(*R);
if (!RelExpr || RelExpr->EvaluateAsAbsolute(Value) == false)
return;
}
uint64_t Addr = Value;
if (const SectionRef *S = findSectionContaining(Addr)) {
StringRef Name; S->getName(Name);
uint64_t SAddr; S->getAddress(SAddr);
if (Name == "__cstring") {
StringRef Contents;
S->getContents(Contents);
Contents = Contents.substr(Addr - SAddr);
cStream << " ## literal pool for: "
<< Contents.substr(0, Contents.find_first_of(0));
}
}
}
void
swift::trimLeadingWhitespaceFromLines(StringRef RawText,
unsigned WhitespaceToTrim,
SmallVectorImpl<StringRef> &OutLines) {
SmallVector<StringRef, 8> Lines;
bool IsFirstLine = true;
while (!RawText.empty()) {
size_t Pos = RawText.find_first_of("\n\r");
if (Pos == StringRef::npos)
Pos = RawText.size();
StringRef Line = RawText.substr(0, Pos);
Lines.push_back(Line);
if (!IsFirstLine) {
size_t NonWhitespacePos = RawText.find_first_not_of(' ');
if (NonWhitespacePos != StringRef::npos)
WhitespaceToTrim =
std::min(WhitespaceToTrim,
static_cast<unsigned>(NonWhitespacePos));
}
IsFirstLine = false;
RawText = RawText.drop_front(Pos);
unsigned NewlineBytes = measureNewline(RawText);
RawText = RawText.drop_front(NewlineBytes);
}
IsFirstLine = true;
for (auto &Line : Lines) {
if (!IsFirstLine) {
Line = Line.drop_front(WhitespaceToTrim);
}
IsFirstLine = false;
}
OutLines.append(Lines.begin(), Lines.end());
}
bool PassBuilder::parseLoopPassPipeline(LoopPassManager &LPM,
StringRef &PipelineText,
bool VerifyEachPass,
bool DebugLogging) {
for (;;) {
// Parse nested pass managers by recursing.
if (PipelineText.startswith("loop(")) {
LoopPassManager NestedLPM(DebugLogging);
// Parse the inner pipeline inte the nested manager.
PipelineText = PipelineText.substr(strlen("loop("));
if (!parseLoopPassPipeline(NestedLPM, PipelineText, VerifyEachPass,
DebugLogging) ||
PipelineText.empty())
return false;
assert(PipelineText[0] == ')');
PipelineText = PipelineText.substr(1);
// Add the nested pass manager with the appropriate adaptor.
LPM.addPass(std::move(NestedLPM));
} else {
// Otherwise try to parse a pass name.
size_t End = PipelineText.find_first_of(",)");
if (!parseLoopPassName(LPM, PipelineText.substr(0, End)))
return false;
// TODO: Ideally, we would run a LoopVerifierPass() here in the
// VerifyEachPass case, but we don't have such a verifier yet.
PipelineText = PipelineText.substr(End);
}
if (PipelineText.empty() || PipelineText[0] == ')')
return true;
assert(PipelineText[0] == ',');
PipelineText = PipelineText.substr(1);
}
}
// Append a #define line to Buf for Macro. Macro should be of the form XXX,
// in which case we emit "#define XXX 1" or "XXX=Y z W" in which case we emit
// "#define XXX Y z W". To get a #define with no value, use "XXX=".
static void DefineBuiltinMacro(MacroBuilder &Builder, StringRef Macro,
DiagnosticsEngine &Diags) {
std::pair<StringRef, StringRef> MacroPair = Macro.split('=');
StringRef MacroName = MacroPair.first;
StringRef MacroBody = MacroPair.second;
if (MacroName.size() != Macro.size()) {
// Per GCC -D semantics, the macro ends at \n if it exists.
StringRef::size_type End = MacroBody.find_first_of("\n\r");
if (End != StringRef::npos)
Diags.Report(diag::warn_fe_macro_contains_embedded_newline)
<< MacroName;
MacroBody = MacroBody.substr(0, End);
// We handle macro bodies which end in a backslash by appending an extra
// backslash+newline. This makes sure we don't accidentally treat the
// backslash as a line continuation marker.
if (MacroBodyEndsInBackslash(MacroBody))
Builder.defineMacro(MacroName, Twine(MacroBody) + "\\\n");
else
Builder.defineMacro(MacroName, MacroBody);
} else {
// Push "macroname 1".
Builder.defineMacro(Macro);
}
}
/// \brief After the file has been processed, check to see if we got all of
/// the expected diagnostics and check to see if there were any unexpected
/// ones.
bool DiagnosticVerifier::verifyFile(unsigned BufferID,
bool shouldAutoApplyFixes) {
using llvm::SMLoc;
const SourceLoc BufferStartLoc = SM.getLocForBufferStart(BufferID);
CharSourceRange EntireRange = SM.getRangeForBuffer(BufferID);
StringRef InputFile = SM.extractText(EntireRange);
StringRef BufferName = SM.getIdentifierForBuffer(BufferID);
// Queue up all of the diagnostics, allowing us to sort them and emit them in
// file order.
std::vector<llvm::SMDiagnostic> Errors;
unsigned PrevExpectedContinuationLine = 0;
std::vector<ExpectedDiagnosticInfo> ExpectedDiagnostics;
auto addError = [&](const char *Loc, std::string message,
ArrayRef<llvm::SMFixIt> FixIts = {}) {
auto loc = SourceLoc(SMLoc::getFromPointer(Loc));
auto diag = SM.GetMessage(loc, llvm::SourceMgr::DK_Error, message,
{}, FixIts);
Errors.push_back(diag);
};
// Scan the memory buffer looking for expected-note/warning/error.
for (size_t Match = InputFile.find("expected-");
Match != StringRef::npos; Match = InputFile.find("expected-", Match+1)) {
// Process this potential match. If we fail to process it, just move on to
// the next match.
StringRef MatchStart = InputFile.substr(Match);
const char *DiagnosticLoc = MatchStart.data();
llvm::SourceMgr::DiagKind ExpectedClassification;
if (MatchStart.startswith("expected-note")) {
ExpectedClassification = llvm::SourceMgr::DK_Note;
MatchStart = MatchStart.substr(strlen("expected-note"));
} else if (MatchStart.startswith("expected-warning")) {
ExpectedClassification = llvm::SourceMgr::DK_Warning;
MatchStart = MatchStart.substr(strlen("expected-warning"));
} else if (MatchStart.startswith("expected-error")) {
ExpectedClassification = llvm::SourceMgr::DK_Error;
MatchStart = MatchStart.substr(strlen("expected-error"));
} else
continue;
// Skip any whitespace before the {{.
MatchStart = MatchStart.substr(MatchStart.find_first_not_of(" \t"));
size_t TextStartIdx = MatchStart.find("{{");
if (TextStartIdx == StringRef::npos) {
addError(MatchStart.data(),
"expected {{ in expected-warning/note/error line");
continue;
}
int LineOffset = 0;
if (TextStartIdx > 0 && MatchStart[0] == '@') {
if (MatchStart[1] != '+' && MatchStart[1] != '-') {
addError(MatchStart.data(), "expected '+'/'-' for line offset");
continue;
}
StringRef Offs;
if (MatchStart[1] == '+')
Offs = MatchStart.slice(2, TextStartIdx).rtrim();
else
Offs = MatchStart.slice(1, TextStartIdx).rtrim();
size_t SpaceIndex = Offs.find(' ');
if (SpaceIndex != StringRef::npos && SpaceIndex < TextStartIdx) {
size_t Delta = Offs.size() - SpaceIndex;
MatchStart = MatchStart.substr(TextStartIdx - Delta);
TextStartIdx = Delta;
Offs = Offs.slice(0, SpaceIndex);
} else {
MatchStart = MatchStart.substr(TextStartIdx);
TextStartIdx = 0;
}
if (Offs.getAsInteger(10, LineOffset)) {
addError(MatchStart.data(), "expected line offset before '{{'");
continue;
}
}
ExpectedDiagnosticInfo Expected(DiagnosticLoc, ExpectedClassification);
unsigned Count = 1;
if (TextStartIdx > 0) {
StringRef CountStr = MatchStart.substr(0, TextStartIdx).trim();
if (CountStr == "*") {
Expected.mayAppear = true;
} else {
if (CountStr.getAsInteger(10, Count)) {
addError(MatchStart.data(), "expected match count before '{{'");
continue;
}
if (Count == 0) {
addError(MatchStart.data(),
"expected positive match count before '{{'");
continue;
}
}
// Resync up to the '{{'.
MatchStart = MatchStart.substr(TextStartIdx);
}
size_t End = MatchStart.find("}}");
if (End == StringRef::npos) {
addError(MatchStart.data(),
"didn't find '}}' to match '{{' in expected-warning/note/error line");
continue;
}
llvm::SmallString<256> Buf;
Expected.MessageRange = MatchStart.slice(2, End);
Expected.MessageStr =
Lexer::getEncodedStringSegment(Expected.MessageRange, Buf);
if (PrevExpectedContinuationLine)
Expected.LineNo = PrevExpectedContinuationLine;
else
Expected.LineNo = SM.getLineAndColumn(
BufferStartLoc.getAdvancedLoc(MatchStart.data() - InputFile.data()),
BufferID).first;
Expected.LineNo += LineOffset;
// Check if the next expected diagnostic should be in the same line.
StringRef AfterEnd = MatchStart.substr(End + strlen("}}"));
AfterEnd = AfterEnd.substr(AfterEnd.find_first_not_of(" \t"));
if (AfterEnd.startswith("\\"))
PrevExpectedContinuationLine = Expected.LineNo;
else
PrevExpectedContinuationLine = 0;
// Scan for fix-its: {{10-14=replacement text}}
StringRef ExtraChecks = MatchStart.substr(End+2).ltrim(" \t");
while (ExtraChecks.startswith("{{")) {
// First make sure we have a closing "}}".
size_t EndLoc = ExtraChecks.find("}}");
if (EndLoc == StringRef::npos) {
addError(ExtraChecks.data(),
"didn't find '}}' to match '{{' in fix-it verification");
break;
}
// Allow for close braces to appear in the replacement text.
while (EndLoc+2 < ExtraChecks.size() && ExtraChecks[EndLoc+2] == '}')
++EndLoc;
StringRef FixItStr = ExtraChecks.slice(2, EndLoc);
// Check for matching a later "}}" on a different line.
if (FixItStr.find_first_of("\r\n") != StringRef::npos) {
addError(ExtraChecks.data(), "didn't find '}}' to match '{{' in "
"fix-it verification");
break;
}
// Prepare for the next round of checks.
ExtraChecks = ExtraChecks.substr(EndLoc+2).ltrim();
// Special case for specifying no fixits should appear.
if (FixItStr == "none") {
Expected.noFixitsMayAppear = true;
continue;
}
// Parse the pieces of the fix-it.
size_t MinusLoc = FixItStr.find('-');
if (MinusLoc == StringRef::npos) {
addError(FixItStr.data(), "expected '-' in fix-it verification");
continue;
}
StringRef StartColStr = FixItStr.slice(0, MinusLoc);
StringRef AfterMinus = FixItStr.substr(MinusLoc+1);
size_t EqualLoc = AfterMinus.find('=');
if (EqualLoc == StringRef::npos) {
addError(AfterMinus.data(),
"expected '=' after '-' in fix-it verification");
continue;
}
StringRef EndColStr = AfterMinus.slice(0, EqualLoc);
StringRef AfterEqual = AfterMinus.substr(EqualLoc+1);
ExpectedFixIt FixIt;
FixIt.StartLoc = StartColStr.data()-2;
FixIt.EndLoc = FixItStr.data()+EndLoc;
if (StartColStr.getAsInteger(10, FixIt.StartCol)) {
addError(StartColStr.data(),
"invalid column number in fix-it verification");
continue;
}
if (EndColStr.getAsInteger(10, FixIt.EndCol)) {
addError(EndColStr.data(),
"invalid column number in fix-it verification");
continue;
}
// Translate literal "\\n" into '\n', inefficiently.
StringRef fixItText = AfterEqual.slice(0, EndLoc);
for (const char *current = fixItText.begin(), *end = fixItText.end();
current != end; /* in loop */) {
if (*current == '\\' && current + 1 < end) {
if (current[1] == 'n') {
FixIt.Text += '\n';
current += 2;
} else { // Handle \}, \\, etc.
FixIt.Text += current[1];
current += 2;
}
} else {
FixIt.Text += *current++;
}
}
Expected.Fixits.push_back(FixIt);
}
Expected.ExpectedEnd = ExtraChecks.data();
// Don't include trailing whitespace in the expected-foo{{}} range.
while (isspace(Expected.ExpectedEnd[-1]))
--Expected.ExpectedEnd;
// Add the diagnostic the expected number of times.
for (; Count; --Count)
ExpectedDiagnostics.push_back(Expected);
}
// Make sure all the expected diagnostics appeared.
std::reverse(ExpectedDiagnostics.begin(), ExpectedDiagnostics.end());
for (unsigned i = ExpectedDiagnostics.size(); i != 0; ) {
--i;
auto &expected = ExpectedDiagnostics[i];
// Check to see if we had this expected diagnostic.
auto FoundDiagnosticIter = findDiagnostic(expected, BufferName);
if (FoundDiagnosticIter == CapturedDiagnostics.end()) {
// Diagnostic didn't exist. If this is a 'mayAppear' diagnostic, then
// we're ok. Otherwise, leave it in the list.
if (expected.mayAppear)
ExpectedDiagnostics.erase(ExpectedDiagnostics.begin()+i);
continue;
}
auto &FoundDiagnostic = *FoundDiagnosticIter;
const char *IncorrectFixit = nullptr;
// Verify that any expected fix-its are present in the diagnostic.
for (auto fixit : expected.Fixits) {
// If we found it, we're ok.
if (!checkForFixIt(fixit, FoundDiagnostic, InputFile))
IncorrectFixit = fixit.StartLoc;
}
// If we have any expected fixits that didn't get matched, then they are
// wrong. Replace the failed fixit with what actually happened.
if (IncorrectFixit) {
if (FoundDiagnostic.getFixIts().empty()) {
addError(IncorrectFixit, "expected fix-it not seen");
continue;
}
// If we had an incorrect expected fixit, render it and produce a fixit
// of our own.
auto actual = renderFixits(FoundDiagnostic.getFixIts(), InputFile);
auto replStartLoc = SMLoc::getFromPointer(expected.Fixits[0].StartLoc);
auto replEndLoc = SMLoc::getFromPointer(expected.Fixits.back().EndLoc);
llvm::SMFixIt fix(llvm::SMRange(replStartLoc, replEndLoc), actual);
addError(IncorrectFixit,
"expected fix-it not seen; actual fix-its: " + actual, fix);
} else if (expected.noFixitsMayAppear &&
!FoundDiagnostic.getFixIts().empty() &&
!expected.mayAppear) {
// If there was no fixit specification, but some were produced, add a
// fixit to add them in.
auto actual = renderFixits(FoundDiagnostic.getFixIts(), InputFile);
auto replStartLoc = SMLoc::getFromPointer(expected.ExpectedEnd - 8); // {{none}} length
auto replEndLoc = SMLoc::getFromPointer(expected.ExpectedEnd - 1);
llvm::SMFixIt fix(llvm::SMRange(replStartLoc, replEndLoc), actual);
addError(replStartLoc.getPointer(), "expected no fix-its; actual fix-it seen: " + actual, fix);
}
// Actually remove the diagnostic from the list, so we don't match it
// again. We do have to do this after checking fix-its, though, because
// the diagnostic owns its fix-its.
CapturedDiagnostics.erase(FoundDiagnosticIter);
// We found the diagnostic, so remove it... unless we allow an arbitrary
// number of diagnostics, in which case we want to reprocess this.
if (expected.mayAppear)
++i;
else
ExpectedDiagnostics.erase(ExpectedDiagnostics.begin()+i);
}
// Check to see if we have any incorrect diagnostics. If so, diagnose them as
// such.
for (unsigned i = ExpectedDiagnostics.size(); i != 0; ) {
--i;
auto &expected = ExpectedDiagnostics[i];
// Check to see if any found diagnostics have the right line and
// classification, but the wrong text.
auto I = CapturedDiagnostics.begin();
for (auto E = CapturedDiagnostics.end(); I != E; ++I) {
// Verify the file and line of the diagnostic.
if (I->getLineNo() != (int)expected.LineNo ||
I->getFilename() != BufferName ||
I->getKind() != expected.Classification)
continue;
// Otherwise, we found it, break out.
break;
}
if (I == CapturedDiagnostics.end()) continue;
auto StartLoc = SMLoc::getFromPointer(expected.MessageRange.begin());
auto EndLoc = SMLoc::getFromPointer(expected.MessageRange.end());
llvm::SMFixIt fixIt(llvm::SMRange{ StartLoc, EndLoc }, I->getMessage());
addError(expected.MessageRange.begin(), "incorrect message found", fixIt);
CapturedDiagnostics.erase(I);
ExpectedDiagnostics.erase(ExpectedDiagnostics.begin()+i);
}
// Diagnose expected diagnostics that didn't appear.
std::reverse(ExpectedDiagnostics.begin(), ExpectedDiagnostics.end());
for (auto const &expected : ExpectedDiagnostics) {
std::string message = "expected "+getDiagKindString(expected.Classification)
+ " not produced";
// Get the range of the expected-foo{{}} diagnostic specifier.
auto StartLoc = expected.ExpectedStart;
auto EndLoc = expected.ExpectedEnd;
// A very common case if for the specifier to be the last thing on the line.
// In this case, eat any trailing whitespace.
while (isspace(*EndLoc) && *EndLoc != '\n' && *EndLoc != '\r')
++EndLoc;
// If we found the end of the line, we can do great things. Otherwise,
// avoid nuking whitespace that might be zapped through other means.
if (*EndLoc != '\n' && *EndLoc != '\r') {
EndLoc = expected.ExpectedEnd;
} else {
// If we hit the end of line, then zap whitespace leading up to it.
auto FileStart = InputFile.data();
while (StartLoc-1 != FileStart && isspace(StartLoc[-1]) &&
StartLoc[-1] != '\n' && StartLoc[-1] != '\r')
--StartLoc;
// If we got to the end of the line, and the thing before this diagnostic
// is a "//" then we can remove it too.
if (StartLoc-2 >= FileStart && StartLoc[-1] == '/' && StartLoc[-2] == '/')
StartLoc -= 2;
// Perform another round of general whitespace nuking to cleanup
// whitespace before the //.
while (StartLoc-1 != FileStart && isspace(StartLoc[-1]) &&
StartLoc[-1] != '\n' && StartLoc[-1] != '\r')
--StartLoc;
// If we found a \n, then we can nuke the entire line.
if (StartLoc-1 != FileStart &&
(StartLoc[-1] == '\n' || StartLoc[-1] == '\r'))
--StartLoc;
}
// Remove the expected-foo{{}} as a fixit.
llvm::SMFixIt fixIt(llvm::SMRange{
SMLoc::getFromPointer(StartLoc),
SMLoc::getFromPointer(EndLoc)
}, "");
addError(expected.ExpectedStart, message, fixIt);
}
// Verify that there are no diagnostics (in MemoryBuffer) left in the list.
for (unsigned i = 0, e = CapturedDiagnostics.size(); i != e; ++i) {
if (CapturedDiagnostics[i].getFilename() != BufferName)
continue;
std::string Message =
"unexpected "+getDiagKindString(CapturedDiagnostics[i].getKind())+
" produced: "+CapturedDiagnostics[i].getMessage().str();
addError(CapturedDiagnostics[i].getLoc().getPointer(),
Message);
}
// Sort the diagnostics by their address in the memory buffer as the primary
// key. This ensures that an "unexpected diagnostic" and
// "expected diagnostic" in the same place are emitted next to each other.
std::sort(Errors.begin(), Errors.end(),
[&](const llvm::SMDiagnostic &lhs,
const llvm::SMDiagnostic &rhs) -> bool {
return lhs.getLoc().getPointer() < rhs.getLoc().getPointer();
});
// Emit all of the queue'd up errors.
for (auto Err : Errors)
SM.getLLVMSourceMgr().PrintMessage(llvm::errs(), Err);
// If auto-apply fixits is on, rewrite the original source file.
if (shouldAutoApplyFixes)
autoApplyFixes(BufferID, Errors);
return !Errors.empty();
}
