mirror of
https://github.com/llvm/llvm-project.git
synced 2025-05-03 12:06:07 +00:00

We have a workaround from D128621 that makes $0 no longer being a placeholder to conform a vscode feature. However, we have to refine the logic as it may suppress the last parameter placeholder for constructor of base class because not all patterns of completion are compound statements. This fixes clangd/clangd#1479 Reviewed By: nridge Differential Revision: https://reviews.llvm.org/D145319
308 lines
12 KiB
C++
308 lines
12 KiB
C++
//===--- CodeCompletionStrings.cpp -------------------------------*- C++-*-===//
|
|
//
|
|
// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
|
|
// See https://llvm.org/LICENSE.txt for license information.
|
|
// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
|
|
//
|
|
//===----------------------------------------------------------------------===//
|
|
|
|
#include "CodeCompletionStrings.h"
|
|
#include "clang-c/Index.h"
|
|
#include "clang/AST/ASTContext.h"
|
|
#include "clang/AST/RawCommentList.h"
|
|
#include "clang/Basic/SourceManager.h"
|
|
#include "clang/Sema/CodeCompleteConsumer.h"
|
|
#include "llvm/Support/JSON.h"
|
|
#include <limits>
|
|
#include <utility>
|
|
|
|
namespace clang {
|
|
namespace clangd {
|
|
namespace {
|
|
|
|
bool isInformativeQualifierChunk(CodeCompletionString::Chunk const &Chunk) {
|
|
return Chunk.Kind == CodeCompletionString::CK_Informative &&
|
|
llvm::StringRef(Chunk.Text).endswith("::");
|
|
}
|
|
|
|
void appendEscapeSnippet(const llvm::StringRef Text, std::string *Out) {
|
|
for (const auto Character : Text) {
|
|
if (Character == '$' || Character == '}' || Character == '\\')
|
|
Out->push_back('\\');
|
|
Out->push_back(Character);
|
|
}
|
|
}
|
|
|
|
void appendOptionalChunk(const CodeCompletionString &CCS, std::string *Out) {
|
|
for (const CodeCompletionString::Chunk &C : CCS) {
|
|
switch (C.Kind) {
|
|
case CodeCompletionString::CK_Optional:
|
|
assert(C.Optional &&
|
|
"Expected the optional code completion string to be non-null.");
|
|
appendOptionalChunk(*C.Optional, Out);
|
|
break;
|
|
default:
|
|
*Out += C.Text;
|
|
break;
|
|
}
|
|
}
|
|
}
|
|
|
|
bool looksLikeDocComment(llvm::StringRef CommentText) {
|
|
// We don't report comments that only contain "special" chars.
|
|
// This avoids reporting various delimiters, like:
|
|
// =================
|
|
// -----------------
|
|
// *****************
|
|
return CommentText.find_first_not_of("/*-= \t\r\n") != llvm::StringRef::npos;
|
|
}
|
|
|
|
// Determine whether the completion string should be patched
|
|
// to replace the last placeholder with $0.
|
|
bool shouldPatchPlaceholder0(CodeCompletionResult::ResultKind ResultKind,
|
|
CXCursorKind CursorKind) {
|
|
bool CompletingPattern = ResultKind == CodeCompletionResult::RK_Pattern;
|
|
|
|
if (!CompletingPattern)
|
|
return false;
|
|
|
|
// If the result kind of CodeCompletionResult(CCR) is `RK_Pattern`, it doesn't
|
|
// always mean we're completing a chunk of statements. Constructors defined
|
|
// in base class, for example, are considered as a type of pattern, with the
|
|
// cursor type set to CXCursor_Constructor.
|
|
if (CursorKind == CXCursorKind::CXCursor_Constructor ||
|
|
CursorKind == CXCursorKind::CXCursor_Destructor)
|
|
return false;
|
|
|
|
return true;
|
|
}
|
|
|
|
} // namespace
|
|
|
|
std::string getDocComment(const ASTContext &Ctx,
|
|
const CodeCompletionResult &Result,
|
|
bool CommentsFromHeaders) {
|
|
// FIXME: clang's completion also returns documentation for RK_Pattern if they
|
|
// contain a pattern for ObjC properties. Unfortunately, there is no API to
|
|
// get this declaration, so we don't show documentation in that case.
|
|
if (Result.Kind != CodeCompletionResult::RK_Declaration)
|
|
return "";
|
|
return Result.getDeclaration() ? getDeclComment(Ctx, *Result.getDeclaration())
|
|
: "";
|
|
}
|
|
|
|
std::string getDeclComment(const ASTContext &Ctx, const NamedDecl &Decl) {
|
|
if (isa<NamespaceDecl>(Decl)) {
|
|
// Namespaces often have too many redecls for any particular redecl comment
|
|
// to be useful. Moreover, we often confuse file headers or generated
|
|
// comments with namespace comments. Therefore we choose to just ignore
|
|
// the comments for namespaces.
|
|
return "";
|
|
}
|
|
const RawComment *RC = getCompletionComment(Ctx, &Decl);
|
|
if (!RC)
|
|
return "";
|
|
// Sanity check that the comment does not come from the PCH. We choose to not
|
|
// write them into PCH, because they are racy and slow to load.
|
|
assert(!Ctx.getSourceManager().isLoadedSourceLocation(RC->getBeginLoc()));
|
|
std::string Doc =
|
|
RC->getFormattedText(Ctx.getSourceManager(), Ctx.getDiagnostics());
|
|
if (!looksLikeDocComment(Doc))
|
|
return "";
|
|
// Clang requires source to be UTF-8, but doesn't enforce this in comments.
|
|
if (!llvm::json::isUTF8(Doc))
|
|
Doc = llvm::json::fixUTF8(Doc);
|
|
return Doc;
|
|
}
|
|
|
|
void getSignature(const CodeCompletionString &CCS, std::string *Signature,
|
|
std::string *Snippet,
|
|
CodeCompletionResult::ResultKind ResultKind,
|
|
CXCursorKind CursorKind, std::string *RequiredQualifiers) {
|
|
// Placeholder with this index will be $0 to mark final cursor position.
|
|
// Usually we do not add $0, so the cursor is placed at end of completed text.
|
|
unsigned CursorSnippetArg = std::numeric_limits<unsigned>::max();
|
|
|
|
// If the snippet contains a group of statements, we replace the
|
|
// last placeholder with $0 to leave the cursor there, e.g.
|
|
// namespace ${1:name} {
|
|
// ${0:decls}
|
|
// }
|
|
// We try to identify such cases using the ResultKind and CursorKind.
|
|
if (shouldPatchPlaceholder0(ResultKind, CursorKind)) {
|
|
CursorSnippetArg =
|
|
llvm::count_if(CCS, [](const CodeCompletionString::Chunk &C) {
|
|
return C.Kind == CodeCompletionString::CK_Placeholder;
|
|
});
|
|
}
|
|
unsigned SnippetArg = 0;
|
|
bool HadObjCArguments = false;
|
|
bool HadInformativeChunks = false;
|
|
for (const auto &Chunk : CCS) {
|
|
// Informative qualifier chunks only clutter completion results, skip
|
|
// them.
|
|
if (isInformativeQualifierChunk(Chunk))
|
|
continue;
|
|
|
|
switch (Chunk.Kind) {
|
|
case CodeCompletionString::CK_TypedText:
|
|
// The typed-text chunk is the actual name. We don't record this chunk.
|
|
// C++:
|
|
// In general our string looks like <qualifiers><name><signature>.
|
|
// So once we see the name, any text we recorded so far should be
|
|
// reclassified as qualifiers.
|
|
//
|
|
// Objective-C:
|
|
// Objective-C methods expressions may have multiple typed-text chunks,
|
|
// so we must treat them carefully. For Objective-C methods, all
|
|
// typed-text and informative chunks will end in ':' (unless there are
|
|
// no arguments, in which case we can safely treat them as C++).
|
|
//
|
|
// Completing a method declaration itself (not a method expression) is
|
|
// similar except that we use the `RequiredQualifiers` to store the
|
|
// text before the selector, e.g. `- (void)`.
|
|
if (!llvm::StringRef(Chunk.Text).endswith(":")) { // Treat as C++.
|
|
if (RequiredQualifiers)
|
|
*RequiredQualifiers = std::move(*Signature);
|
|
Signature->clear();
|
|
Snippet->clear();
|
|
} else { // Objective-C method with args.
|
|
// If this is the first TypedText to the Objective-C method, discard any
|
|
// text that we've previously seen (such as previous parameter selector,
|
|
// which will be marked as Informative text).
|
|
//
|
|
// TODO: Make previous parameters part of the signature for Objective-C
|
|
// methods.
|
|
if (!HadObjCArguments) {
|
|
HadObjCArguments = true;
|
|
// If we have no previous informative chunks (informative selector
|
|
// fragments in practice), we treat any previous chunks as
|
|
// `RequiredQualifiers` so they will be added as a prefix during the
|
|
// completion.
|
|
//
|
|
// e.g. to complete `- (void)doSomething:(id)argument`:
|
|
// - Completion name: `doSomething:`
|
|
// - RequiredQualifiers: `- (void)`
|
|
// - Snippet/Signature suffix: `(id)argument`
|
|
//
|
|
// This differs from the case when we're completing a method
|
|
// expression with a previous informative selector fragment.
|
|
//
|
|
// e.g. to complete `[self doSomething:nil ^somethingElse:(id)]`:
|
|
// - Previous Informative Chunk: `doSomething:`
|
|
// - Completion name: `somethingElse:`
|
|
// - Snippet/Signature suffix: `(id)`
|
|
if (!HadInformativeChunks) {
|
|
if (RequiredQualifiers)
|
|
*RequiredQualifiers = std::move(*Signature);
|
|
Snippet->clear();
|
|
}
|
|
Signature->clear();
|
|
} else { // Subsequent argument, considered part of snippet/signature.
|
|
*Signature += Chunk.Text;
|
|
*Snippet += Chunk.Text;
|
|
}
|
|
}
|
|
break;
|
|
case CodeCompletionString::CK_Text:
|
|
*Signature += Chunk.Text;
|
|
*Snippet += Chunk.Text;
|
|
break;
|
|
case CodeCompletionString::CK_Optional:
|
|
assert(Chunk.Optional);
|
|
// No need to create placeholders for default arguments in Snippet.
|
|
appendOptionalChunk(*Chunk.Optional, Signature);
|
|
break;
|
|
case CodeCompletionString::CK_Placeholder:
|
|
*Signature += Chunk.Text;
|
|
++SnippetArg;
|
|
if (SnippetArg == CursorSnippetArg) {
|
|
// We'd like to make $0 a placeholder too, but vscode does not support
|
|
// this (https://github.com/microsoft/vscode/issues/152837).
|
|
*Snippet += "$0";
|
|
} else {
|
|
*Snippet += "${" + std::to_string(SnippetArg) + ':';
|
|
appendEscapeSnippet(Chunk.Text, Snippet);
|
|
*Snippet += '}';
|
|
}
|
|
break;
|
|
case CodeCompletionString::CK_Informative:
|
|
HadInformativeChunks = true;
|
|
// For example, the word "const" for a const method, or the name of
|
|
// the base class for methods that are part of the base class.
|
|
*Signature += Chunk.Text;
|
|
// Don't put the informative chunks in the snippet.
|
|
break;
|
|
case CodeCompletionString::CK_ResultType:
|
|
// This is not part of the signature.
|
|
break;
|
|
case CodeCompletionString::CK_CurrentParameter:
|
|
// This should never be present while collecting completion items,
|
|
// only while collecting overload candidates.
|
|
llvm_unreachable("Unexpected CK_CurrentParameter while collecting "
|
|
"CompletionItems");
|
|
break;
|
|
case CodeCompletionString::CK_LeftParen:
|
|
case CodeCompletionString::CK_RightParen:
|
|
case CodeCompletionString::CK_LeftBracket:
|
|
case CodeCompletionString::CK_RightBracket:
|
|
case CodeCompletionString::CK_LeftBrace:
|
|
case CodeCompletionString::CK_RightBrace:
|
|
case CodeCompletionString::CK_LeftAngle:
|
|
case CodeCompletionString::CK_RightAngle:
|
|
case CodeCompletionString::CK_Comma:
|
|
case CodeCompletionString::CK_Colon:
|
|
case CodeCompletionString::CK_SemiColon:
|
|
case CodeCompletionString::CK_Equal:
|
|
case CodeCompletionString::CK_HorizontalSpace:
|
|
*Signature += Chunk.Text;
|
|
*Snippet += Chunk.Text;
|
|
break;
|
|
case CodeCompletionString::CK_VerticalSpace:
|
|
*Snippet += Chunk.Text;
|
|
// Don't even add a space to the signature.
|
|
break;
|
|
}
|
|
}
|
|
}
|
|
|
|
std::string formatDocumentation(const CodeCompletionString &CCS,
|
|
llvm::StringRef DocComment) {
|
|
// Things like __attribute__((nonnull(1,3))) and [[noreturn]]. Present this
|
|
// information in the documentation field.
|
|
std::string Result;
|
|
const unsigned AnnotationCount = CCS.getAnnotationCount();
|
|
if (AnnotationCount > 0) {
|
|
Result += "Annotation";
|
|
if (AnnotationCount == 1) {
|
|
Result += ": ";
|
|
} else /* AnnotationCount > 1 */ {
|
|
Result += "s: ";
|
|
}
|
|
for (unsigned I = 0; I < AnnotationCount; ++I) {
|
|
Result += CCS.getAnnotation(I);
|
|
Result.push_back(I == AnnotationCount - 1 ? '\n' : ' ');
|
|
}
|
|
}
|
|
// Add brief documentation (if there is any).
|
|
if (!DocComment.empty()) {
|
|
if (!Result.empty()) {
|
|
// This means we previously added annotations. Add an extra newline
|
|
// character to make the annotations stand out.
|
|
Result.push_back('\n');
|
|
}
|
|
Result += DocComment;
|
|
}
|
|
return Result;
|
|
}
|
|
|
|
std::string getReturnType(const CodeCompletionString &CCS) {
|
|
for (const auto &Chunk : CCS)
|
|
if (Chunk.Kind == CodeCompletionString::CK_ResultType)
|
|
return Chunk.Text;
|
|
return "";
|
|
}
|
|
|
|
} // namespace clangd
|
|
} // namespace clang
|