1 //===--- CodeComplete.cpp ----------------------------------------*- C++-*-===//
2 //
3 // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
4 // See https://llvm.org/LICENSE.txt for license information.
5 // SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
6 //
7 //===----------------------------------------------------------------------===//
8 //
9 // Code completion has several moving parts:
10 //  - AST-based completions are provided using the completion hooks in Sema.
11 //  - external completions are retrieved from the index (using hints from Sema)
12 //  - the two sources overlap, and must be merged and overloads bundled
13 //  - results must be scored and ranked (see Quality.h) before rendering
14 //
15 // Signature help works in a similar way as code completion, but it is simpler:
16 // it's purely AST-based, and there are few candidates.
17 //
18 //===----------------------------------------------------------------------===//
19 
20 #include "CodeComplete.h"
21 #include "AST.h"
22 #include "ClangdUnit.h"
23 #include "CodeCompletionStrings.h"
24 #include "Compiler.h"
25 #include "Diagnostics.h"
26 #include "ExpectedTypes.h"
27 #include "FileDistance.h"
28 #include "FuzzyMatch.h"
29 #include "Headers.h"
30 #include "Logger.h"
31 #include "Protocol.h"
32 #include "Quality.h"
33 #include "SourceCode.h"
34 #include "TUScheduler.h"
35 #include "Trace.h"
36 #include "URI.h"
37 #include "index/Index.h"
38 #include "index/Symbol.h"
39 #include "index/SymbolOrigin.h"
40 #include "clang/AST/Decl.h"
41 #include "clang/AST/DeclBase.h"
42 #include "clang/Basic/CharInfo.h"
43 #include "clang/Basic/LangOptions.h"
44 #include "clang/Basic/SourceLocation.h"
45 #include "clang/Format/Format.h"
46 #include "clang/Frontend/CompilerInstance.h"
47 #include "clang/Frontend/FrontendActions.h"
48 #include "clang/Lex/ExternalPreprocessorSource.h"
49 #include "clang/Lex/Preprocessor.h"
50 #include "clang/Lex/PreprocessorOptions.h"
51 #include "clang/Sema/CodeCompleteConsumer.h"
52 #include "clang/Sema/DeclSpec.h"
53 #include "clang/Sema/Sema.h"
54 #include "llvm/ADT/ArrayRef.h"
55 #include "llvm/ADT/None.h"
56 #include "llvm/ADT/Optional.h"
57 #include "llvm/ADT/SmallVector.h"
58 #include "llvm/ADT/StringExtras.h"
59 #include "llvm/ADT/StringRef.h"
60 #include "llvm/Support/Compiler.h"
61 #include "llvm/Support/Debug.h"
62 #include "llvm/Support/Error.h"
63 #include "llvm/Support/Format.h"
64 #include "llvm/Support/FormatVariadic.h"
65 #include "llvm/Support/ScopedPrinter.h"
66 #include <algorithm>
67 #include <iterator>
68 
69 // We log detailed candidate here if you run with -debug-only=codecomplete.
70 #define DEBUG_TYPE "CodeComplete"
71 
72 namespace clang {
73 namespace clangd {
74 namespace {
75 
76 CompletionItemKind toCompletionItemKind(index::SymbolKind Kind) {
77   using SK = index::SymbolKind;
78   switch (Kind) {
79   case SK::Unknown:
80     return CompletionItemKind::Missing;
81   case SK::Module:
82   case SK::Namespace:
83   case SK::NamespaceAlias:
84     return CompletionItemKind::Module;
85   case SK::Macro:
86     return CompletionItemKind::Text;
87   case SK::Enum:
88     return CompletionItemKind::Enum;
89   // FIXME(ioeric): use LSP struct instead of class when it is suppoted in the
90   // protocol.
91   case SK::Struct:
92   case SK::Class:
93   case SK::Protocol:
94   case SK::Extension:
95   case SK::Union:
96     return CompletionItemKind::Class;
97   case SK::TypeAlias:
98     // We use the same kind as the VSCode C++ extension.
99     // FIXME: pick a better option when we have one.
100     return CompletionItemKind::Interface;
101   case SK::Using:
102     return CompletionItemKind::Reference;
103   case SK::Function:
104   // FIXME(ioeric): this should probably be an operator. This should be fixed
105   // when `Operator` is support type in the protocol.
106   case SK::ConversionFunction:
107     return CompletionItemKind::Function;
108   case SK::Variable:
109   case SK::Parameter:
110     return CompletionItemKind::Variable;
111   case SK::Field:
112     return CompletionItemKind::Field;
113   // FIXME(ioeric): use LSP enum constant when it is supported in the protocol.
114   case SK::EnumConstant:
115     return CompletionItemKind::Value;
116   case SK::InstanceMethod:
117   case SK::ClassMethod:
118   case SK::StaticMethod:
119   case SK::Destructor:
120     return CompletionItemKind::Method;
121   case SK::InstanceProperty:
122   case SK::ClassProperty:
123   case SK::StaticProperty:
124     return CompletionItemKind::Property;
125   case SK::Constructor:
126     return CompletionItemKind::Constructor;
127   }
128   llvm_unreachable("Unhandled clang::index::SymbolKind.");
129 }
130 
131 CompletionItemKind
132 toCompletionItemKind(CodeCompletionResult::ResultKind ResKind,
133                      const NamedDecl *Decl,
134                      CodeCompletionContext::Kind CtxKind) {
135   if (Decl)
136     return toCompletionItemKind(index::getSymbolInfo(Decl).Kind);
137   if (CtxKind == CodeCompletionContext::CCC_IncludedFile)
138     return CompletionItemKind::File;
139   switch (ResKind) {
140   case CodeCompletionResult::RK_Declaration:
141     llvm_unreachable("RK_Declaration without Decl");
142   case CodeCompletionResult::RK_Keyword:
143     return CompletionItemKind::Keyword;
144   case CodeCompletionResult::RK_Macro:
145     return CompletionItemKind::Text; // unfortunately, there's no 'Macro'
146                                      // completion items in LSP.
147   case CodeCompletionResult::RK_Pattern:
148     return CompletionItemKind::Snippet;
149   }
150   llvm_unreachable("Unhandled CodeCompletionResult::ResultKind.");
151 }
152 
153 // Identifier code completion result.
154 struct RawIdentifier {
155   llvm::StringRef Name;
156   unsigned References; // # of usages in file.
157 };
158 
159 /// A code completion result, in clang-native form.
160 /// It may be promoted to a CompletionItem if it's among the top-ranked results.
161 struct CompletionCandidate {
162   llvm::StringRef Name; // Used for filtering and sorting.
163   // We may have a result from Sema, from the index, or both.
164   const CodeCompletionResult *SemaResult = nullptr;
165   const Symbol *IndexResult = nullptr;
166   const RawIdentifier *IdentifierResult = nullptr;
167   llvm::SmallVector<llvm::StringRef, 1> RankedIncludeHeaders;
168 
169   // Returns a token identifying the overload set this is part of.
170   // 0 indicates it's not part of any overload set.
171   size_t overloadSet(const CodeCompleteOptions &Opts) const {
172     if (!Opts.BundleOverloads.getValueOr(false))
173       return 0;
174     llvm::SmallString<256> Scratch;
175     if (IndexResult) {
176       switch (IndexResult->SymInfo.Kind) {
177       case index::SymbolKind::ClassMethod:
178       case index::SymbolKind::InstanceMethod:
179       case index::SymbolKind::StaticMethod:
180 #ifndef NDEBUG
181         llvm_unreachable("Don't expect members from index in code completion");
182 #else
183         LLVM_FALLTHROUGH;
184 #endif
185       case index::SymbolKind::Function:
186         // We can't group overloads together that need different #includes.
187         // This could break #include insertion.
188         return llvm::hash_combine(
189             (IndexResult->Scope + IndexResult->Name).toStringRef(Scratch),
190             headerToInsertIfAllowed(Opts).getValueOr(""));
191       default:
192         return 0;
193       }
194     }
195     if (SemaResult) {
196       // We need to make sure we're consistent with the IndexResult case!
197       const NamedDecl *D = SemaResult->Declaration;
198       if (!D || !D->isFunctionOrFunctionTemplate())
199         return 0;
200       {
201         llvm::raw_svector_ostream OS(Scratch);
202         D->printQualifiedName(OS);
203       }
204       return llvm::hash_combine(Scratch,
205                                 headerToInsertIfAllowed(Opts).getValueOr(""));
206     }
207     assert(IdentifierResult);
208     return 0;
209   }
210 
211   // The best header to include if include insertion is allowed.
212   llvm::Optional<llvm::StringRef>
213   headerToInsertIfAllowed(const CodeCompleteOptions &Opts) const {
214     if (Opts.InsertIncludes == CodeCompleteOptions::NeverInsert ||
215         RankedIncludeHeaders.empty())
216       return None;
217     if (SemaResult && SemaResult->Declaration) {
218       // Avoid inserting new #include if the declaration is found in the current
219       // file e.g. the symbol is forward declared.
220       auto &SM = SemaResult->Declaration->getASTContext().getSourceManager();
221       for (const Decl *RD : SemaResult->Declaration->redecls())
222         if (SM.isInMainFile(SM.getExpansionLoc(RD->getBeginLoc())))
223           return None;
224     }
225     return RankedIncludeHeaders[0];
226   }
227 
228   using Bundle = llvm::SmallVector<CompletionCandidate, 4>;
229 };
230 using ScoredBundle =
231     std::pair<CompletionCandidate::Bundle, CodeCompletion::Scores>;
232 struct ScoredBundleGreater {
233   bool operator()(const ScoredBundle &L, const ScoredBundle &R) {
234     if (L.second.Total != R.second.Total)
235       return L.second.Total > R.second.Total;
236     return L.first.front().Name <
237            R.first.front().Name; // Earlier name is better.
238   }
239 };
240 
241 // Assembles a code completion out of a bundle of >=1 completion candidates.
242 // Many of the expensive strings are only computed at this point, once we know
243 // the candidate bundle is going to be returned.
244 //
245 // Many fields are the same for all candidates in a bundle (e.g. name), and are
246 // computed from the first candidate, in the constructor.
247 // Others vary per candidate, so add() must be called for remaining candidates.
248 struct CodeCompletionBuilder {
249   CodeCompletionBuilder(ASTContext *ASTCtx, const CompletionCandidate &C,
250                         CodeCompletionString *SemaCCS,
251                         llvm::ArrayRef<std::string> QueryScopes,
252                         const IncludeInserter &Includes,
253                         llvm::StringRef FileName,
254                         CodeCompletionContext::Kind ContextKind,
255                         const CodeCompleteOptions &Opts)
256       : ASTCtx(ASTCtx), ExtractDocumentation(Opts.IncludeComments),
257         EnableFunctionArgSnippets(Opts.EnableFunctionArgSnippets) {
258     add(C, SemaCCS);
259     if (C.SemaResult) {
260       assert(ASTCtx);
261       Completion.Origin |= SymbolOrigin::AST;
262       Completion.Name = llvm::StringRef(SemaCCS->getTypedText());
263       if (Completion.Scope.empty()) {
264         if ((C.SemaResult->Kind == CodeCompletionResult::RK_Declaration) ||
265             (C.SemaResult->Kind == CodeCompletionResult::RK_Pattern))
266           if (const auto *D = C.SemaResult->getDeclaration())
267             if (const auto *ND = dyn_cast<NamedDecl>(D))
268               Completion.Scope =
269                   splitQualifiedName(printQualifiedName(*ND)).first;
270       }
271       Completion.Kind = toCompletionItemKind(
272           C.SemaResult->Kind, C.SemaResult->Declaration, ContextKind);
273       // Sema could provide more info on whether the completion was a file or
274       // folder.
275       if (Completion.Kind == CompletionItemKind::File &&
276           Completion.Name.back() == '/')
277         Completion.Kind = CompletionItemKind::Folder;
278       for (const auto &FixIt : C.SemaResult->FixIts) {
279         Completion.FixIts.push_back(toTextEdit(
280             FixIt, ASTCtx->getSourceManager(), ASTCtx->getLangOpts()));
281       }
282       llvm::sort(Completion.FixIts, [](const TextEdit &X, const TextEdit &Y) {
283         return std::tie(X.range.start.line, X.range.start.character) <
284                std::tie(Y.range.start.line, Y.range.start.character);
285       });
286       Completion.Deprecated |=
287           (C.SemaResult->Availability == CXAvailability_Deprecated);
288     }
289     if (C.IndexResult) {
290       Completion.Origin |= C.IndexResult->Origin;
291       if (Completion.Scope.empty())
292         Completion.Scope = C.IndexResult->Scope;
293       if (Completion.Kind == CompletionItemKind::Missing)
294         Completion.Kind = toCompletionItemKind(C.IndexResult->SymInfo.Kind);
295       if (Completion.Name.empty())
296         Completion.Name = C.IndexResult->Name;
297       // If the completion was visible to Sema, no qualifier is needed. This
298       // avoids unneeded qualifiers in cases like with `using ns::X`.
299       if (Completion.RequiredQualifier.empty() && !C.SemaResult) {
300         llvm::StringRef ShortestQualifier = C.IndexResult->Scope;
301         for (llvm::StringRef Scope : QueryScopes) {
302           llvm::StringRef Qualifier = C.IndexResult->Scope;
303           if (Qualifier.consume_front(Scope) &&
304               Qualifier.size() < ShortestQualifier.size())
305             ShortestQualifier = Qualifier;
306         }
307         Completion.RequiredQualifier = ShortestQualifier;
308       }
309       Completion.Deprecated |= (C.IndexResult->Flags & Symbol::Deprecated);
310     }
311     if (C.IdentifierResult) {
312       Completion.Origin |= SymbolOrigin::Identifier;
313       Completion.Kind = CompletionItemKind::Text;
314       Completion.Name = C.IdentifierResult->Name;
315     }
316 
317     // Turn absolute path into a literal string that can be #included.
318     auto Inserted = [&](llvm::StringRef Header)
319         -> llvm::Expected<std::pair<std::string, bool>> {
320       auto DeclaringURI =
321           URI::parse(C.IndexResult->CanonicalDeclaration.FileURI);
322       if (!DeclaringURI)
323         return DeclaringURI.takeError();
324       auto ResolvedDeclaring = URI::resolve(*DeclaringURI, FileName);
325       if (!ResolvedDeclaring)
326         return ResolvedDeclaring.takeError();
327       auto ResolvedInserted = toHeaderFile(Header, FileName);
328       if (!ResolvedInserted)
329         return ResolvedInserted.takeError();
330       auto Spelled = Includes.calculateIncludePath(*ResolvedInserted, FileName);
331       if (!Spelled)
332         return llvm::createStringError(llvm::inconvertibleErrorCode(),
333                                        "Header not on include path");
334       return std::make_pair(
335           std::move(*Spelled),
336           Includes.shouldInsertInclude(*ResolvedDeclaring, *ResolvedInserted));
337     };
338     bool ShouldInsert = C.headerToInsertIfAllowed(Opts).hasValue();
339     // Calculate include paths and edits for all possible headers.
340     for (const auto &Inc : C.RankedIncludeHeaders) {
341       if (auto ToInclude = Inserted(Inc)) {
342         CodeCompletion::IncludeCandidate Include;
343         Include.Header = ToInclude->first;
344         if (ToInclude->second && ShouldInsert)
345           Include.Insertion = Includes.insert(ToInclude->first);
346         Completion.Includes.push_back(std::move(Include));
347       } else
348         log("Failed to generate include insertion edits for adding header "
349             "(FileURI='{0}', IncludeHeader='{1}') into {2}",
350             C.IndexResult->CanonicalDeclaration.FileURI, Inc, FileName);
351     }
352     // Prefer includes that do not need edits (i.e. already exist).
353     std::stable_partition(Completion.Includes.begin(),
354                           Completion.Includes.end(),
355                           [](const CodeCompletion::IncludeCandidate &I) {
356                             return !I.Insertion.hasValue();
357                           });
358   }
359 
360   void add(const CompletionCandidate &C, CodeCompletionString *SemaCCS) {
361     assert(bool(C.SemaResult) == bool(SemaCCS));
362     Bundled.emplace_back();
363     BundledEntry &S = Bundled.back();
364     if (C.SemaResult) {
365       bool IsPattern = C.SemaResult->Kind == CodeCompletionResult::RK_Pattern;
366       getSignature(*SemaCCS, &S.Signature, &S.SnippetSuffix,
367                    &Completion.RequiredQualifier, IsPattern);
368       S.ReturnType = getReturnType(*SemaCCS);
369     } else if (C.IndexResult) {
370       S.Signature = C.IndexResult->Signature;
371       S.SnippetSuffix = C.IndexResult->CompletionSnippetSuffix;
372       S.ReturnType = C.IndexResult->ReturnType;
373     }
374     if (ExtractDocumentation && Completion.Documentation.empty()) {
375       if (C.IndexResult)
376         Completion.Documentation = C.IndexResult->Documentation;
377       else if (C.SemaResult)
378         Completion.Documentation = getDocComment(*ASTCtx, *C.SemaResult,
379                                                  /*CommentsFromHeader=*/false);
380     }
381   }
382 
383   CodeCompletion build() {
384     Completion.ReturnType = summarizeReturnType();
385     Completion.Signature = summarizeSignature();
386     Completion.SnippetSuffix = summarizeSnippet();
387     Completion.BundleSize = Bundled.size();
388     return std::move(Completion);
389   }
390 
391 private:
392   struct BundledEntry {
393     std::string SnippetSuffix;
394     std::string Signature;
395     std::string ReturnType;
396   };
397 
398   // If all BundledEntrys have the same value for a property, return it.
399   template <std::string BundledEntry::*Member>
400   const std::string *onlyValue() const {
401     auto B = Bundled.begin(), E = Bundled.end();
402     for (auto I = B + 1; I != E; ++I)
403       if (I->*Member != B->*Member)
404         return nullptr;
405     return &(B->*Member);
406   }
407 
408   template <bool BundledEntry::*Member> const bool *onlyValue() const {
409     auto B = Bundled.begin(), E = Bundled.end();
410     for (auto I = B + 1; I != E; ++I)
411       if (I->*Member != B->*Member)
412         return nullptr;
413     return &(B->*Member);
414   }
415 
416   std::string summarizeReturnType() const {
417     if (auto *RT = onlyValue<&BundledEntry::ReturnType>())
418       return *RT;
419     return "";
420   }
421 
422   std::string summarizeSnippet() const {
423     auto *Snippet = onlyValue<&BundledEntry::SnippetSuffix>();
424     if (!Snippet)
425       // All bundles are function calls.
426       // FIXME(ibiryukov): sometimes add template arguments to a snippet, e.g.
427       // we need to complete 'forward<$1>($0)'.
428       return "($0)";
429     if (EnableFunctionArgSnippets)
430       return *Snippet;
431 
432     // Replace argument snippets with a simplified pattern.
433     if (Snippet->empty())
434       return "";
435     if (Completion.Kind == CompletionItemKind::Function ||
436         Completion.Kind == CompletionItemKind::Method) {
437       // Functions snippets can be of 2 types:
438       // - containing only function arguments, e.g.
439       //   foo(${1:int p1}, ${2:int p2});
440       //   We transform this pattern to '($0)' or '()'.
441       // - template arguments and function arguments, e.g.
442       //   foo<${1:class}>(${2:int p1}).
443       //   We transform this pattern to '<$1>()$0' or '<$0>()'.
444 
445       bool EmptyArgs = llvm::StringRef(*Snippet).endswith("()");
446       if (Snippet->front() == '<')
447         return EmptyArgs ? "<$1>()$0" : "<$1>($0)";
448       if (Snippet->front() == '(')
449         return EmptyArgs ? "()" : "($0)";
450       return *Snippet; // Not an arg snippet?
451     }
452     // 'CompletionItemKind::Interface' matches template type aliases.
453     if (Completion.Kind == CompletionItemKind::Interface ||
454         Completion.Kind == CompletionItemKind::Class) {
455       if (Snippet->front() != '<')
456         return *Snippet; // Not an arg snippet?
457 
458       // Classes and template using aliases can only have template arguments,
459       // e.g. Foo<${1:class}>.
460       if (llvm::StringRef(*Snippet).endswith("<>"))
461         return "<>"; // can happen with defaulted template arguments.
462       return "<$0>";
463     }
464     return *Snippet;
465   }
466 
467   std::string summarizeSignature() const {
468     if (auto *Signature = onlyValue<&BundledEntry::Signature>())
469       return *Signature;
470     // All bundles are function calls.
471     return "(…)";
472   }
473 
474   // ASTCtx can be nullptr if not run with sema.
475   ASTContext *ASTCtx;
476   CodeCompletion Completion;
477   llvm::SmallVector<BundledEntry, 1> Bundled;
478   bool ExtractDocumentation;
479   bool EnableFunctionArgSnippets;
480 };
481 
482 // Determine the symbol ID for a Sema code completion result, if possible.
483 llvm::Optional<SymbolID> getSymbolID(const CodeCompletionResult &R,
484                                      const SourceManager &SM) {
485   switch (R.Kind) {
486   case CodeCompletionResult::RK_Declaration:
487   case CodeCompletionResult::RK_Pattern: {
488     return clang::clangd::getSymbolID(R.Declaration);
489   }
490   case CodeCompletionResult::RK_Macro:
491     return clang::clangd::getSymbolID(*R.Macro, R.MacroDefInfo, SM);
492   case CodeCompletionResult::RK_Keyword:
493     return None;
494   }
495   llvm_unreachable("unknown CodeCompletionResult kind");
496 }
497 
498 // Scopes of the paritial identifier we're trying to complete.
499 // It is used when we query the index for more completion results.
500 struct SpecifiedScope {
501   // The scopes we should look in, determined by Sema.
502   //
503   // If the qualifier was fully resolved, we look for completions in these
504   // scopes; if there is an unresolved part of the qualifier, it should be
505   // resolved within these scopes.
506   //
507   // Examples of qualified completion:
508   //
509   //   "::vec"                                      => {""}
510   //   "using namespace std; ::vec^"                => {"", "std::"}
511   //   "namespace ns {using namespace std;} ns::^"  => {"ns::", "std::"}
512   //   "std::vec^"                                  => {""}  // "std" unresolved
513   //
514   // Examples of unqualified completion:
515   //
516   //   "vec^"                                       => {""}
517   //   "using namespace std; vec^"                  => {"", "std::"}
518   //   "using namespace std; namespace ns { vec^ }" => {"ns::", "std::", ""}
519   //
520   // "" for global namespace, "ns::" for normal namespace.
521   std::vector<std::string> AccessibleScopes;
522   // The full scope qualifier as typed by the user (without the leading "::").
523   // Set if the qualifier is not fully resolved by Sema.
524   llvm::Optional<std::string> UnresolvedQualifier;
525 
526   // Construct scopes being queried in indexes. The results are deduplicated.
527   // This method format the scopes to match the index request representation.
528   std::vector<std::string> scopesForIndexQuery() {
529     std::set<std::string> Results;
530     for (llvm::StringRef AS : AccessibleScopes)
531       Results.insert(
532           (AS + (UnresolvedQualifier ? *UnresolvedQualifier : "")).str());
533     return {Results.begin(), Results.end()};
534   }
535 };
536 
537 // Get all scopes that will be queried in indexes and whether symbols from
538 // any scope is allowed. The first scope in the list is the preferred scope
539 // (e.g. enclosing namespace).
540 std::pair<std::vector<std::string>, bool>
541 getQueryScopes(CodeCompletionContext &CCContext, const Sema &CCSema,
542                const CompletionPrefix &HeuristicPrefix,
543                const CodeCompleteOptions &Opts) {
544   SpecifiedScope Scopes;
545   for (auto *Context : CCContext.getVisitedContexts()) {
546     if (isa<TranslationUnitDecl>(Context))
547       Scopes.AccessibleScopes.push_back(""); // global namespace
548     else if (isa<NamespaceDecl>(Context))
549       Scopes.AccessibleScopes.push_back(printNamespaceScope(*Context));
550   }
551 
552   const CXXScopeSpec *SemaSpecifier =
553       CCContext.getCXXScopeSpecifier().getValueOr(nullptr);
554   // Case 1: unqualified completion.
555   if (!SemaSpecifier) {
556     // Case 2 (exception): sema saw no qualifier, but there appears to be one!
557     // This can happen e.g. in incomplete macro expansions. Use heuristics.
558     if (!HeuristicPrefix.Qualifier.empty()) {
559       vlog("Sema said no scope specifier, but we saw {0} in the source code",
560            HeuristicPrefix.Qualifier);
561       StringRef SpelledSpecifier = HeuristicPrefix.Qualifier;
562       if (SpelledSpecifier.consume_front("::"))
563         Scopes.AccessibleScopes = {""};
564       Scopes.UnresolvedQualifier = SpelledSpecifier;
565       return {Scopes.scopesForIndexQuery(), false};
566     }
567     // The enclosing namespace must be first, it gets a quality boost.
568     std::vector<std::string> EnclosingAtFront;
569     std::string EnclosingScope = printNamespaceScope(*CCSema.CurContext);
570     EnclosingAtFront.push_back(EnclosingScope);
571     for (auto &S : Scopes.scopesForIndexQuery()) {
572       if (EnclosingScope != S)
573         EnclosingAtFront.push_back(std::move(S));
574     }
575     // Allow AllScopes completion as there is no explicit scope qualifier.
576     return {EnclosingAtFront, Opts.AllScopes};
577   }
578   // Case 3: sema saw and resolved a scope qualifier.
579   if (SemaSpecifier && SemaSpecifier->isValid())
580     return {Scopes.scopesForIndexQuery(), false};
581 
582   // Case 4: There was a qualifier, and Sema didn't resolve it.
583   Scopes.AccessibleScopes.push_back(""); // Make sure global scope is included.
584   llvm::StringRef SpelledSpecifier = Lexer::getSourceText(
585       CharSourceRange::getCharRange(SemaSpecifier->getRange()),
586       CCSema.SourceMgr, clang::LangOptions());
587   if (SpelledSpecifier.consume_front("::"))
588     Scopes.AccessibleScopes = {""};
589   Scopes.UnresolvedQualifier = SpelledSpecifier;
590   // Sema excludes the trailing "::".
591   if (!Scopes.UnresolvedQualifier->empty())
592     *Scopes.UnresolvedQualifier += "::";
593 
594   return {Scopes.scopesForIndexQuery(), false};
595 }
596 
597 // Should we perform index-based completion in a context of the specified kind?
598 // FIXME: consider allowing completion, but restricting the result types.
599 bool contextAllowsIndex(enum CodeCompletionContext::Kind K) {
600   switch (K) {
601   case CodeCompletionContext::CCC_TopLevel:
602   case CodeCompletionContext::CCC_ObjCInterface:
603   case CodeCompletionContext::CCC_ObjCImplementation:
604   case CodeCompletionContext::CCC_ObjCIvarList:
605   case CodeCompletionContext::CCC_ClassStructUnion:
606   case CodeCompletionContext::CCC_Statement:
607   case CodeCompletionContext::CCC_Expression:
608   case CodeCompletionContext::CCC_ObjCMessageReceiver:
609   case CodeCompletionContext::CCC_EnumTag:
610   case CodeCompletionContext::CCC_UnionTag:
611   case CodeCompletionContext::CCC_ClassOrStructTag:
612   case CodeCompletionContext::CCC_ObjCProtocolName:
613   case CodeCompletionContext::CCC_Namespace:
614   case CodeCompletionContext::CCC_Type:
615   case CodeCompletionContext::CCC_ParenthesizedExpression:
616   case CodeCompletionContext::CCC_ObjCInterfaceName:
617   case CodeCompletionContext::CCC_ObjCCategoryName:
618   case CodeCompletionContext::CCC_Symbol:
619   case CodeCompletionContext::CCC_SymbolOrNewName:
620     return true;
621   case CodeCompletionContext::CCC_OtherWithMacros:
622   case CodeCompletionContext::CCC_DotMemberAccess:
623   case CodeCompletionContext::CCC_ArrowMemberAccess:
624   case CodeCompletionContext::CCC_ObjCPropertyAccess:
625   case CodeCompletionContext::CCC_MacroName:
626   case CodeCompletionContext::CCC_MacroNameUse:
627   case CodeCompletionContext::CCC_PreprocessorExpression:
628   case CodeCompletionContext::CCC_PreprocessorDirective:
629   case CodeCompletionContext::CCC_SelectorName:
630   case CodeCompletionContext::CCC_TypeQualifiers:
631   case CodeCompletionContext::CCC_ObjCInstanceMessage:
632   case CodeCompletionContext::CCC_ObjCClassMessage:
633   case CodeCompletionContext::CCC_IncludedFile:
634   // FIXME: Provide identifier based completions for the following contexts:
635   case CodeCompletionContext::CCC_Other: // Be conservative.
636   case CodeCompletionContext::CCC_NaturalLanguage:
637   case CodeCompletionContext::CCC_Recovery:
638   case CodeCompletionContext::CCC_NewName:
639     return false;
640   }
641   llvm_unreachable("unknown code completion context");
642 }
643 
644 static bool isInjectedClass(const NamedDecl &D) {
645   if (auto *R = dyn_cast_or_null<RecordDecl>(&D))
646     if (R->isInjectedClassName())
647       return true;
648   return false;
649 }
650 
651 // Some member calls are blacklisted because they're so rarely useful.
652 static bool isBlacklistedMember(const NamedDecl &D) {
653   // Destructor completion is rarely useful, and works inconsistently.
654   // (s.^ completes ~string, but s.~st^ is an error).
655   if (D.getKind() == Decl::CXXDestructor)
656     return true;
657   // Injected name may be useful for A::foo(), but who writes A::A::foo()?
658   if (isInjectedClass(D))
659     return true;
660   // Explicit calls to operators are also rare.
661   auto NameKind = D.getDeclName().getNameKind();
662   if (NameKind == DeclarationName::CXXOperatorName ||
663       NameKind == DeclarationName::CXXLiteralOperatorName ||
664       NameKind == DeclarationName::CXXConversionFunctionName)
665     return true;
666   return false;
667 }
668 
669 // The CompletionRecorder captures Sema code-complete output, including context.
670 // It filters out ignored results (but doesn't apply fuzzy-filtering yet).
671 // It doesn't do scoring or conversion to CompletionItem yet, as we want to
672 // merge with index results first.
673 // Generally the fields and methods of this object should only be used from
674 // within the callback.
675 struct CompletionRecorder : public CodeCompleteConsumer {
676   CompletionRecorder(const CodeCompleteOptions &Opts,
677                      llvm::unique_function<void()> ResultsCallback)
678       : CodeCompleteConsumer(Opts.getClangCompleteOpts()),
679         CCContext(CodeCompletionContext::CCC_Other), Opts(Opts),
680         CCAllocator(std::make_shared<GlobalCodeCompletionAllocator>()),
681         CCTUInfo(CCAllocator), ResultsCallback(std::move(ResultsCallback)) {
682     assert(this->ResultsCallback);
683   }
684 
685   std::vector<CodeCompletionResult> Results;
686   CodeCompletionContext CCContext;
687   Sema *CCSema = nullptr; // Sema that created the results.
688   // FIXME: Sema is scary. Can we store ASTContext and Preprocessor, instead?
689 
690   void ProcessCodeCompleteResults(class Sema &S, CodeCompletionContext Context,
691                                   CodeCompletionResult *InResults,
692                                   unsigned NumResults) override final {
693     // Results from recovery mode are generally useless, and the callback after
694     // recovery (if any) is usually more interesting. To make sure we handle the
695     // future callback from sema, we just ignore all callbacks in recovery mode,
696     // as taking only results from recovery mode results in poor completion
697     // results.
698     // FIXME: in case there is no future sema completion callback after the
699     // recovery mode, we might still want to provide some results (e.g. trivial
700     // identifier-based completion).
701     if (Context.getKind() == CodeCompletionContext::CCC_Recovery) {
702       log("Code complete: Ignoring sema code complete callback with Recovery "
703           "context.");
704       return;
705     }
706     // If a callback is called without any sema result and the context does not
707     // support index-based completion, we simply skip it to give way to
708     // potential future callbacks with results.
709     if (NumResults == 0 && !contextAllowsIndex(Context.getKind()))
710       return;
711     if (CCSema) {
712       log("Multiple code complete callbacks (parser backtracked?). "
713           "Dropping results from context {0}, keeping results from {1}.",
714           getCompletionKindString(Context.getKind()),
715           getCompletionKindString(this->CCContext.getKind()));
716       return;
717     }
718     // Record the completion context.
719     CCSema = &S;
720     CCContext = Context;
721 
722     // Retain the results we might want.
723     for (unsigned I = 0; I < NumResults; ++I) {
724       auto &Result = InResults[I];
725       // Class members that are shadowed by subclasses are usually noise.
726       if (Result.Hidden && Result.Declaration &&
727           Result.Declaration->isCXXClassMember())
728         continue;
729       if (!Opts.IncludeIneligibleResults &&
730           (Result.Availability == CXAvailability_NotAvailable ||
731            Result.Availability == CXAvailability_NotAccessible))
732         continue;
733       if (Result.Declaration &&
734           !Context.getBaseType().isNull() // is this a member-access context?
735           && isBlacklistedMember(*Result.Declaration))
736         continue;
737       // Skip injected class name when no class scope is not explicitly set.
738       // E.g. show injected A::A in `using A::A^` but not in "A^".
739       if (Result.Declaration && !Context.getCXXScopeSpecifier().hasValue() &&
740           isInjectedClass(*Result.Declaration))
741         continue;
742       // We choose to never append '::' to completion results in clangd.
743       Result.StartsNestedNameSpecifier = false;
744       Results.push_back(Result);
745     }
746     ResultsCallback();
747   }
748 
749   CodeCompletionAllocator &getAllocator() override { return *CCAllocator; }
750   CodeCompletionTUInfo &getCodeCompletionTUInfo() override { return CCTUInfo; }
751 
752   // Returns the filtering/sorting name for Result, which must be from Results.
753   // Returned string is owned by this recorder (or the AST).
754   llvm::StringRef getName(const CodeCompletionResult &Result) {
755     switch (Result.Kind) {
756     case CodeCompletionResult::RK_Declaration:
757       if (auto *ID = Result.Declaration->getIdentifier())
758         return ID->getName();
759       break;
760     case CodeCompletionResult::RK_Keyword:
761       return Result.Keyword;
762     case CodeCompletionResult::RK_Macro:
763       return Result.Macro->getName();
764     case CodeCompletionResult::RK_Pattern:
765       return Result.Pattern->getTypedText();
766     }
767     auto *CCS = codeCompletionString(Result);
768     return CCS->getTypedText();
769   }
770 
771   // Build a CodeCompletion string for R, which must be from Results.
772   // The CCS will be owned by this recorder.
773   CodeCompletionString *codeCompletionString(const CodeCompletionResult &R) {
774     // CodeCompletionResult doesn't seem to be const-correct. We own it, anyway.
775     return const_cast<CodeCompletionResult &>(R).CreateCodeCompletionString(
776         *CCSema, CCContext, *CCAllocator, CCTUInfo,
777         /*IncludeBriefComments=*/false);
778   }
779 
780 private:
781   CodeCompleteOptions Opts;
782   std::shared_ptr<GlobalCodeCompletionAllocator> CCAllocator;
783   CodeCompletionTUInfo CCTUInfo;
784   llvm::unique_function<void()> ResultsCallback;
785 };
786 
787 struct ScoredSignature {
788   // When set, requires documentation to be requested from the index with this
789   // ID.
790   llvm::Optional<SymbolID> IDForDoc;
791   SignatureInformation Signature;
792   SignatureQualitySignals Quality;
793 };
794 
795 class SignatureHelpCollector final : public CodeCompleteConsumer {
796 public:
797   SignatureHelpCollector(const clang::CodeCompleteOptions &CodeCompleteOpts,
798                          const SymbolIndex *Index, SignatureHelp &SigHelp)
799       : CodeCompleteConsumer(CodeCompleteOpts), SigHelp(SigHelp),
800         Allocator(std::make_shared<clang::GlobalCodeCompletionAllocator>()),
801         CCTUInfo(Allocator), Index(Index) {}
802 
803   void ProcessOverloadCandidates(Sema &S, unsigned CurrentArg,
804                                  OverloadCandidate *Candidates,
805                                  unsigned NumCandidates,
806                                  SourceLocation OpenParLoc) override {
807     assert(!OpenParLoc.isInvalid());
808     SourceManager &SrcMgr = S.getSourceManager();
809     OpenParLoc = SrcMgr.getFileLoc(OpenParLoc);
810     if (SrcMgr.isInMainFile(OpenParLoc))
811       SigHelp.argListStart = sourceLocToPosition(SrcMgr, OpenParLoc);
812     else
813       elog("Location oustide main file in signature help: {0}",
814            OpenParLoc.printToString(SrcMgr));
815 
816     std::vector<ScoredSignature> ScoredSignatures;
817     SigHelp.signatures.reserve(NumCandidates);
818     ScoredSignatures.reserve(NumCandidates);
819     // FIXME(rwols): How can we determine the "active overload candidate"?
820     // Right now the overloaded candidates seem to be provided in a "best fit"
821     // order, so I'm not too worried about this.
822     SigHelp.activeSignature = 0;
823     assert(CurrentArg <= (unsigned)std::numeric_limits<int>::max() &&
824            "too many arguments");
825     SigHelp.activeParameter = static_cast<int>(CurrentArg);
826     for (unsigned I = 0; I < NumCandidates; ++I) {
827       OverloadCandidate Candidate = Candidates[I];
828       // We want to avoid showing instantiated signatures, because they may be
829       // long in some cases (e.g. when 'T' is substituted with 'std::string', we
830       // would get 'std::basic_string<char>').
831       if (auto *Func = Candidate.getFunction()) {
832         if (auto *Pattern = Func->getTemplateInstantiationPattern())
833           Candidate = OverloadCandidate(Pattern);
834       }
835 
836       const auto *CCS = Candidate.CreateSignatureString(
837           CurrentArg, S, *Allocator, CCTUInfo, true);
838       assert(CCS && "Expected the CodeCompletionString to be non-null");
839       ScoredSignatures.push_back(processOverloadCandidate(
840           Candidate, *CCS,
841           Candidate.getFunction()
842               ? getDeclComment(S.getASTContext(), *Candidate.getFunction())
843               : ""));
844     }
845 
846     // Sema does not load the docs from the preamble, so we need to fetch extra
847     // docs from the index instead.
848     llvm::DenseMap<SymbolID, std::string> FetchedDocs;
849     if (Index) {
850       LookupRequest IndexRequest;
851       for (const auto &S : ScoredSignatures) {
852         if (!S.IDForDoc)
853           continue;
854         IndexRequest.IDs.insert(*S.IDForDoc);
855       }
856       Index->lookup(IndexRequest, [&](const Symbol &S) {
857         if (!S.Documentation.empty())
858           FetchedDocs[S.ID] = S.Documentation;
859       });
860       log("SigHelp: requested docs for {0} symbols from the index, got {1} "
861           "symbols with non-empty docs in the response",
862           IndexRequest.IDs.size(), FetchedDocs.size());
863     }
864 
865     llvm::sort(ScoredSignatures, [](const ScoredSignature &L,
866                                     const ScoredSignature &R) {
867       // Ordering follows:
868       // - Less number of parameters is better.
869       // - Function is better than FunctionType which is better than
870       // Function Template.
871       // - High score is better.
872       // - Shorter signature is better.
873       // - Alphebatically smaller is better.
874       if (L.Quality.NumberOfParameters != R.Quality.NumberOfParameters)
875         return L.Quality.NumberOfParameters < R.Quality.NumberOfParameters;
876       if (L.Quality.NumberOfOptionalParameters !=
877           R.Quality.NumberOfOptionalParameters)
878         return L.Quality.NumberOfOptionalParameters <
879                R.Quality.NumberOfOptionalParameters;
880       if (L.Quality.Kind != R.Quality.Kind) {
881         using OC = CodeCompleteConsumer::OverloadCandidate;
882         switch (L.Quality.Kind) {
883         case OC::CK_Function:
884           return true;
885         case OC::CK_FunctionType:
886           return R.Quality.Kind != OC::CK_Function;
887         case OC::CK_FunctionTemplate:
888           return false;
889         }
890         llvm_unreachable("Unknown overload candidate type.");
891       }
892       if (L.Signature.label.size() != R.Signature.label.size())
893         return L.Signature.label.size() < R.Signature.label.size();
894       return L.Signature.label < R.Signature.label;
895     });
896 
897     for (auto &SS : ScoredSignatures) {
898       auto IndexDocIt =
899           SS.IDForDoc ? FetchedDocs.find(*SS.IDForDoc) : FetchedDocs.end();
900       if (IndexDocIt != FetchedDocs.end())
901         SS.Signature.documentation = IndexDocIt->second;
902 
903       SigHelp.signatures.push_back(std::move(SS.Signature));
904     }
905   }
906 
907   GlobalCodeCompletionAllocator &getAllocator() override { return *Allocator; }
908 
909   CodeCompletionTUInfo &getCodeCompletionTUInfo() override { return CCTUInfo; }
910 
911 private:
912   void processParameterChunk(llvm::StringRef ChunkText,
913                              SignatureInformation &Signature) const {
914     // (!) this is O(n), should still be fast compared to building ASTs.
915     unsigned ParamStartOffset = lspLength(Signature.label);
916     unsigned ParamEndOffset = ParamStartOffset + lspLength(ChunkText);
917     // A piece of text that describes the parameter that corresponds to
918     // the code-completion location within a function call, message send,
919     // macro invocation, etc.
920     Signature.label += ChunkText;
921     ParameterInformation Info;
922     Info.labelOffsets.emplace(ParamStartOffset, ParamEndOffset);
923     // FIXME: only set 'labelOffsets' when all clients migrate out of it.
924     Info.labelString = ChunkText;
925 
926     Signature.parameters.push_back(std::move(Info));
927   }
928 
929   void processOptionalChunk(const CodeCompletionString &CCS,
930                             SignatureInformation &Signature,
931                             SignatureQualitySignals &Signal) const {
932     for (const auto &Chunk : CCS) {
933       switch (Chunk.Kind) {
934       case CodeCompletionString::CK_Optional:
935         assert(Chunk.Optional &&
936                "Expected the optional code completion string to be non-null.");
937         processOptionalChunk(*Chunk.Optional, Signature, Signal);
938         break;
939       case CodeCompletionString::CK_VerticalSpace:
940         break;
941       case CodeCompletionString::CK_CurrentParameter:
942       case CodeCompletionString::CK_Placeholder:
943         processParameterChunk(Chunk.Text, Signature);
944         Signal.NumberOfOptionalParameters++;
945         break;
946       default:
947         Signature.label += Chunk.Text;
948         break;
949       }
950     }
951   }
952 
953   // FIXME(ioeric): consider moving CodeCompletionString logic here to
954   // CompletionString.h.
955   ScoredSignature processOverloadCandidate(const OverloadCandidate &Candidate,
956                                            const CodeCompletionString &CCS,
957                                            llvm::StringRef DocComment) const {
958     SignatureInformation Signature;
959     SignatureQualitySignals Signal;
960     const char *ReturnType = nullptr;
961 
962     Signature.documentation = formatDocumentation(CCS, DocComment);
963     Signal.Kind = Candidate.getKind();
964 
965     for (const auto &Chunk : CCS) {
966       switch (Chunk.Kind) {
967       case CodeCompletionString::CK_ResultType:
968         // A piece of text that describes the type of an entity or,
969         // for functions and methods, the return type.
970         assert(!ReturnType && "Unexpected CK_ResultType");
971         ReturnType = Chunk.Text;
972         break;
973       case CodeCompletionString::CK_CurrentParameter:
974       case CodeCompletionString::CK_Placeholder:
975         processParameterChunk(Chunk.Text, Signature);
976         Signal.NumberOfParameters++;
977         break;
978       case CodeCompletionString::CK_Optional: {
979         // The rest of the parameters are defaulted/optional.
980         assert(Chunk.Optional &&
981                "Expected the optional code completion string to be non-null.");
982         processOptionalChunk(*Chunk.Optional, Signature, Signal);
983         break;
984       }
985       case CodeCompletionString::CK_VerticalSpace:
986         break;
987       default:
988         Signature.label += Chunk.Text;
989         break;
990       }
991     }
992     if (ReturnType) {
993       Signature.label += " -> ";
994       Signature.label += ReturnType;
995     }
996     dlog("Signal for {0}: {1}", Signature, Signal);
997     ScoredSignature Result;
998     Result.Signature = std::move(Signature);
999     Result.Quality = Signal;
1000     Result.IDForDoc =
1001         Result.Signature.documentation.empty() && Candidate.getFunction()
1002             ? clangd::getSymbolID(Candidate.getFunction())
1003             : None;
1004     return Result;
1005   }
1006 
1007   SignatureHelp &SigHelp;
1008   std::shared_ptr<clang::GlobalCodeCompletionAllocator> Allocator;
1009   CodeCompletionTUInfo CCTUInfo;
1010   const SymbolIndex *Index;
1011 }; // SignatureHelpCollector
1012 
1013 struct SemaCompleteInput {
1014   PathRef FileName;
1015   const tooling::CompileCommand &Command;
1016   const PreambleData *Preamble;
1017   llvm::StringRef Contents;
1018   size_t Offset;
1019   llvm::IntrusiveRefCntPtr<llvm::vfs::FileSystem> VFS;
1020 };
1021 
1022 void loadMainFilePreambleMacros(const Preprocessor &PP,
1023                                 const PreambleData &Preamble) {
1024   // The ExternalPreprocessorSource has our macros, if we know where to look.
1025   // We can read all the macros using PreambleMacros->ReadDefinedMacros(),
1026   // but this includes transitively included files, so may deserialize a lot.
1027   ExternalPreprocessorSource *PreambleMacros = PP.getExternalSource();
1028   // As we have the names of the macros, we can look up their IdentifierInfo
1029   // and then use this to load just the macros we want.
1030   IdentifierInfoLookup *PreambleIdentifiers =
1031       PP.getIdentifierTable().getExternalIdentifierLookup();
1032   if (!PreambleIdentifiers || !PreambleMacros)
1033     return;
1034   for (const auto &MacroName : Preamble.MainFileMacros)
1035     if (auto *II = PreambleIdentifiers->get(MacroName))
1036       if (II->isOutOfDate())
1037         PreambleMacros->updateOutOfDateIdentifier(*II);
1038 }
1039 
1040 // Invokes Sema code completion on a file.
1041 // If \p Includes is set, it will be updated based on the compiler invocation.
1042 bool semaCodeComplete(std::unique_ptr<CodeCompleteConsumer> Consumer,
1043                       const clang::CodeCompleteOptions &Options,
1044                       const SemaCompleteInput &Input,
1045                       IncludeStructure *Includes = nullptr) {
1046   trace::Span Tracer("Sema completion");
1047   llvm::IntrusiveRefCntPtr<llvm::vfs::FileSystem> VFS = Input.VFS;
1048   if (Input.Preamble && Input.Preamble->StatCache)
1049     VFS = Input.Preamble->StatCache->getConsumingFS(std::move(VFS));
1050   ParseInputs ParseInput;
1051   ParseInput.CompileCommand = Input.Command;
1052   ParseInput.FS = VFS;
1053   ParseInput.Contents = Input.Contents;
1054   ParseInput.Opts = ParseOptions();
1055   auto CI = buildCompilerInvocation(ParseInput);
1056   if (!CI) {
1057     elog("Couldn't create CompilerInvocation");
1058     return false;
1059   }
1060   auto &FrontendOpts = CI->getFrontendOpts();
1061   FrontendOpts.SkipFunctionBodies = true;
1062   // Disable typo correction in Sema.
1063   CI->getLangOpts()->SpellChecking = false;
1064   // Setup code completion.
1065   FrontendOpts.CodeCompleteOpts = Options;
1066   FrontendOpts.CodeCompletionAt.FileName = Input.FileName;
1067   std::tie(FrontendOpts.CodeCompletionAt.Line,
1068            FrontendOpts.CodeCompletionAt.Column) =
1069       offsetToClangLineColumn(Input.Contents, Input.Offset);
1070 
1071   std::unique_ptr<llvm::MemoryBuffer> ContentsBuffer =
1072       llvm::MemoryBuffer::getMemBufferCopy(Input.Contents, Input.FileName);
1073   // The diagnostic options must be set before creating a CompilerInstance.
1074   CI->getDiagnosticOpts().IgnoreWarnings = true;
1075   // We reuse the preamble whether it's valid or not. This is a
1076   // correctness/performance tradeoff: building without a preamble is slow, and
1077   // completion is latency-sensitive.
1078   // However, if we're completing *inside* the preamble section of the draft,
1079   // overriding the preamble will break sema completion. Fortunately we can just
1080   // skip all includes in this case; these completions are really simple.
1081   PreambleBounds PreambleRegion =
1082       ComputePreambleBounds(*CI->getLangOpts(), ContentsBuffer.get(), 0);
1083   bool CompletingInPreamble = PreambleRegion.Size > Input.Offset;
1084   // NOTE: we must call BeginSourceFile after prepareCompilerInstance. Otherwise
1085   // the remapped buffers do not get freed.
1086   IgnoreDiagnostics DummyDiagsConsumer;
1087   auto Clang = prepareCompilerInstance(
1088       std::move(CI),
1089       (Input.Preamble && !CompletingInPreamble) ? &Input.Preamble->Preamble
1090                                                 : nullptr,
1091       std::move(ContentsBuffer), std::move(VFS), DummyDiagsConsumer);
1092   Clang->getPreprocessorOpts().SingleFileParseMode = CompletingInPreamble;
1093   Clang->setCodeCompletionConsumer(Consumer.release());
1094 
1095   SyntaxOnlyAction Action;
1096   if (!Action.BeginSourceFile(*Clang, Clang->getFrontendOpts().Inputs[0])) {
1097     log("BeginSourceFile() failed when running codeComplete for {0}",
1098         Input.FileName);
1099     return false;
1100   }
1101   // Macros can be defined within the preamble region of the main file.
1102   // They don't fall nicely into our index/Sema dichotomy:
1103   //  - they're not indexed for completion (they're not available across files)
1104   //  - but Sema code complete won't see them: as part of the preamble, they're
1105   //    deserialized only when mentioned.
1106   // Force them to be deserialized so SemaCodeComplete sees them.
1107   if (Input.Preamble)
1108     loadMainFilePreambleMacros(Clang->getPreprocessor(), *Input.Preamble);
1109   if (Includes)
1110     Clang->getPreprocessor().addPPCallbacks(
1111         collectIncludeStructureCallback(Clang->getSourceManager(), Includes));
1112   if (llvm::Error Err = Action.Execute()) {
1113     log("Execute() failed when running codeComplete for {0}: {1}",
1114         Input.FileName, toString(std::move(Err)));
1115     return false;
1116   }
1117   Action.EndSourceFile();
1118 
1119   return true;
1120 }
1121 
1122 // Should we allow index completions in the specified context?
1123 bool allowIndex(CodeCompletionContext &CC) {
1124   if (!contextAllowsIndex(CC.getKind()))
1125     return false;
1126   // We also avoid ClassName::bar (but allow namespace::bar).
1127   auto Scope = CC.getCXXScopeSpecifier();
1128   if (!Scope)
1129     return true;
1130   NestedNameSpecifier *NameSpec = (*Scope)->getScopeRep();
1131   if (!NameSpec)
1132     return true;
1133   // We only query the index when qualifier is a namespace.
1134   // If it's a class, we rely solely on sema completions.
1135   switch (NameSpec->getKind()) {
1136   case NestedNameSpecifier::Global:
1137   case NestedNameSpecifier::Namespace:
1138   case NestedNameSpecifier::NamespaceAlias:
1139     return true;
1140   case NestedNameSpecifier::Super:
1141   case NestedNameSpecifier::TypeSpec:
1142   case NestedNameSpecifier::TypeSpecWithTemplate:
1143   // Unresolved inside a template.
1144   case NestedNameSpecifier::Identifier:
1145     return false;
1146   }
1147   llvm_unreachable("invalid NestedNameSpecifier kind");
1148 }
1149 
1150 std::future<SymbolSlab> startAsyncFuzzyFind(const SymbolIndex &Index,
1151                                             const FuzzyFindRequest &Req) {
1152   return runAsync<SymbolSlab>([&Index, Req]() {
1153     trace::Span Tracer("Async fuzzyFind");
1154     SymbolSlab::Builder Syms;
1155     Index.fuzzyFind(Req, [&Syms](const Symbol &Sym) { Syms.insert(Sym); });
1156     return std::move(Syms).build();
1157   });
1158 }
1159 
1160 // Creates a `FuzzyFindRequest` based on the cached index request from the
1161 // last completion, if any, and the speculated completion filter text in the
1162 // source code.
1163 FuzzyFindRequest speculativeFuzzyFindRequestForCompletion(
1164     FuzzyFindRequest CachedReq, const CompletionPrefix &HeuristicPrefix) {
1165   CachedReq.Query = HeuristicPrefix.Name;
1166   return CachedReq;
1167 }
1168 
1169 // Runs Sema-based (AST) and Index-based completion, returns merged results.
1170 //
1171 // There are a few tricky considerations:
1172 //   - the AST provides information needed for the index query (e.g. which
1173 //     namespaces to search in). So Sema must start first.
1174 //   - we only want to return the top results (Opts.Limit).
1175 //     Building CompletionItems for everything else is wasteful, so we want to
1176 //     preserve the "native" format until we're done with scoring.
1177 //   - the data underlying Sema completion items is owned by the AST and various
1178 //     other arenas, which must stay alive for us to build CompletionItems.
1179 //   - we may get duplicate results from Sema and the Index, we need to merge.
1180 //
1181 // So we start Sema completion first, and do all our work in its callback.
1182 // We use the Sema context information to query the index.
1183 // Then we merge the two result sets, producing items that are Sema/Index/Both.
1184 // These items are scored, and the top N are synthesized into the LSP response.
1185 // Finally, we can clean up the data structures created by Sema completion.
1186 //
1187 // Main collaborators are:
1188 //   - semaCodeComplete sets up the compiler machinery to run code completion.
1189 //   - CompletionRecorder captures Sema completion results, including context.
1190 //   - SymbolIndex (Opts.Index) provides index completion results as Symbols
1191 //   - CompletionCandidates are the result of merging Sema and Index results.
1192 //     Each candidate points to an underlying CodeCompletionResult (Sema), a
1193 //     Symbol (Index), or both. It computes the result quality score.
1194 //     CompletionCandidate also does conversion to CompletionItem (at the end).
1195 //   - FuzzyMatcher scores how the candidate matches the partial identifier.
1196 //     This score is combined with the result quality score for the final score.
1197 //   - TopN determines the results with the best score.
1198 class CodeCompleteFlow {
1199   PathRef FileName;
1200   IncludeStructure Includes;           // Complete once the compiler runs.
1201   SpeculativeFuzzyFind *SpecFuzzyFind; // Can be nullptr.
1202   const CodeCompleteOptions &Opts;
1203 
1204   // Sema takes ownership of Recorder. Recorder is valid until Sema cleanup.
1205   CompletionRecorder *Recorder = nullptr;
1206   CodeCompletionContext::Kind CCContextKind = CodeCompletionContext::CCC_Other;
1207   // Counters for logging.
1208   int NSema = 0, NIndex = 0, NSemaAndIndex = 0, NIdent = 0;
1209   bool Incomplete = false; // Would more be available with a higher limit?
1210   CompletionPrefix HeuristicPrefix;
1211   llvm::Optional<FuzzyMatcher> Filter; // Initialized once Sema runs.
1212   Range ReplacedRange;
1213   std::vector<std::string> QueryScopes; // Initialized once Sema runs.
1214   // Initialized once QueryScopes is initialized, if there are scopes.
1215   llvm::Optional<ScopeDistance> ScopeProximity;
1216   llvm::Optional<OpaqueType> PreferredType; // Initialized once Sema runs.
1217   // Whether to query symbols from any scope. Initialized once Sema runs.
1218   bool AllScopes = false;
1219   llvm::StringSet<> ContextWords;
1220   // Include-insertion and proximity scoring rely on the include structure.
1221   // This is available after Sema has run.
1222   llvm::Optional<IncludeInserter> Inserter;  // Available during runWithSema.
1223   llvm::Optional<URIDistance> FileProximity; // Initialized once Sema runs.
1224   /// Speculative request based on the cached request and the filter text before
1225   /// the cursor.
1226   /// Initialized right before sema run. This is only set if `SpecFuzzyFind` is
1227   /// set and contains a cached request.
1228   llvm::Optional<FuzzyFindRequest> SpecReq;
1229 
1230 public:
1231   // A CodeCompleteFlow object is only useful for calling run() exactly once.
1232   CodeCompleteFlow(PathRef FileName, const IncludeStructure &Includes,
1233                    SpeculativeFuzzyFind *SpecFuzzyFind,
1234                    const CodeCompleteOptions &Opts)
1235       : FileName(FileName), Includes(Includes), SpecFuzzyFind(SpecFuzzyFind),
1236         Opts(Opts) {}
1237 
1238   CodeCompleteResult run(const SemaCompleteInput &SemaCCInput) && {
1239     trace::Span Tracer("CodeCompleteFlow");
1240     HeuristicPrefix =
1241         guessCompletionPrefix(SemaCCInput.Contents, SemaCCInput.Offset);
1242     populateContextWords(SemaCCInput.Contents);
1243     if (Opts.Index && SpecFuzzyFind && SpecFuzzyFind->CachedReq.hasValue()) {
1244       assert(!SpecFuzzyFind->Result.valid());
1245       SpecReq = speculativeFuzzyFindRequestForCompletion(
1246           *SpecFuzzyFind->CachedReq, HeuristicPrefix);
1247       SpecFuzzyFind->Result = startAsyncFuzzyFind(*Opts.Index, *SpecReq);
1248     }
1249 
1250     // We run Sema code completion first. It builds an AST and calculates:
1251     //   - completion results based on the AST.
1252     //   - partial identifier and context. We need these for the index query.
1253     CodeCompleteResult Output;
1254     auto RecorderOwner = llvm::make_unique<CompletionRecorder>(Opts, [&]() {
1255       assert(Recorder && "Recorder is not set");
1256       CCContextKind = Recorder->CCContext.getKind();
1257       auto Style = getFormatStyleForFile(
1258           SemaCCInput.FileName, SemaCCInput.Contents, SemaCCInput.VFS.get());
1259       // If preprocessor was run, inclusions from preprocessor callback should
1260       // already be added to Includes.
1261       Inserter.emplace(
1262           SemaCCInput.FileName, SemaCCInput.Contents, Style,
1263           SemaCCInput.Command.Directory,
1264           &Recorder->CCSema->getPreprocessor().getHeaderSearchInfo());
1265       for (const auto &Inc : Includes.MainFileIncludes)
1266         Inserter->addExisting(Inc);
1267 
1268       // Most of the cost of file proximity is in initializing the FileDistance
1269       // structures based on the observed includes, once per query. Conceptually
1270       // that happens here (though the per-URI-scheme initialization is lazy).
1271       // The per-result proximity scoring is (amortized) very cheap.
1272       FileDistanceOptions ProxOpts{}; // Use defaults.
1273       const auto &SM = Recorder->CCSema->getSourceManager();
1274       llvm::StringMap<SourceParams> ProxSources;
1275       for (auto &Entry : Includes.includeDepth(
1276                SM.getFileEntryForID(SM.getMainFileID())->getName())) {
1277         auto &Source = ProxSources[Entry.getKey()];
1278         Source.Cost = Entry.getValue() * ProxOpts.IncludeCost;
1279         // Symbols near our transitive includes are good, but only consider
1280         // things in the same directory or below it. Otherwise there can be
1281         // many false positives.
1282         if (Entry.getValue() > 0)
1283           Source.MaxUpTraversals = 1;
1284       }
1285       FileProximity.emplace(ProxSources, ProxOpts);
1286 
1287       Output = runWithSema();
1288       Inserter.reset(); // Make sure this doesn't out-live Clang.
1289       SPAN_ATTACH(Tracer, "sema_completion_kind",
1290                   getCompletionKindString(CCContextKind));
1291       log("Code complete: sema context {0}, query scopes [{1}] (AnyScope={2}), "
1292           "expected type {3}",
1293           getCompletionKindString(CCContextKind),
1294           llvm::join(QueryScopes.begin(), QueryScopes.end(), ","), AllScopes,
1295           PreferredType ? Recorder->CCContext.getPreferredType().getAsString()
1296                         : "<none>");
1297     });
1298 
1299     Recorder = RecorderOwner.get();
1300 
1301     semaCodeComplete(std::move(RecorderOwner), Opts.getClangCompleteOpts(),
1302                      SemaCCInput, &Includes);
1303     logResults(Output, Tracer);
1304     return Output;
1305   }
1306 
1307   void logResults(const CodeCompleteResult &Output, const trace::Span &Tracer) {
1308     SPAN_ATTACH(Tracer, "sema_results", NSema);
1309     SPAN_ATTACH(Tracer, "index_results", NIndex);
1310     SPAN_ATTACH(Tracer, "merged_results", NSemaAndIndex);
1311     SPAN_ATTACH(Tracer, "identifier_results", NIdent);
1312     SPAN_ATTACH(Tracer, "returned_results", int64_t(Output.Completions.size()));
1313     SPAN_ATTACH(Tracer, "incomplete", Output.HasMore);
1314     log("Code complete: {0} results from Sema, {1} from Index, "
1315         "{2} matched, {3} from identifiers, {4} returned{5}.",
1316         NSema, NIndex, NSemaAndIndex, NIdent, Output.Completions.size(),
1317         Output.HasMore ? " (incomplete)" : "");
1318     assert(!Opts.Limit || Output.Completions.size() <= Opts.Limit);
1319     // We don't assert that isIncomplete means we hit a limit.
1320     // Indexes may choose to impose their own limits even if we don't have one.
1321   }
1322 
1323   CodeCompleteResult
1324   runWithoutSema(llvm::StringRef Content, size_t Offset,
1325                  llvm::IntrusiveRefCntPtr<llvm::vfs::FileSystem> VFS) && {
1326     trace::Span Tracer("CodeCompleteWithoutSema");
1327     // Fill in fields normally set by runWithSema()
1328     HeuristicPrefix = guessCompletionPrefix(Content, Offset);
1329     populateContextWords(Content);
1330     CCContextKind = CodeCompletionContext::CCC_Recovery;
1331     Filter = FuzzyMatcher(HeuristicPrefix.Name);
1332     auto Pos = offsetToPosition(Content, Offset);
1333     ReplacedRange.start = ReplacedRange.end = Pos;
1334     ReplacedRange.start.character -= HeuristicPrefix.Name.size();
1335 
1336     llvm::StringMap<SourceParams> ProxSources;
1337     ProxSources[FileName].Cost = 0;
1338     FileProximity.emplace(ProxSources);
1339 
1340     auto Style = getFormatStyleForFile(FileName, Content, VFS.get());
1341     // This will only insert verbatim headers.
1342     Inserter.emplace(FileName, Content, Style,
1343                      /*BuildDir=*/"", /*HeaderSearchInfo=*/nullptr);
1344 
1345     auto Identifiers = collectIdentifiers(Content, Style);
1346     std::vector<RawIdentifier> IdentifierResults;
1347     for (const auto &IDAndCount : Identifiers) {
1348       RawIdentifier ID;
1349       ID.Name = IDAndCount.first();
1350       ID.References = IDAndCount.second;
1351       // Avoid treating typed filter as an identifier.
1352       if (ID.Name == HeuristicPrefix.Name)
1353         --ID.References;
1354       if (ID.References > 0)
1355         IdentifierResults.push_back(std::move(ID));
1356     }
1357 
1358     // Simplified version of getQueryScopes():
1359     //  - accessible scopes are determined heuristically.
1360     //  - all-scopes query if no qualifier was typed (and it's allowed).
1361     SpecifiedScope Scopes;
1362     Scopes.AccessibleScopes =
1363         visibleNamespaces(Content.take_front(Offset), Style);
1364     for (std::string &S : Scopes.AccessibleScopes)
1365       if (!S.empty())
1366         S.append("::"); // visibleNamespaces doesn't include trailing ::.
1367     if (HeuristicPrefix.Qualifier.empty())
1368       AllScopes = Opts.AllScopes;
1369     else if (HeuristicPrefix.Qualifier.startswith("::")) {
1370       Scopes.AccessibleScopes = {""};
1371       Scopes.UnresolvedQualifier = HeuristicPrefix.Qualifier.drop_front(2);
1372     } else
1373       Scopes.UnresolvedQualifier = HeuristicPrefix.Qualifier;
1374     // First scope is the (modified) enclosing scope.
1375     QueryScopes = Scopes.scopesForIndexQuery();
1376     ScopeProximity.emplace(QueryScopes);
1377 
1378     SymbolSlab IndexResults = Opts.Index ? queryIndex() : SymbolSlab();
1379 
1380     CodeCompleteResult Output = toCodeCompleteResult(mergeResults(
1381         /*SemaResults=*/{}, IndexResults, IdentifierResults));
1382     Output.RanParser = false;
1383     logResults(Output, Tracer);
1384     return Output;
1385   }
1386 
1387 private:
1388   void populateContextWords(llvm::StringRef Content) {
1389     // Take last 3 lines before the completion point.
1390     unsigned RangeEnd = HeuristicPrefix.Qualifier.begin() - Content.data(),
1391              RangeBegin = RangeEnd;
1392     for (size_t I = 0; I < 3 && RangeBegin > 0; ++I) {
1393       auto PrevNL = Content.rfind('\n', RangeBegin - 1);
1394       if (PrevNL == StringRef::npos) {
1395         RangeBegin = 0;
1396         break;
1397       }
1398       RangeBegin = PrevNL + 1;
1399     }
1400 
1401     ContextWords = collectWords(Content.slice(RangeBegin, RangeEnd));
1402     dlog("Completion context words: {0}",
1403          llvm::join(ContextWords.keys(), ", "));
1404   }
1405 
1406   // This is called by run() once Sema code completion is done, but before the
1407   // Sema data structures are torn down. It does all the real work.
1408   CodeCompleteResult runWithSema() {
1409     const auto &CodeCompletionRange = CharSourceRange::getCharRange(
1410         Recorder->CCSema->getPreprocessor().getCodeCompletionTokenRange());
1411     // When we are getting completions with an empty identifier, for example
1412     //    std::vector<int> asdf;
1413     //    asdf.^;
1414     // Then the range will be invalid and we will be doing insertion, use
1415     // current cursor position in such cases as range.
1416     if (CodeCompletionRange.isValid()) {
1417       ReplacedRange = halfOpenToRange(Recorder->CCSema->getSourceManager(),
1418                                       CodeCompletionRange);
1419     } else {
1420       const auto &Pos = sourceLocToPosition(
1421           Recorder->CCSema->getSourceManager(),
1422           Recorder->CCSema->getPreprocessor().getCodeCompletionLoc());
1423       ReplacedRange.start = ReplacedRange.end = Pos;
1424     }
1425     Filter = FuzzyMatcher(
1426         Recorder->CCSema->getPreprocessor().getCodeCompletionFilter());
1427     std::tie(QueryScopes, AllScopes) = getQueryScopes(
1428         Recorder->CCContext, *Recorder->CCSema, HeuristicPrefix, Opts);
1429     if (!QueryScopes.empty())
1430       ScopeProximity.emplace(QueryScopes);
1431     PreferredType =
1432         OpaqueType::fromType(Recorder->CCSema->getASTContext(),
1433                              Recorder->CCContext.getPreferredType());
1434     // Sema provides the needed context to query the index.
1435     // FIXME: in addition to querying for extra/overlapping symbols, we should
1436     //        explicitly request symbols corresponding to Sema results.
1437     //        We can use their signals even if the index can't suggest them.
1438     // We must copy index results to preserve them, but there are at most Limit.
1439     auto IndexResults = (Opts.Index && allowIndex(Recorder->CCContext))
1440                             ? queryIndex()
1441                             : SymbolSlab();
1442     trace::Span Tracer("Populate CodeCompleteResult");
1443     // Merge Sema and Index results, score them, and pick the winners.
1444     auto Top =
1445         mergeResults(Recorder->Results, IndexResults, /*Identifiers*/ {});
1446     return toCodeCompleteResult(Top);
1447   }
1448 
1449   CodeCompleteResult
1450   toCodeCompleteResult(const std::vector<ScoredBundle> &Scored) {
1451     CodeCompleteResult Output;
1452 
1453     // Convert the results to final form, assembling the expensive strings.
1454     for (auto &C : Scored) {
1455       Output.Completions.push_back(toCodeCompletion(C.first));
1456       Output.Completions.back().Score = C.second;
1457       Output.Completions.back().CompletionTokenRange = ReplacedRange;
1458     }
1459     Output.HasMore = Incomplete;
1460     Output.Context = CCContextKind;
1461     return Output;
1462   }
1463 
1464   SymbolSlab queryIndex() {
1465     trace::Span Tracer("Query index");
1466     SPAN_ATTACH(Tracer, "limit", int64_t(Opts.Limit));
1467 
1468     // Build the query.
1469     FuzzyFindRequest Req;
1470     if (Opts.Limit)
1471       Req.Limit = Opts.Limit;
1472     Req.Query = Filter->pattern();
1473     Req.RestrictForCodeCompletion = true;
1474     Req.Scopes = QueryScopes;
1475     Req.AnyScope = AllScopes;
1476     // FIXME: we should send multiple weighted paths here.
1477     Req.ProximityPaths.push_back(FileName);
1478     if (PreferredType)
1479       Req.PreferredTypes.push_back(PreferredType->raw());
1480     vlog("Code complete: fuzzyFind({0:2})", toJSON(Req));
1481 
1482     if (SpecFuzzyFind)
1483       SpecFuzzyFind->NewReq = Req;
1484     if (SpecFuzzyFind && SpecFuzzyFind->Result.valid() && (*SpecReq == Req)) {
1485       vlog("Code complete: speculative fuzzy request matches the actual index "
1486            "request. Waiting for the speculative index results.");
1487       SPAN_ATTACH(Tracer, "Speculative results", true);
1488 
1489       trace::Span WaitSpec("Wait speculative results");
1490       return SpecFuzzyFind->Result.get();
1491     }
1492 
1493     SPAN_ATTACH(Tracer, "Speculative results", false);
1494 
1495     // Run the query against the index.
1496     SymbolSlab::Builder ResultsBuilder;
1497     if (Opts.Index->fuzzyFind(
1498             Req, [&](const Symbol &Sym) { ResultsBuilder.insert(Sym); }))
1499       Incomplete = true;
1500     return std::move(ResultsBuilder).build();
1501   }
1502 
1503   // Merges Sema and Index results where possible, to form CompletionCandidates.
1504   // \p Identifiers is raw idenfiers that can also be completion condidates.
1505   // Identifiers are not merged with results from index or sema.
1506   // Groups overloads if desired, to form CompletionCandidate::Bundles. The
1507   // bundles are scored and top results are returned, best to worst.
1508   std::vector<ScoredBundle>
1509   mergeResults(const std::vector<CodeCompletionResult> &SemaResults,
1510                const SymbolSlab &IndexResults,
1511                const std::vector<RawIdentifier> &IdentifierResults) {
1512     trace::Span Tracer("Merge and score results");
1513     std::vector<CompletionCandidate::Bundle> Bundles;
1514     llvm::DenseMap<size_t, size_t> BundleLookup;
1515     auto AddToBundles = [&](const CodeCompletionResult *SemaResult,
1516                             const Symbol *IndexResult,
1517                             const RawIdentifier *IdentifierResult) {
1518       CompletionCandidate C;
1519       C.SemaResult = SemaResult;
1520       C.IndexResult = IndexResult;
1521       C.IdentifierResult = IdentifierResult;
1522       if (C.IndexResult) {
1523         C.Name = IndexResult->Name;
1524         C.RankedIncludeHeaders = getRankedIncludes(*C.IndexResult);
1525       } else if (C.SemaResult) {
1526         C.Name = Recorder->getName(*SemaResult);
1527       } else {
1528         assert(IdentifierResult);
1529         C.Name = IdentifierResult->Name;
1530       }
1531       if (auto OverloadSet = C.overloadSet(Opts)) {
1532         auto Ret = BundleLookup.try_emplace(OverloadSet, Bundles.size());
1533         if (Ret.second)
1534           Bundles.emplace_back();
1535         Bundles[Ret.first->second].push_back(std::move(C));
1536       } else {
1537         Bundles.emplace_back();
1538         Bundles.back().push_back(std::move(C));
1539       }
1540     };
1541     llvm::DenseSet<const Symbol *> UsedIndexResults;
1542     auto CorrespondingIndexResult =
1543         [&](const CodeCompletionResult &SemaResult) -> const Symbol * {
1544       if (auto SymID =
1545               getSymbolID(SemaResult, Recorder->CCSema->getSourceManager())) {
1546         auto I = IndexResults.find(*SymID);
1547         if (I != IndexResults.end()) {
1548           UsedIndexResults.insert(&*I);
1549           return &*I;
1550         }
1551       }
1552       return nullptr;
1553     };
1554     // Emit all Sema results, merging them with Index results if possible.
1555     for (auto &SemaResult : SemaResults)
1556       AddToBundles(&SemaResult, CorrespondingIndexResult(SemaResult), nullptr);
1557     // Now emit any Index-only results.
1558     for (const auto &IndexResult : IndexResults) {
1559       if (UsedIndexResults.count(&IndexResult))
1560         continue;
1561       AddToBundles(/*SemaResult=*/nullptr, &IndexResult, nullptr);
1562     }
1563     // Emit identifier results.
1564     for (const auto &Ident : IdentifierResults)
1565       AddToBundles(/*SemaResult=*/nullptr, /*IndexResult=*/nullptr, &Ident);
1566     // We only keep the best N results at any time, in "native" format.
1567     TopN<ScoredBundle, ScoredBundleGreater> Top(
1568         Opts.Limit == 0 ? std::numeric_limits<size_t>::max() : Opts.Limit);
1569     for (auto &Bundle : Bundles)
1570       addCandidate(Top, std::move(Bundle));
1571     return std::move(Top).items();
1572   }
1573 
1574   llvm::Optional<float> fuzzyScore(const CompletionCandidate &C) {
1575     // Macros can be very spammy, so we only support prefix completion.
1576     // We won't end up with underfull index results, as macros are sema-only.
1577     if (C.SemaResult && C.SemaResult->Kind == CodeCompletionResult::RK_Macro &&
1578         !C.Name.startswith_lower(Filter->pattern()))
1579       return None;
1580     return Filter->match(C.Name);
1581   }
1582 
1583   // Scores a candidate and adds it to the TopN structure.
1584   void addCandidate(TopN<ScoredBundle, ScoredBundleGreater> &Candidates,
1585                     CompletionCandidate::Bundle Bundle) {
1586     SymbolQualitySignals Quality;
1587     SymbolRelevanceSignals Relevance;
1588     Relevance.Context = CCContextKind;
1589     Relevance.Name = Bundle.front().Name;
1590     Relevance.Query = SymbolRelevanceSignals::CodeComplete;
1591     Relevance.FileProximityMatch = FileProximity.getPointer();
1592     if (ScopeProximity)
1593       Relevance.ScopeProximityMatch = ScopeProximity.getPointer();
1594     if (PreferredType)
1595       Relevance.HadContextType = true;
1596     Relevance.ContextWords = &ContextWords;
1597 
1598     auto &First = Bundle.front();
1599     if (auto FuzzyScore = fuzzyScore(First))
1600       Relevance.NameMatch = *FuzzyScore;
1601     else
1602       return;
1603     SymbolOrigin Origin = SymbolOrigin::Unknown;
1604     bool FromIndex = false;
1605     for (const auto &Candidate : Bundle) {
1606       if (Candidate.IndexResult) {
1607         Quality.merge(*Candidate.IndexResult);
1608         Relevance.merge(*Candidate.IndexResult);
1609         Origin |= Candidate.IndexResult->Origin;
1610         FromIndex = true;
1611         if (!Candidate.IndexResult->Type.empty())
1612           Relevance.HadSymbolType |= true;
1613         if (PreferredType &&
1614             PreferredType->raw() == Candidate.IndexResult->Type) {
1615           Relevance.TypeMatchesPreferred = true;
1616         }
1617       }
1618       if (Candidate.SemaResult) {
1619         Quality.merge(*Candidate.SemaResult);
1620         Relevance.merge(*Candidate.SemaResult);
1621         if (PreferredType) {
1622           if (auto CompletionType = OpaqueType::fromCompletionResult(
1623                   Recorder->CCSema->getASTContext(), *Candidate.SemaResult)) {
1624             Relevance.HadSymbolType |= true;
1625             if (PreferredType == CompletionType)
1626               Relevance.TypeMatchesPreferred = true;
1627           }
1628         }
1629         Origin |= SymbolOrigin::AST;
1630       }
1631       if (Candidate.IdentifierResult) {
1632         Quality.References = Candidate.IdentifierResult->References;
1633         Relevance.Scope = SymbolRelevanceSignals::FileScope;
1634         Origin |= SymbolOrigin::Identifier;
1635       }
1636     }
1637 
1638     CodeCompletion::Scores Scores;
1639     Scores.Quality = Quality.evaluate();
1640     Scores.Relevance = Relevance.evaluate();
1641     Scores.Total = evaluateSymbolAndRelevance(Scores.Quality, Scores.Relevance);
1642     // NameMatch is in fact a multiplier on total score, so rescoring is sound.
1643     Scores.ExcludingName = Relevance.NameMatch
1644                                ? Scores.Total / Relevance.NameMatch
1645                                : Scores.Quality;
1646 
1647     dlog("CodeComplete: {0} ({1}) = {2}\n{3}{4}\n", First.Name,
1648          llvm::to_string(Origin), Scores.Total, llvm::to_string(Quality),
1649          llvm::to_string(Relevance));
1650 
1651     NSema += bool(Origin & SymbolOrigin::AST);
1652     NIndex += FromIndex;
1653     NSemaAndIndex += bool(Origin & SymbolOrigin::AST) && FromIndex;
1654     NIdent += bool(Origin & SymbolOrigin::Identifier);
1655     if (Candidates.push({std::move(Bundle), Scores}))
1656       Incomplete = true;
1657   }
1658 
1659   CodeCompletion toCodeCompletion(const CompletionCandidate::Bundle &Bundle) {
1660     llvm::Optional<CodeCompletionBuilder> Builder;
1661     for (const auto &Item : Bundle) {
1662       CodeCompletionString *SemaCCS =
1663           Item.SemaResult ? Recorder->codeCompletionString(*Item.SemaResult)
1664                           : nullptr;
1665       if (!Builder)
1666         Builder.emplace(Recorder ? &Recorder->CCSema->getASTContext() : nullptr,
1667                         Item, SemaCCS, QueryScopes, *Inserter, FileName,
1668                         CCContextKind, Opts);
1669       else
1670         Builder->add(Item, SemaCCS);
1671     }
1672     return Builder->build();
1673   }
1674 };
1675 
1676 template <class T> bool isExplicitTemplateSpecialization(const NamedDecl &ND) {
1677   if (const auto *TD = dyn_cast<T>(&ND))
1678     if (TD->getTemplateSpecializationKind() == TSK_ExplicitSpecialization)
1679       return true;
1680   return false;
1681 }
1682 
1683 } // namespace
1684 
1685 clang::CodeCompleteOptions CodeCompleteOptions::getClangCompleteOpts() const {
1686   clang::CodeCompleteOptions Result;
1687   Result.IncludeCodePatterns = EnableSnippets && IncludeCodePatterns;
1688   Result.IncludeMacros = IncludeMacros;
1689   Result.IncludeGlobals = true;
1690   // We choose to include full comments and not do doxygen parsing in
1691   // completion.
1692   // FIXME: ideally, we should support doxygen in some form, e.g. do markdown
1693   // formatting of the comments.
1694   Result.IncludeBriefComments = false;
1695 
1696   // When an is used, Sema is responsible for completing the main file,
1697   // the index can provide results from the preamble.
1698   // Tell Sema not to deserialize the preamble to look for results.
1699   Result.LoadExternal = !Index;
1700   Result.IncludeFixIts = IncludeFixIts;
1701 
1702   return Result;
1703 }
1704 
1705 CompletionPrefix guessCompletionPrefix(llvm::StringRef Content,
1706                                        unsigned Offset) {
1707   assert(Offset <= Content.size());
1708   StringRef Rest = Content.take_front(Offset);
1709   CompletionPrefix Result;
1710 
1711   // Consume the unqualified name. We only handle ASCII characters.
1712   // isIdentifierBody will let us match "0invalid", but we don't mind.
1713   while (!Rest.empty() && isIdentifierBody(Rest.back()))
1714     Rest = Rest.drop_back();
1715   Result.Name = Content.slice(Rest.size(), Offset);
1716 
1717   // Consume qualifiers.
1718   while (Rest.consume_back("::") && !Rest.endswith(":")) // reject ::::
1719     while (!Rest.empty() && isIdentifierBody(Rest.back()))
1720       Rest = Rest.drop_back();
1721   Result.Qualifier =
1722       Content.slice(Rest.size(), Result.Name.begin() - Content.begin());
1723 
1724   return Result;
1725 }
1726 
1727 CodeCompleteResult
1728 codeComplete(PathRef FileName, const tooling::CompileCommand &Command,
1729              const PreambleData *Preamble, llvm::StringRef Contents,
1730              Position Pos, llvm::IntrusiveRefCntPtr<llvm::vfs::FileSystem> VFS,
1731              CodeCompleteOptions Opts, SpeculativeFuzzyFind *SpecFuzzyFind) {
1732   auto Offset = positionToOffset(Contents, Pos);
1733   if (!Offset) {
1734     elog("Code completion position was invalid {0}", Offset.takeError());
1735     return CodeCompleteResult();
1736   }
1737   auto Flow = CodeCompleteFlow(
1738       FileName, Preamble ? Preamble->Includes : IncludeStructure(),
1739       SpecFuzzyFind, Opts);
1740   return (!Preamble || Opts.RunParser == CodeCompleteOptions::NeverParse)
1741              ? std::move(Flow).runWithoutSema(Contents, *Offset, VFS)
1742              : std::move(Flow).run(
1743                    {FileName, Command, Preamble, Contents, *Offset, VFS});
1744 }
1745 
1746 SignatureHelp signatureHelp(PathRef FileName,
1747                             const tooling::CompileCommand &Command,
1748                             const PreambleData *Preamble,
1749                             llvm::StringRef Contents, Position Pos,
1750                             llvm::IntrusiveRefCntPtr<llvm::vfs::FileSystem> VFS,
1751                             const SymbolIndex *Index) {
1752   auto Offset = positionToOffset(Contents, Pos);
1753   if (!Offset) {
1754     elog("Code completion position was invalid {0}", Offset.takeError());
1755     return SignatureHelp();
1756   }
1757   SignatureHelp Result;
1758   clang::CodeCompleteOptions Options;
1759   Options.IncludeGlobals = false;
1760   Options.IncludeMacros = false;
1761   Options.IncludeCodePatterns = false;
1762   Options.IncludeBriefComments = false;
1763   IncludeStructure PreambleInclusions; // Unused for signatureHelp
1764   semaCodeComplete(
1765       llvm::make_unique<SignatureHelpCollector>(Options, Index, Result),
1766       Options,
1767       {FileName, Command, Preamble, Contents, *Offset, std::move(VFS)});
1768   return Result;
1769 }
1770 
1771 bool isIndexedForCodeCompletion(const NamedDecl &ND, ASTContext &ASTCtx) {
1772   auto InTopLevelScope = [](const NamedDecl &ND) {
1773     switch (ND.getDeclContext()->getDeclKind()) {
1774     case Decl::TranslationUnit:
1775     case Decl::Namespace:
1776     case Decl::LinkageSpec:
1777       return true;
1778     default:
1779       break;
1780     };
1781     return false;
1782   };
1783   // We only complete symbol's name, which is the same as the name of the
1784   // *primary* template in case of template specializations.
1785   if (isExplicitTemplateSpecialization<FunctionDecl>(ND) ||
1786       isExplicitTemplateSpecialization<CXXRecordDecl>(ND) ||
1787       isExplicitTemplateSpecialization<VarDecl>(ND))
1788     return false;
1789 
1790   if (InTopLevelScope(ND))
1791     return true;
1792 
1793   if (const auto *EnumDecl = dyn_cast<clang::EnumDecl>(ND.getDeclContext()))
1794     return InTopLevelScope(*EnumDecl) && !EnumDecl->isScoped();
1795 
1796   return false;
1797 }
1798 
1799 CompletionItem CodeCompletion::render(const CodeCompleteOptions &Opts) const {
1800   CompletionItem LSP;
1801   const auto *InsertInclude = Includes.empty() ? nullptr : &Includes[0];
1802   LSP.label = ((InsertInclude && InsertInclude->Insertion)
1803                    ? Opts.IncludeIndicator.Insert
1804                    : Opts.IncludeIndicator.NoInsert) +
1805               (Opts.ShowOrigins ? "[" + llvm::to_string(Origin) + "]" : "") +
1806               RequiredQualifier + Name + Signature;
1807 
1808   LSP.kind = Kind;
1809   LSP.detail = BundleSize > 1 ? llvm::formatv("[{0} overloads]", BundleSize)
1810                               : ReturnType;
1811   LSP.deprecated = Deprecated;
1812   if (InsertInclude)
1813     LSP.detail += "\n" + InsertInclude->Header;
1814   LSP.documentation = Documentation;
1815   LSP.sortText = sortText(Score.Total, Name);
1816   LSP.filterText = Name;
1817   LSP.textEdit = {CompletionTokenRange, RequiredQualifier + Name};
1818   // Merge continuous additionalTextEdits into main edit. The main motivation
1819   // behind this is to help LSP clients, it seems most of them are confused when
1820   // they are provided with additionalTextEdits that are consecutive to main
1821   // edit.
1822   // Note that we store additional text edits from back to front in a line. That
1823   // is mainly to help LSP clients again, so that changes do not effect each
1824   // other.
1825   for (const auto &FixIt : FixIts) {
1826     if (isRangeConsecutive(FixIt.range, LSP.textEdit->range)) {
1827       LSP.textEdit->newText = FixIt.newText + LSP.textEdit->newText;
1828       LSP.textEdit->range.start = FixIt.range.start;
1829     } else {
1830       LSP.additionalTextEdits.push_back(FixIt);
1831     }
1832   }
1833   if (Opts.EnableSnippets)
1834     LSP.textEdit->newText += SnippetSuffix;
1835 
1836   // FIXME(kadircet): Do not even fill insertText after making sure textEdit is
1837   // compatible with most of the editors.
1838   LSP.insertText = LSP.textEdit->newText;
1839   LSP.insertTextFormat = Opts.EnableSnippets ? InsertTextFormat::Snippet
1840                                              : InsertTextFormat::PlainText;
1841   if (InsertInclude && InsertInclude->Insertion)
1842     LSP.additionalTextEdits.push_back(*InsertInclude->Insertion);
1843 
1844   return LSP;
1845 }
1846 
1847 llvm::raw_ostream &operator<<(llvm::raw_ostream &OS, const CodeCompletion &C) {
1848   // For now just lean on CompletionItem.
1849   return OS << C.render(CodeCompleteOptions());
1850 }
1851 
1852 llvm::raw_ostream &operator<<(llvm::raw_ostream &OS,
1853                               const CodeCompleteResult &R) {
1854   OS << "CodeCompleteResult: " << R.Completions.size() << (R.HasMore ? "+" : "")
1855      << " (" << getCompletionKindString(R.Context) << ")"
1856      << " items:\n";
1857   for (const auto &C : R.Completions)
1858     OS << C << "\n";
1859   return OS;
1860 }
1861 
1862 } // namespace clangd
1863 } // namespace clang
1864