[clang] [clang][ExtractAPI] Remove symbols defined in categories to external types unless requested (PR #92522)

[Daniel Grumberg via cfe-commits]

https://github.com/daniel-grumberg created https://github.com/llvm/llvm-project/pull/92522

rdar://128259890

>From 7650c18c883bb14e5a4b17d6b6d61297f2fa3c44 Mon Sep 17 00:00:00 2001
From: Daniel Grumberg <dgrumberg at apple.com>
Date: Fri, 17 May 2024 11:58:18 +0100
Subject: [PATCH] [clang][ExtractAPI] Remove symbols defined in categories to
 external types unless requested

rdar://128259890
---
 .../Serialization/SymbolGraphSerializer.h      |  9 +++++++--
 .../Serialization/SymbolGraphSerializer.cpp    | 11 +++++++++--
 clang/test/ExtractAPI/objc_external_category.m | 18 +++++++++++++-----
 3 files changed, 29 insertions(+), 9 deletions(-)

diff --git a/clang/include/clang/ExtractAPI/Serialization/SymbolGraphSerializer.h b/clang/include/clang/ExtractAPI/Serialization/SymbolGraphSerializer.h
index 724b087f7aea9..27e9167ca1ad0 100644
--- a/clang/include/clang/ExtractAPI/Serialization/SymbolGraphSerializer.h
+++ b/clang/include/clang/ExtractAPI/Serialization/SymbolGraphSerializer.h
@@ -102,6 +102,8 @@ class SymbolGraphSerializer : public APISetVisitor<SymbolGraphSerializer> {
 
   const bool EmitSymbolLabelsForTesting = false;
 
+  const bool SkipSymbolsInCategoriesToExternalTypes = false;
+
   /// The object instantiated by the last call to serializeAPIRecord.
   Object *CurrentSymbol = nullptr;
 
@@ -271,10 +273,13 @@ class SymbolGraphSerializer : public APISetVisitor<SymbolGraphSerializer> {
 
   SymbolGraphSerializer(const APISet &API, const APIIgnoresList &IgnoresList,
                         bool EmitSymbolLabelsForTesting = false,
-                        bool ForceEmitToMainModule = false)
+                        bool ForceEmitToMainModule = false,
+                        bool SkipSymbolsInCategoriesToExternalTypes = false)
       : Base(API), ForceEmitToMainModule(ForceEmitToMainModule),
         IgnoresList(IgnoresList),
-        EmitSymbolLabelsForTesting(EmitSymbolLabelsForTesting) {}
+        EmitSymbolLabelsForTesting(EmitSymbolLabelsForTesting),
+        SkipSymbolsInCategoriesToExternalTypes(
+            SkipSymbolsInCategoriesToExternalTypes) {}
 };
 
 } // namespace extractapi
diff --git a/clang/lib/ExtractAPI/Serialization/SymbolGraphSerializer.cpp b/clang/lib/ExtractAPI/Serialization/SymbolGraphSerializer.cpp
index c16d4623f115d..08e711cafae28 100644
--- a/clang/lib/ExtractAPI/Serialization/SymbolGraphSerializer.cpp
+++ b/clang/lib/ExtractAPI/Serialization/SymbolGraphSerializer.cpp
@@ -925,6 +925,10 @@ bool SymbolGraphSerializer::visitObjCInterfaceRecord(
 
 bool SymbolGraphSerializer::traverseObjCCategoryRecord(
     const ObjCCategoryRecord *Record) {
+  if (SkipSymbolsInCategoriesToExternalTypes &&
+      !API.findRecordForUSR(Record->Interface.USR))
+    return true;
+
   auto *CurrentModule = ModuleForCurrentSymbol;
   if (Record->isExtendingExternalModule())
     ModuleForCurrentSymbol = &ExtendedModules[Record->Interface.Source];
@@ -1040,8 +1044,11 @@ void SymbolGraphSerializer::serializeGraphToStream(
 void SymbolGraphSerializer::serializeMainSymbolGraph(
     raw_ostream &OS, const APISet &API, const APIIgnoresList &IgnoresList,
     SymbolGraphSerializerOption Options) {
-  SymbolGraphSerializer Serializer(API, IgnoresList,
-                                   Options.EmitSymbolLabelsForTesting);
+  SymbolGraphSerializer Serializer(
+      API, IgnoresList, Options.EmitSymbolLabelsForTesting,
+      /*ForceEmitToMainModule=*/true,
+      /*SkipSymbolsInCategoriesToExternalTypes=*/true);
+
   Serializer.traverseAPISet();
   Serializer.serializeGraphToStream(OS, Options, API.ProductName,
                                     std::move(Serializer.MainModule));
diff --git a/clang/test/ExtractAPI/objc_external_category.m b/clang/test/ExtractAPI/objc_external_category.m
index 47e699cb91c0e..8afc92489f28b 100644
--- a/clang/test/ExtractAPI/objc_external_category.m
+++ b/clang/test/ExtractAPI/objc_external_category.m
@@ -4,6 +4,9 @@
 // RUN:   --emit-extension-symbol-graphs --symbol-graph-dir=%t/symbols \
 // RUN:   --product-name=Module -fmodules -fimplicit-module-maps -fmodules-cache-path=%t/modules-cache \
 // RUN:   -triple arm64-apple-macosx -x objective-c-header %t/input.h -verify
+// RUN: %clang_cc1 -extract-api --pretty-sgf --emit-sgf-symbol-labels-for-testing \
+// RUN:   --product-name=Module -o %t/ModuleNoExt.symbols.json -triple arm64-apple-macosx \
+// RUN:   -x objective-c-header %t/input.h
 
 //--- input.h
 #include "ExternalModule.h"
@@ -28,15 +31,20 @@ @interface ExtInterface
     header "ExternalModule.h"
 }
 
+// Main symbol graph from the build with extension SGFs
 // RUN: FileCheck %s --input-file  %t/symbols/Module.symbols.json --check-prefix MOD
+
 // MOD-NOT: "!testRelLabel": "memberOf $ c:objc(cs)ExtInterface(py)Property $ c:objc(cs)ExtInterface"
 // MOD-NOT: "!testRelLabel": "memberOf $ c:objc(cs)ExtInterface(im)InstanceMethod $ c:objc(cs)ExtInterface"
 // MOD-NOT: "!testRelLabel": "memberOf $ c:objc(cs)ExtInterface(cm)ClassMethod $ c:objc(cs)ExtInterface"
-// MOD-NOT: "!testLabel": "c:objc(cs)ExtInterface(py)Property"
-// MOD-NOT: "!testLabel": "c:objc(cs)ExtInterface(im)InstanceMethod"
-// MOD-NOT: "!testLabel": "c:objc(cs)ExtInterface(cm)ClassMethod"
-// MOD-NOT: "!testLabel": "c:objc(cs)ExtInterface"
-// MOD-DAG: "!testLabel": "c:objc(cs)ModInterface"
+// MOD-NOT: "c:objc(cs)ExtInterface(py)Property"
+// MOD-NOT: "c:objc(cs)ExtInterface(im)InstanceMethod"
+// MOD-NOT: "c:objc(cs)ExtInterface(cm)ClassMethod"
+// MOD-NOT: "c:objc(cs)ExtInterface"
+// MOD-DAG: "c:objc(cs)ModInterface"
+
+// Symbol graph from the build without extension SGFs should be identical to main symbol graph with extension SGFs
+// RUN: diff %t/symbols/Module.symbols.json %t/ModuleNoExt.symbols.json
 
 // RUN: FileCheck %s --input-file %t/symbols/ExternalModule at Module.symbols.json --check-prefix EXT
 // EXT-DAG: "!testRelLabel": "memberOf $ c:objc(cs)ExtInterface(py)Property $ c:objc(cs)ExtInterface"