ccls/src/indexer.cc

868 lines
28 KiB
C++
Raw Normal View History

2017-02-22 08:52:00 +00:00
#include "indexer.h"
2017-02-17 09:57:44 +00:00
2018-05-27 19:24:56 +00:00
#include "log.hh"
2017-04-08 22:54:36 +00:00
#include "platform.h"
#include "serializer.h"
using ccls::Intern;
#include <clang/AST/AST.h>
#include <clang/Frontend/ASTUnit.h>
#include <clang/Frontend/CompilerInstance.h>
#include <clang/Frontend/FrontendAction.h>
#include <clang/Index/IndexDataConsumer.h>
#include <clang/Index/IndexingAction.h>
#include <clang/Index/USRGeneration.h>
#include <clang/Lex/PreprocessorOptions.h>
#include <llvm/ADT/DenseSet.h>
#include <llvm/Support/Timer.h>
#include <llvm/Support/CrashRecoveryContext.h>
using namespace clang;
using llvm::Timer;
#include <assert.h>
#include <inttypes.h>
2018-03-31 20:59:27 +00:00
#include <limits.h>
2017-04-22 07:39:55 +00:00
#include <algorithm>
#include <chrono>
#include <unordered_set>
2017-04-22 07:39:55 +00:00
namespace {
struct IndexParam {
llvm::DenseSet<unsigned> SeenUID;
std::vector<std::string> seen_files;
std::unordered_map<std::string, FileContents> file_contents;
std::unordered_map<std::string, int64_t> file2write_time;
2017-04-12 07:57:12 +00:00
// Only use this when strictly needed (ie, primary translation unit is
// needed). Most logic should get the IndexFile instance via
// |file_consumer|.
//
// This can be null if we're not generating an index for the primary
// translation unit.
IndexFile* primary_file = nullptr;
ASTUnit& Unit;
FileConsumer* file_consumer = nullptr;
NamespaceHelper ns;
IndexParam(ASTUnit& Unit, FileConsumer* file_consumer)
: Unit(Unit), file_consumer(file_consumer) {}
};
IndexFile *ConsumeFile(IndexParam &param, const FileEntry &File) {
IndexFile *db =
param.file_consumer->TryConsumeFile(File, &param.file_contents);
// If this is the first time we have seen the file (ignoring if we are
// generating an index for it):
if (param.SeenUID.insert(File.getUID()).second) {
std::string file_name = FileName(File);
// Add to all files we have seen so we can generate proper dependency
// graph.
param.seen_files.push_back(file_name);
// Set modification time.
std::optional<int64_t> write_time = LastWriteTime(file_name);
LOG_IF_S(ERROR, !write_time)
<< "failed to fetch write time for " << file_name;
if (write_time)
param.file2write_time[file_name] = *write_time;
2018-01-07 09:07:39 +00:00
}
return db;
2018-01-07 09:07:39 +00:00
}
Range FromSourceRange(const SourceManager &SM, const LangOptions &LangOpts,
SourceRange R, unsigned *UID, bool token) {
SourceLocation BLoc = R.getBegin(), ELoc = R.getEnd();
std::pair<FileID, unsigned> BInfo = SM.getDecomposedLoc(BLoc);
std::pair<FileID, unsigned> EInfo = SM.getDecomposedLoc(ELoc);
if (token)
EInfo.second += Lexer::MeasureTokenLength(ELoc, SM, LangOpts);
unsigned l0 = SM.getLineNumber(BInfo.first, BInfo.second) - 1,
c0 = SM.getColumnNumber(BInfo.first, BInfo.second) - 1,
l1 = SM.getLineNumber(EInfo.first, EInfo.second) - 1,
c1 = SM.getColumnNumber(EInfo.first, EInfo.second) - 1;
if (l0 > INT16_MAX) l0 = 0;
if (c0 > INT16_MAX) c0 = 0;
if (l1 > INT16_MAX) l1 = 0;
if (c1 > INT16_MAX) c1 = 0;
if (UID) {
if (const FileEntry *F = SM.getFileEntryForID(BInfo.first))
*UID = F->getUID();
else
*UID = 0;
}
return {{int16_t(l0), int16_t(c0)}, {int16_t(l1), int16_t(c1)}};
}
Range FromCharRange(const SourceManager &SM, const LangOptions &LangOpts,
SourceRange R) {
return FromSourceRange(SM, LangOpts, R, nullptr, false);
}
Range FromTokenRange(const SourceManager &SM, const LangOptions &LangOpts,
SourceRange R, unsigned *UID = nullptr) {
return FromSourceRange(SM, LangOpts, R, UID, true);
}
Position ResolveSourceLocation(const SourceManager &SM, SourceLocation Loc,
FileID *F = nullptr) {
std::pair<FileID, unsigned> D = SM.getDecomposedLoc(Loc);
if (F)
*F = D.first;
unsigned line = SM.getLineNumber(D.first, D.second) - 1;
unsigned col = SM.getColumnNumber(D.first, D.second) - 1;
return {int16_t(line > INT16_MAX ? 0 : line),
int16_t(col > INT16_MAX ? 0 : col)};
}
SymbolKind GetSymbolKind(const Decl& D) {
switch (D.getKind()) {
case Decl::TranslationUnit:
return SymbolKind::File;
case Decl::FunctionTemplate:
case Decl::Function:
case Decl::CXXMethod:
case Decl::CXXConstructor:
case Decl::CXXConversion:
case Decl::CXXDestructor:
return SymbolKind::Func;
case Decl::Namespace:
case Decl::NamespaceAlias:
case Decl::ClassTemplate:
case Decl::TypeAliasTemplate:
case Decl::Enum:
case Decl::Record:
case Decl::CXXRecord:
case Decl::TypeAlias:
case Decl::Typedef:
case Decl::UnresolvedUsingTypename:
return SymbolKind::Type;
case Decl::Field:
case Decl::Var:
case Decl::ParmVar:
case Decl::ImplicitParam:
case Decl::Decomposition:
case Decl::EnumConstant:
return SymbolKind::Var;
default:
return SymbolKind::Invalid;
}
}
const Decl* GetTypeDecl(QualType T) {
Decl *D = nullptr;
const Type *TP;
for(;;) {
TP = T.getTypePtrOrNull();
if (!TP)
return D;
switch (TP->getTypeClass()) {
case Type::Pointer:
T = cast<PointerType>(TP)->getPointeeType();
continue;
case Type::BlockPointer:
T = cast<BlockPointerType>(TP)->getPointeeType();
continue;
case Type::LValueReference:
case Type::RValueReference:
T = cast<ReferenceType>(TP)->getPointeeType();
continue;
case Type::ObjCObjectPointer:
T = cast<ObjCObjectPointerType>(TP)->getPointeeType();
continue;
case Type::MemberPointer:
T = cast<MemberPointerType>(TP)->getPointeeType();
continue;
default:
break;
}
break;
}
try_again:
switch (TP->getTypeClass()) {
case Type::Typedef:
D = cast<TypedefType>(TP)->getDecl();
break;
case Type::ObjCObject:
D = cast<ObjCObjectType>(TP)->getInterface();
break;
case Type::ObjCInterface:
D = cast<ObjCInterfaceType>(TP)->getDecl();
break;
case Type::Record:
case Type::Enum:
D = cast<TagType>(TP)->getDecl();
break;
case Type::TemplateSpecialization:
if (const RecordType *Record = TP->getAs<RecordType>())
D = Record->getDecl();
else
D = cast<TemplateSpecializationType>(TP)
->getTemplateName()
.getAsTemplateDecl();
break;
case Type::Auto:
case Type::DeducedTemplateSpecialization:
TP = cast<DeducedType>(TP)->getDeducedType().getTypePtrOrNull();
if (TP)
goto try_again;
break;
case Type::InjectedClassName:
D = cast<InjectedClassNameType>(TP)->getDecl();
break;
// FIXME: Template type parameters!
case Type::Elaborated:
TP = cast<ElaboratedType>(TP)->getNamedType().getTypePtrOrNull();
goto try_again;
default:
break;
}
return D;
}
const Decl* GetSpecialized(const Decl* D) {
if (!D)
return D;
Decl *Template = nullptr;
if (const CXXRecordDecl *CXXRecord = dyn_cast<CXXRecordDecl>(D)) {
if (const ClassTemplatePartialSpecializationDecl *PartialSpec
= dyn_cast<ClassTemplatePartialSpecializationDecl>(CXXRecord))
Template = PartialSpec->getSpecializedTemplate();
else if (const ClassTemplateSpecializationDecl *ClassSpec
= dyn_cast<ClassTemplateSpecializationDecl>(CXXRecord)) {
llvm::PointerUnion<ClassTemplateDecl *,
ClassTemplatePartialSpecializationDecl *> Result
= ClassSpec->getSpecializedTemplateOrPartial();
if (Result.is<ClassTemplateDecl *>())
Template = Result.get<ClassTemplateDecl *>();
else
Template = Result.get<ClassTemplatePartialSpecializationDecl *>();
} else
Template = CXXRecord->getInstantiatedFromMemberClass();
} else if (const FunctionDecl *Function = dyn_cast<FunctionDecl>(D)) {
Template = Function->getPrimaryTemplate();
if (!Template)
Template = Function->getInstantiatedFromMemberFunction();
} else if (const VarDecl *Var = dyn_cast<VarDecl>(D)) {
if (Var->isStaticDataMember())
Template = Var->getInstantiatedFromStaticDataMember();
} else if (const RedeclarableTemplateDecl *Tmpl
= dyn_cast<RedeclarableTemplateDecl>(D))
Template = Tmpl->getInstantiatedFromMemberTemplate();
else
return nullptr;
return Template;
}
class IndexDataConsumer : public index::IndexDataConsumer {
ASTContext *Ctx;
IndexParam& param;
llvm::DenseMap<const Decl*, Usr> Decl2usr;
2018-07-07 22:25:25 +00:00
std::string GetComment(const Decl* D) {
SourceManager &SM = Ctx->getSourceManager();
const RawComment *RC = Ctx->getRawCommentForAnyRedecl(D);
if (!RC) return "";
StringRef Raw = RC->getRawText(Ctx->getSourceManager());
SourceRange R = RC->getSourceRange();
std::pair<FileID, unsigned> BInfo = SM.getDecomposedLoc(R.getBegin());
unsigned start_column = SM.getLineNumber(BInfo.first, BInfo.second);
std::string ret;
int pad = -1;
for (const char *p = Raw.data(), *E = Raw.end(); p < E;) {
// The first line starts with a comment marker, but the rest needs
// un-indenting.
unsigned skip = start_column - 1;
for (; skip > 0 && p < E && (*p == ' ' || *p == '\t'); p++)
skip--;
const char *q = p;
while (q < E && *q != '\n')
q++;
if (q < E)
q++;
// A minimalist approach to skip Doxygen comment markers.
// See https://www.stack.nl/~dimitri/doxygen/manual/docblocks.html
if (pad < 0) {
// First line, detect the length of comment marker and put into |pad|
const char *begin = p;
while (p < E && (*p == '/' || *p == '*'))
p++;
if (p < E && (*p == '<' || *p == '!'))
p++;
if (p < E && *p == ' ')
p++;
pad = int(p - begin);
} else {
// Other lines, skip |pad| bytes
int prefix = pad;
while (prefix > 0 && p < E &&
(*p == ' ' || *p == '/' || *p == '*' || *p == '<' || *p == '!'))
prefix--, p++;
}
ret.insert(ret.end(), p, q);
p = q;
}
while (ret.size() && isspace(ret.back()))
ret.pop_back();
if (EndsWith(ret, "*/")) {
ret.resize(ret.size() - 2);
} else if (EndsWith(ret, "\n/")) {
ret.resize(ret.size() - 2);
}
while (ret.size() && isspace(ret.back()))
ret.pop_back();
return ret;
}
Usr GetUsr(const Decl* D) {
D = D->getCanonicalDecl();
auto R = Decl2usr.try_emplace(D);
if (R.second) {
SmallString<256> USR;
index::generateUSRForDecl(D, USR);
R.first->second = HashUsr({USR.data(), USR.size()});
}
return R.first->second;
}
template <typename Def>
void SetName(const Decl *D, std::string_view short_name,
std::string_view qualified, Def &def,
PrintingPolicy *Policy = nullptr) {
SmallString<256> Str;
llvm::raw_svector_ostream OS(Str);
if (Policy) {
D->print(OS, *Policy);
} else {
PrintingPolicy PP(Ctx->getLangOpts());
PP.AnonymousTagLocations = false;
PP.TerseOutput = true;
// PP.PolishForDeclaration = true;
PP.ConstantsAsWritten = true;
PP.SuppressTagKeyword = false;
PP.FullyQualifiedName = false;
D->print(OS, PP);
}
std::string name = OS.str();
for (std::string::size_type i = 0;;) {
if ((i = name.find("(anonymous ", i)) == std::string::npos)
break;
i++;
if (name.size() > 10 + 9 && name.compare(10, 9, "namespace"))
name.replace(i, 10 + 9, "anon ns");
else
name.replace(i, 10, "anon");
}
auto i = name.find(short_name);
if (i == std::string::npos) {
// e.g. operator type-parameter-1
i = 0;
def.short_name_offset = 0;
} else if (short_name.size()) {
name.replace(i, short_name.size(), qualified);
def.short_name_offset = i + qualified.size() - short_name.size();
} else {
def.short_name_offset = i;
}
def.short_name_size = short_name.size();
for (int paren = 0; i; i--) {
// Skip parentheses in "(anon struct)::name"
if (name[i - 1] == ')')
paren++;
else if (name[i - 1] == '(')
paren--;
else if (!(paren > 0 || isalnum(name[i - 1]) ||
name[i - 1] == '_' || name[i - 1] == ':'))
break;
}
def.qual_name_offset = i;
def.detailed_name = Intern(name);
2017-07-30 04:24:02 +00:00
}
Use GetUse(IndexFile* db, Range range, const DeclContext *DC, Role role) {
if (!DC)
return Use{{range, 0, SymbolKind::File, role}};
const Decl *D = cast<Decl>(DC);
switch (GetSymbolKind(*D)) {
case SymbolKind::Func:
return Use{{range, db->ToFunc(GetUsr(D)).usr, SymbolKind::Func, role}};
case SymbolKind::Type:
return Use{{range, db->ToType(GetUsr(D)).usr, SymbolKind::Type, role}};
case SymbolKind::Var:
return Use{{range, db->ToVar(GetUsr(D)).usr, SymbolKind::Var, role}};
2018-02-18 18:07:13 +00:00
default:
return Use{{range, 0, SymbolKind::File, role}};
}
}
public:
IndexDataConsumer(IndexParam& param) : param(param) {}
void initialize(ASTContext &Ctx) override {
this->Ctx = &Ctx;
}
bool handleDeclOccurence(const Decl *D, index::SymbolRoleSet Roles,
ArrayRef<index::SymbolRelation> Relations,
#if LLVM_VERSION_MAJOR >= 7
SourceLocation Loc,
#else
FileID LocFID, unsigned LocOffset,
#endif
ASTNodeInfo ASTNode) override {
SourceManager &SM = Ctx->getSourceManager();
const LangOptions &Lang = Ctx->getLangOpts();
#if LLVM_VERSION_MAJOR < 7
SourceLocation Loc;
{
const SrcMgr::SLocEntry &Entry = SM.getSLocEntry(LocFID);
unsigned off = Entry.getOffset() + LocOffset;
if (!Entry.isFile())
off |= 1u << 31;
Loc = SourceLocation::getFromRawEncoding(off);
}
#endif
SourceLocation Spell = SM.getSpellingLoc(Loc);
Range spell = FromTokenRange(SM, Ctx->getLangOpts(), SourceRange(Spell, Spell));
const FileEntry *FE = SM.getFileEntryForID(SM.getFileID(Spell));
if (!FE) {
#if LLVM_VERSION_MAJOR < 7
auto P = SM.getExpansionRange(Loc);
spell = FromCharRange(SM, Ctx->getLangOpts(), SourceRange(P.first, P.second));
FE = SM.getFileEntryForID(SM.getFileID(P.first));
#else
auto R = SM.getExpansionRange(Loc);
spell = FromTokenRange(SM, Ctx->getLangOpts(), R.getAsRange());
FE = SM.getFileEntryForID(SM.getFileID(R.getBegin()));
#endif
if (!FE)
return true;
}
IndexFile *db = ConsumeFile(param, *FE);
if (!db)
return true;
const DeclContext *SemDC = D->getDeclContext();
const DeclContext *LexDC = D->getLexicalDeclContext();
(void)SemDC;
(void)LexDC;
Range extent = FromTokenRange(SM, Lang, D->getSourceRange());
Role role = static_cast<Role>(Roles);
bool is_decl = Roles & uint32_t(index::SymbolRole::Declaration);
bool is_def = Roles & uint32_t(index::SymbolRole::Definition);
std::string short_name, qualified;
if (auto* ND = dyn_cast<NamedDecl>(D)) {
short_name = ND->getNameAsString();
qualified = ND->getQualifiedNameAsString();
}
IndexFunc *func = nullptr;
IndexType *type = nullptr;
IndexVar *var = nullptr;
SymbolKind kind = GetSymbolKind(*D);
Usr usr = GetUsr(D);
switch (kind) {
case SymbolKind::Invalid:
LOG_S(INFO) << "Unhandled " << int(D->getKind());
return true;
case SymbolKind::File:
return true;
case SymbolKind::Func:
func = &db->ToFunc(usr);
if (!func->def.detailed_name[0]) {
SetName(D, short_name, qualified, func->def);
if (g_config->index.comments)
2018-07-07 22:25:25 +00:00
func->def.comments = Intern(GetComment(D));
}
if (is_def || (is_decl && !func->def.spell)) {
if (func->def.spell)
func->declarations.push_back(*func->def.spell);
func->def.spell = GetUse(db, spell, LexDC, role);
func->def.extent = GetUse(db, extent, LexDC, Role::None);
if (auto *FD = dyn_cast<FunctionDecl>(D)) {
DeclarationNameInfo Info = FD->getNameInfo();
func->def.spell = GetUse(
db, FromTokenRange(SM, Ctx->getLangOpts(), Info.getSourceRange()),
LexDC, role);
}
} else
func->uses.push_back(GetUse(db, spell, LexDC, role));
break;
case SymbolKind::Type:
type = &db->ToType(usr);
if (!type->def.detailed_name[0]) {
SetName(D, short_name, qualified, type->def);
if (g_config->index.comments)
2018-07-07 22:25:25 +00:00
type->def.comments = Intern(GetComment(D));
}
if (is_def || (is_decl && !type->def.spell)) {
if (type->def.spell)
type->declarations.push_back(*type->def.spell);
type->def.spell = GetUse(db, spell, LexDC, role);
type->def.extent = GetUse(db, extent, LexDC, Role::None);
} else
type->uses.push_back(GetUse(db, spell, LexDC, role));
break;
case SymbolKind::Var:
var = &db->ToVar(usr);
if (!var->def.detailed_name[0]) {
SetName(D, short_name, qualified, var->def);
if (g_config->index.comments)
2018-07-07 22:25:25 +00:00
var->def.comments = Intern(GetComment(D));
}
if (is_def || (is_decl && !var->def.spell)) {
if (var->def.spell)
var->declarations.push_back(*var->def.spell);
var->def.spell = GetUse(db, spell, LexDC, role);
var->def.extent = GetUse(db, extent, LexDC, Role::None);
if (auto *VD = dyn_cast<VarDecl>(D)) {
var->def.storage = VD->getStorageClass();
QualType T = VD->getType();
for (const Decl* D1 = GetTypeDecl(T); D1; D1 = GetSpecialized(D1)) {
Usr usr1 = GetUsr(D1);
if (db->usr2type.count(usr1))
var->def.type = usr1;
}
}
} else
var->uses.push_back(GetUse(db, spell, LexDC, role));
break;
}
switch (D->getKind()) {
case Decl::Namespace:
type->def.kind = lsSymbolKind::Namespace;
break;
case Decl::NamespaceAlias: {
type->def.kind = lsSymbolKind::TypeAlias;
auto* NAD = cast<NamespaceAliasDecl>(D);
if (const NamespaceDecl* ND = NAD->getNamespace()) {
Usr usr1 = GetUsr(ND);
if (db->usr2type.count(usr1))
type->def.alias_of = usr1;
}
break;
}
case Decl::Enum:
type->def.kind = lsSymbolKind::Enum;
break;
case Decl::Record:
case Decl::CXXRecord:
type->def.kind = lsSymbolKind::Struct;
if (is_def) {
auto *RD = cast<RecordDecl>(D);
bool can_get_offset =
RD->isCompleteDefinition() && !RD->isDependentType();
for (FieldDecl *FD : RD->fields())
type->def.vars.emplace_back(
GetUsr(FD), can_get_offset ? Ctx->getFieldOffset(FD) : -1);
}
break;
case Decl::ClassTemplate:
type->def.kind = lsSymbolKind::Class;
break;
case Decl::FunctionTemplate:
type->def.kind = lsSymbolKind::Function;
break;
case Decl::TypeAliasTemplate:
type->def.kind = lsSymbolKind::TypeAlias;
break;
case Decl::TypeAlias:
case Decl::Typedef:
case Decl::UnresolvedUsingTypename:
type->def.kind = lsSymbolKind::TypeAlias;
if (auto *TD = dyn_cast<TypedefNameDecl>(D)) {
QualType T = TD->getUnderlyingType();
if (const Decl* D1 = GetTypeDecl(T)) {
Usr usr1 = GetUsr(D1);
if (db->usr2type.count(usr1))
type->def.alias_of = usr1;
}
}
break;
case Decl::Function:
func->def.kind = lsSymbolKind::Function;
break;
case Decl::CXXMethod:
func->def.kind = lsSymbolKind::Method;
break;
case Decl::CXXConstructor:
case Decl::CXXConversion:
func->def.kind = lsSymbolKind::Constructor;
break;
case Decl::CXXDestructor:
func->def.kind = lsSymbolKind::Method;
break;
case Decl::Var:
case Decl::ParmVar:
var->def.kind = lsSymbolKind::Variable;
if (is_def) {
if (auto* FD = dyn_cast<FunctionDecl>(SemDC))
db->ToFunc(GetUsr(FD)).def.vars.push_back(usr);
else if (auto* ND = dyn_cast<NamespaceDecl>(SemDC))
db->ToType(GetUsr(ND)).def.vars.emplace_back(usr, -1);
}
[[fallthrough]];
case Decl::Field:
case Decl::ImplicitParam:
case Decl::Decomposition:
break;
case Decl::EnumConstant: {
auto *ECD = cast<EnumConstantDecl>(D);
const auto& Val = ECD->getInitVal();
std::string init =
" = " + (Val.isSigned() ? std::to_string(Val.getSExtValue())
: std::to_string(Val.getZExtValue()));
var->def.detailed_name = Intern(var->def.detailed_name + init);
break;
}
default:
LOG_S(INFO) << "Unhandled " << int(D->getKind());
break;
}
return true;
}
};
}
const int IndexFile::kMajorVersion = 16;
const int IndexFile::kMinorVersion = 1;
IndexFile::IndexFile(unsigned UID, const std::string &path,
const std::string &contents)
: UID(UID), path(path), file_contents(contents) {}
2017-02-20 00:56:56 +00:00
IndexFunc& IndexFile::ToFunc(Usr usr) {
auto ret = usr2func.try_emplace(usr);
if (ret.second)
ret.first->second.usr = usr;
return ret.first->second;
2017-02-20 00:56:56 +00:00
}
IndexType& IndexFile::ToType(Usr usr) {
auto ret = usr2type.try_emplace(usr);
if (ret.second)
ret.first->second.usr = usr;
return ret.first->second;
2017-02-20 00:56:56 +00:00
}
IndexVar& IndexFile::ToVar(Usr usr) {
auto ret = usr2var.try_emplace(usr);
if (ret.second)
ret.first->second.usr = usr;
return ret.first->second;
2017-02-17 09:57:44 +00:00
}
std::string IndexFile::ToString() {
return ccls::Serialize(SerializeFormat::Json, *this);
2017-02-17 09:57:44 +00:00
}
2018-05-05 03:40:52 +00:00
void Uniquify(std::vector<Usr>& usrs) {
std::unordered_set<Usr> seen;
size_t n = 0;
2018-05-05 03:40:52 +00:00
for (size_t i = 0; i < usrs.size(); i++)
if (seen.insert(usrs[i]).second)
usrs[n++] = usrs[i];
usrs.resize(n);
}
void Uniquify(std::vector<Use>& uses) {
2018-05-05 03:40:52 +00:00
std::unordered_set<Range> seen;
size_t n = 0;
for (size_t i = 0; i < uses.size(); i++) {
2018-05-05 03:40:52 +00:00
if (seen.insert(uses[i].range).second)
uses[n++] = uses[i];
}
uses.resize(n);
2017-02-24 08:39:25 +00:00
}
2017-02-20 06:40:55 +00:00
2018-05-09 05:01:58 +00:00
std::vector<std::unique_ptr<IndexFile>> ClangIndexer::Index(
VFS* vfs,
std::string file,
2017-09-27 06:03:43 +00:00
const std::vector<std::string>& args,
2018-06-01 04:21:34 +00:00
const std::vector<FileContents>& file_contents) {
2018-04-04 06:05:41 +00:00
if (!g_config->index.enabled)
2018-04-02 07:22:12 +00:00
return {};
2017-04-19 00:05:14 +00:00
file = NormalizePath(file);
2017-03-11 02:24:51 +00:00
std::vector<const char *> Args;
for (auto& arg: args)
Args.push_back(arg.c_str());
Args.push_back("-Xclang");
Args.push_back("-detailed-preprocessing-record");
Args.push_back("-fno-spell-checking");
auto PCHCO = std::make_shared<PCHContainerOperations>();
IntrusiveRefCntPtr<DiagnosticsEngine>
Diags(CompilerInstance::createDiagnostics(new DiagnosticOptions));
std::shared_ptr<CompilerInvocation> CI =
createInvocationFromCommandLine(Args, Diags);
if (!CI)
return {};
CI->getLangOpts()->CommentOpts.ParseAllComments = true;
std::vector<std::unique_ptr<llvm::MemoryBuffer>> BufOwner;
for (auto &c : file_contents) {
std::unique_ptr<llvm::MemoryBuffer> MB =
llvm::MemoryBuffer::getMemBufferCopy(c.content, c.path);
CI->getPreprocessorOpts().addRemappedFile(c.path, MB.get());
BufOwner.push_back(std::move(MB));
}
2017-07-30 04:24:02 +00:00
auto Unit = ASTUnit::create(CI, Diags, true, true);
if (!Unit)
2018-04-02 07:22:12 +00:00
return {};
2017-02-16 09:35:30 +00:00
FileConsumer file_consumer(vfs, file);
IndexParam param(*Unit, &file_consumer);
auto DataConsumer = std::make_shared<IndexDataConsumer>(param);
index::IndexingOptions IndexOpts;
memset(&IndexOpts, 1, sizeof IndexOpts);
IndexOpts.SystemSymbolFilter =
index::IndexingOptions::SystemSymbolFilterKind::All;
IndexOpts.IndexFunctionLocals = true;
std::unique_ptr<FrontendAction> IndexAction =
createIndexingAction(DataConsumer, IndexOpts, nullptr);
llvm::CrashRecoveryContextCleanupRegistrar<FrontendAction> IndexActionCleanup(
IndexAction.get());
DiagnosticErrorTrap DiagTrap(*Diags);
bool Success = ASTUnit::LoadFromCompilerInvocationAction(
std::move(CI), PCHCO, Diags, IndexAction.get(), Unit.get(),
/*Persistent=*/true, "/home/maskray/Dev/llvm/release/lib/clang/7.0.0",
/*OnlyLocalDecls=*/true,
/*CaptureDiagnostics=*/true, 0, false, false, true);
if (!Unit) {
LOG_S(ERROR) << "failed to index " << file;
2018-04-02 07:22:12 +00:00
return {};
2017-12-01 17:57:03 +00:00
}
if (!Success)
return {};
// ClangCursor(clang_getTranslationUnitCursor(tu->cx_tu))
// .VisitChildren(&VisitMacroDefinitionAndExpansions, &param);
2018-01-06 17:28:55 +00:00
std::unordered_map<std::string, int> inc_to_line;
// TODO
if (param.primary_file)
2018-01-11 02:43:01 +00:00
for (auto& inc : param.primary_file->includes)
inc_to_line[inc.resolved_path] = inc.line;
2018-01-06 17:28:55 +00:00
llvm::DenseMap<unsigned, std::vector<Range>> UID2skipped;
{
const SourceManager& SM = Unit->getSourceManager();
PreprocessingRecord *PPRec = Unit->getPreprocessor().getPreprocessingRecord();
if (PPRec) {
const std::vector<SourceRange>& Skipped = PPRec->getSkippedRanges();
for (auto& R : Skipped) {
unsigned UID;
Range range = FromTokenRange(SM, Unit->getLangOpts(), R, &UID);
UID2skipped[UID].push_back(range);
}
}
}
2017-04-08 22:54:36 +00:00
auto result = param.file_consumer->TakeLocalState();
2017-07-30 04:24:02 +00:00
for (std::unique_ptr<IndexFile>& entry : result) {
entry->import_file = file;
entry->args = args;
auto it = UID2skipped.find(entry->UID);
if (it != UID2skipped.end())
entry->skipped_ranges = std::move(it->second);
for (auto& it : entry->usr2func) {
2018-03-07 23:08:50 +00:00
// e.g. declaration + out-of-line definition
Uniquify(it.second.derived);
Uniquify(it.second.uses);
2018-03-07 23:08:50 +00:00
}
for (auto& it : entry->usr2type) {
Uniquify(it.second.derived);
Uniquify(it.second.uses);
// e.g. declaration + out-of-line definition
Uniquify(it.second.def.funcs);
}
for (auto& it : entry->usr2var)
Uniquify(it.second.uses);
2018-01-06 17:28:55 +00:00
if (param.primary_file) {
// If there are errors, show at least one at the include position.
auto it = inc_to_line.find(entry->path);
if (it != inc_to_line.end()) {
int line = it->second;
for (auto ls_diagnostic : entry->diagnostics_) {
if (ls_diagnostic.severity != lsDiagnosticSeverity::Error)
continue;
ls_diagnostic.range =
2018-04-08 00:10:54 +00:00
lsRange{lsPosition{line, 10}, lsPosition{line, 10}};
param.primary_file->diagnostics_.push_back(ls_diagnostic);
break;
}
2018-01-06 17:28:55 +00:00
}
}
2017-07-30 04:24:02 +00:00
// Update file contents and modification time.
entry->last_write_time = param.file2write_time[entry->path];
2017-07-30 04:24:02 +00:00
// Update dependencies for the file. Do not include the file in its own
// dependency set.
for (const std::string& path : param.seen_files)
if (path != entry->path && path != entry->import_file)
entry->dependencies[path] = param.file2write_time[path];
2017-04-08 22:54:36 +00:00
}
2017-04-19 00:05:14 +00:00
2018-04-02 07:22:12 +00:00
return result;
2017-03-22 17:16:09 +00:00
}
void IndexInit() {
// InitLLVM
CXIndex CXIdx = clang_createIndex(0, 0);
clang_disposeIndex(CXIdx);
2017-04-18 23:49:55 +00:00
}
// |SymbolRef| is serialized this way.
// |Use| also uses this though it has an extra field |file|,
// which is not used by Index* so it does not need to be serialized.
void Reflect(Reader& visitor, Reference& value) {
if (visitor.Format() == SerializeFormat::Json) {
std::string t = visitor.GetString();
char* s = const_cast<char*>(t.c_str());
value.range = Range::FromString(s);
s = strchr(s, '|');
value.usr = strtoull(s + 1, &s, 10);
value.kind = static_cast<SymbolKind>(strtol(s + 1, &s, 10));
value.role = static_cast<Role>(strtol(s + 1, &s, 10));
} else {
Reflect(visitor, value.range);
Reflect(visitor, value.usr);
2018-02-09 17:42:10 +00:00
Reflect(visitor, value.kind);
Reflect(visitor, value.role);
}
}
void Reflect(Writer& visitor, Reference& value) {
if (visitor.Format() == SerializeFormat::Json) {
char buf[99];
snprintf(buf, sizeof buf, "%s|%" PRIu64 "|%d|%d",
value.range.ToString().c_str(), value.usr, int(value.kind),
int(value.role));
std::string s(buf);
Reflect(visitor, s);
} else {
Reflect(visitor, value.range);
Reflect(visitor, value.usr);
2018-02-09 17:42:10 +00:00
Reflect(visitor, value.kind);
Reflect(visitor, value.role);
}
}