From: Douglas Gregor Date: Wed, 20 Oct 2010 01:32:02 +0000 (+0000) Subject: Introduce a simple cache for unqualified typo corrections, so that we X-Git-Url: https://granicus.if.org/sourcecode?a=commitdiff_plain;h=3eedbb06dd35b021bb614250585d4a6a5b7936e7;p=clang Introduce a simple cache for unqualified typo corrections, so that we don't repeatedly loop through identifiers, correcting the same typo'd identifier over and over again. We still bail out after 20 typo corrections, but this should help improve performance in the common case where we're typo-correcting because the user forgot to include a header. git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@116901 91177308-0d34-0410-b5e6-96231b3b80d8 --- diff --git a/include/clang/Sema/Sema.h b/include/clang/Sema/Sema.h index c0def12156..5e6497fe1b 100644 --- a/include/clang/Sema/Sema.h +++ b/include/clang/Sema/Sema.h @@ -3518,6 +3518,17 @@ public: /// \brief The number of typos corrected by CorrectTypo. unsigned TyposCorrected; + typedef llvm::DenseMap > + UnqualifiedTyposCorrectedMap; + + /// \brief A cache containing the results of typo correction for unqualified + /// name lookup. + /// + /// The string is the string that we corrected to (which may be empty, if + /// there was no correction), while the boolean will be true when the + /// string represents a keyword. + UnqualifiedTyposCorrectedMap UnqualifiedTyposCorrected; + /// \brief Worker object for performing CFG-based warnings. sema::AnalysisBasedWarnings AnalysisWarnings; diff --git a/lib/Sema/SemaLookup.cpp b/lib/Sema/SemaLookup.cpp index ab810a4682..1294f4f013 100644 --- a/lib/Sema/SemaLookup.cpp +++ b/lib/Sema/SemaLookup.cpp @@ -2822,14 +2822,6 @@ DeclarationName Sema::CorrectTypo(LookupResult &Res, Scope *S, CXXScopeSpec *SS, const ObjCObjectPointerType *OPT) { if (Diags.hasFatalErrorOccurred() || !getLangOptions().SpellChecking) return DeclarationName(); - - // Provide a stop gap for files that are just seriously broken. Trying - // to correct all typos can turn into a HUGE performance penalty, causing - // some files to take minutes to get rejected by the parser. - // FIXME: Is this the right solution? - if (TyposCorrected == 20) - return DeclarationName(); - ++TyposCorrected; // We only attempt to correct typos for identifiers. IdentifierInfo *Typo = Res.getLookupName().getAsIdentifierInfo(); @@ -2849,6 +2841,7 @@ DeclarationName Sema::CorrectTypo(LookupResult &Res, Scope *S, CXXScopeSpec *SS, TypoCorrectionConsumer Consumer(Typo); // Perform name lookup to find visible, similarly-named entities. + bool IsUnqualifiedLookup = false; if (MemberContext) { LookupVisibleDecls(MemberContext, Res.getLookupKind(), Consumer); @@ -2864,26 +2857,52 @@ DeclarationName Sema::CorrectTypo(LookupResult &Res, Scope *S, CXXScopeSpec *SS, if (!DC) return DeclarationName(); + // Provide a stop gap for files that are just seriously broken. Trying + // to correct all typos can turn into a HUGE performance penalty, causing + // some files to take minutes to get rejected by the parser. + if (TyposCorrected + UnqualifiedTyposCorrected.size() >= 20) + return DeclarationName(); + ++TyposCorrected; + LookupVisibleDecls(DC, Res.getLookupKind(), Consumer); } else { - // For unqualified lookup, look through all of the names that we have - // seen in this translation unit. - for (IdentifierTable::iterator I = Context.Idents.begin(), - IEnd = Context.Idents.end(); - I != IEnd; ++I) - Consumer.FoundName(I->getKey()); - - // Walk through identifiers in external identifier sources. - if (IdentifierInfoLookup *External + IsUnqualifiedLookup = true; + UnqualifiedTyposCorrectedMap::iterator Cached + = UnqualifiedTyposCorrected.find(Typo); + if (Cached == UnqualifiedTyposCorrected.end()) { + // Provide a stop gap for files that are just seriously broken. Trying + // to correct all typos can turn into a HUGE performance penalty, causing + // some files to take minutes to get rejected by the parser. + if (TyposCorrected + UnqualifiedTyposCorrected.size() >= 20) + return DeclarationName(); + + // For unqualified lookup, look through all of the names that we have + // seen in this translation unit. + for (IdentifierTable::iterator I = Context.Idents.begin(), + IEnd = Context.Idents.end(); + I != IEnd; ++I) + Consumer.FoundName(I->getKey()); + + // Walk through identifiers in external identifier sources. + if (IdentifierInfoLookup *External = Context.Idents.getExternalIdentifierLookup()) { - IdentifierIterator *Iter = External->getIdentifiers(); - do { - llvm::StringRef Name = Iter->Next(); - if (Name.empty()) - break; + IdentifierIterator *Iter = External->getIdentifiers(); + do { + llvm::StringRef Name = Iter->Next(); + if (Name.empty()) + break; - Consumer.FoundName(Name); - } while (true); + Consumer.FoundName(Name); + } while (true); + } + } else { + // Use the cached value, unless it's a keyword. In the keyword case, we'll + // end up adding the keyword below. + if (Cached->second.first.empty()) + return DeclarationName(); + + if (!Cached->second.second) + Consumer.FoundName(Cached->second.first); } } @@ -3053,14 +3072,24 @@ DeclarationName Sema::CorrectTypo(LookupResult &Res, Scope *S, CXXScopeSpec *SS, } // If we haven't found anything, we're done. - if (Consumer.empty()) + if (Consumer.empty()) { + // If this was an unqualified lookup, note that no correction was found. + if (IsUnqualifiedLookup) + (void)UnqualifiedTyposCorrected[Typo]; + return DeclarationName(); + } // Make sure that the user typed at least 3 characters for each correction // made. Otherwise, we don't even both looking at the results. unsigned ED = Consumer.getBestEditDistance(); - if (ED > 0 && Typo->getName().size() / ED < 3) + if (ED > 0 && Typo->getName().size() / ED < 3) { + // If this was an unqualified lookup, note that no correction was found. + if (IsUnqualifiedLookup) + (void)UnqualifiedTyposCorrected[Typo]; + return DeclarationName(); + } // Weed out any names that could not be found by name lookup. bool LastLookupWasAccepted = false; @@ -3172,6 +3201,11 @@ DeclarationName Sema::CorrectTypo(LookupResult &Res, Scope *S, CXXScopeSpec *SS, } } + // Record the correction for unqualified lookup. + if (IsUnqualifiedLookup) + UnqualifiedTyposCorrected[Typo] + = std::make_pair(Consumer.begin()->getKey(), Consumer.begin()->second); + return &Context.Idents.get(Consumer.begin()->getKey()); } else if (Consumer.size() > 1 && CTC == CTC_ObjCMessageReceiver @@ -3180,11 +3214,21 @@ DeclarationName Sema::CorrectTypo(LookupResult &Res, Scope *S, CXXScopeSpec *SS, // context. Res.suppressDiagnostics(); Res.clear(); + + // Record the correction for unqualified lookup. + if (IsUnqualifiedLookup) + UnqualifiedTyposCorrected[Typo] + = std::make_pair(Consumer.begin()->getKey(), Consumer.begin()->second); + return &Context.Idents.get("super"); } Res.suppressDiagnostics(); Res.setLookupName(Typo); Res.clear(); + // Record the correction for unqualified lookup. + if (IsUnqualifiedLookup) + (void)UnqualifiedTyposCorrected[Typo]; + return DeclarationName(); } diff --git a/test/SemaObjC/super.m b/test/SemaObjC/super.m index 31d8db1702..c15df26f34 100644 --- a/test/SemaObjC/super.m +++ b/test/SemaObjC/super.m @@ -54,8 +54,8 @@ void f0(int super) { [super m]; // expected-warning{{receiver type 'int' is not 'id'}} \ expected-warning {{method '-m' not found (return type defaults to 'id')}} } -void f1(id puper) { // expected-note {{'puper' declared here}} - [super m]; // expected-error{{use of undeclared identifier 'super'; did you mean 'puper'?}} +void f1(id puper) { + [super m]; // expected-error{{use of undeclared identifier 'super'}} } // radar 7400691