Extract SemanticdbSymbols to shared module

jupblb · jupblb · commit 1a391dc30438 · 2026-06-03T14:06:38.000+02:00
Move the proto-free symbol-encoding helper to a new semanticdb-shared
module so semanticdb-java and semanticdb-kotlinc can both depend on it.
The Kotlin Symbol value class and SemanticdbSymbolDescriptor data class
now delegate descriptor encoding to the shared Java implementation
instead of duplicating it.
diff --git a/build.sbt b/build.sbt
@@ -72,6 +72,17 @@ commands +=
       "scalafixAll --check" :: "publishLocal" :: s
   }
 
+// Shared, proto-free utility code consumed by both the Java and Kotlin
+// SemanticDB compiler plugins. Kept intentionally small so each plugin's
+// fat-jar pays only for the encoding helpers, not for the full Java proto
+// schema.
+lazy val semanticdbShared = project
+  .in(file("semanticdb-shared"))
+  .settings(
+    moduleName := "semanticdb-shared",
+    javaOnlySettings
+  )
+
 lazy val semanticdb = project
   .in(file("semanticdb-java"))
   .settings(
@@ -80,6 +91,7 @@ lazy val semanticdb = project
     (Compile / PB.targets) :=
       Seq(PB.gens.java(V.protobuf) -> (Compile / sourceManaged).value)
   )
+  .dependsOn(semanticdbShared)
 
 lazy val agent = project
   .in(file("semanticdb-agent"))
@@ -431,6 +443,7 @@ lazy val semanticdbKotlinc = project
       Attributed.blank(dir)
     }
   )
+  .dependsOn(semanticdbShared)
 
 // `semanticdbKotlincMinimized` mirrors the (still-present) Gradle build at
 // semanticdb-kotlinc/minimized/build.gradle.kts. It compiles a small set of
diff --git a/scip-java/src/main/scala/com/sourcegraph/scip_java/ScipSymbol.scala b/scip-java/src/main/scala/com/sourcegraph/scip_java/ScipSymbol.scala
@@ -1,7 +1,7 @@
 package com.sourcegraph.scip_java
 
 import com.sourcegraph.scip_semanticdb.SymbolDescriptor
-import com.sourcegraph.semanticdb_javac.SemanticdbSymbols
+import com.sourcegraph.semanticdb.SemanticdbSymbols
 
 sealed abstract class ScipSymbol {}
 final case class LocalScipSymbol(identifier: String) extends ScipSymbol
diff --git a/scip-semanticdb/BUILD b/scip-semanticdb/BUILD
@@ -21,6 +21,7 @@ java_library(
         ":all_java_proto",
         "//semanticdb-java",
         "//semanticdb-java/src/main/protobuf:semanticdb_java_proto",
+        "//semanticdb-shared",
         "@maven//:com_google_code_findbugs_jsr305",
         "@maven//:com_google_protobuf_protobuf_java",
         "@maven//:com_google_protobuf_protobuf_java_util",
diff --git a/scip-semanticdb/src/main/java/com/sourcegraph/scip_semanticdb/ScipSemanticdb.java b/scip-semanticdb/src/main/java/com/sourcegraph/scip_semanticdb/ScipSemanticdb.java
@@ -4,7 +4,7 @@
 import com.sourcegraph.semanticdb_javac.Semanticdb;
 import com.sourcegraph.semanticdb_javac.Semanticdb.SymbolOccurrence;
 import com.sourcegraph.semanticdb_javac.Semanticdb.SymbolOccurrence.Role;
-import com.sourcegraph.semanticdb_javac.SemanticdbSymbols;
+import com.sourcegraph.semanticdb.SemanticdbSymbols;
 import org.scip_code.scip.Document;
 import org.scip_code.scip.Index;
 import org.scip_code.scip.Metadata;
diff --git a/scip-semanticdb/src/main/java/com/sourcegraph/scip_semanticdb/SignatureFormatter.java b/scip-semanticdb/src/main/java/com/sourcegraph/scip_semanticdb/SignatureFormatter.java
@@ -3,7 +3,7 @@
 import com.sourcegraph.semanticdb_javac.Semanticdb.SymbolInformation.Property;
 import com.sourcegraph.semanticdb_javac.Semanticdb.*;
 
-import com.sourcegraph.semanticdb_javac.SemanticdbSymbols;
+import com.sourcegraph.semanticdb.SemanticdbSymbols;
 
 import java.util.*;
 import java.util.concurrent.atomic.AtomicInteger;
diff --git a/scip-semanticdb/src/main/java/com/sourcegraph/scip_semanticdb/SymbolDescriptor.java b/scip-semanticdb/src/main/java/com/sourcegraph/scip_semanticdb/SymbolDescriptor.java
@@ -1,8 +1,8 @@
 package com.sourcegraph.scip_semanticdb;
 
-import com.sourcegraph.semanticdb_javac.SemanticdbSymbols;
-import com.sourcegraph.semanticdb_javac.SemanticdbSymbols.Descriptor;
-import com.sourcegraph.semanticdb_javac.SemanticdbSymbols.Descriptor.Kind;
+import com.sourcegraph.semanticdb.SemanticdbSymbols;
+import com.sourcegraph.semanticdb.SemanticdbSymbols.Descriptor;
+import com.sourcegraph.semanticdb.SemanticdbSymbols.Descriptor.Kind;
 import java.util.Optional;
 
 public class SymbolDescriptor {
diff --git a/semanticdb-javac/BUILD b/semanticdb-javac/BUILD
@@ -33,5 +33,6 @@ java_library(
     deps = [
         "//semanticdb-java",
         "//semanticdb-java/src/main/protobuf:semanticdb_java_proto",
+        "//semanticdb-shared",
     ],
 )
diff --git a/semanticdb-javac/src/main/java/com/sourcegraph/semanticdb_javac/GlobalSymbolsCache.java b/semanticdb-javac/src/main/java/com/sourcegraph/semanticdb_javac/GlobalSymbolsCache.java
@@ -1,5 +1,7 @@
 package com.sourcegraph.semanticdb_javac;
 
+import com.sourcegraph.semanticdb.SemanticdbSymbols;
+
 import javax.lang.model.element.Element;
 import javax.lang.model.element.ExecutableElement;
 import javax.lang.model.element.TypeElement;
diff --git a/semanticdb-javac/src/main/java/com/sourcegraph/semanticdb_javac/LocalSymbolsCache.java b/semanticdb-javac/src/main/java/com/sourcegraph/semanticdb_javac/LocalSymbolsCache.java
@@ -1,5 +1,7 @@
 package com.sourcegraph.semanticdb_javac;
 
+import com.sourcegraph.semanticdb.SemanticdbSymbols;
+
 import javax.lang.model.element.Element;
 
 import java.util.IdentityHashMap;
diff --git a/semanticdb-javac/src/main/java/com/sourcegraph/semanticdb_javac/SemanticdbVisitor.java b/semanticdb-javac/src/main/java/com/sourcegraph/semanticdb_javac/SemanticdbVisitor.java
@@ -1,5 +1,7 @@
 package com.sourcegraph.semanticdb_javac;
 
+import com.sourcegraph.semanticdb.SemanticdbSymbols;
+
 import com.sun.source.util.SourcePositions;
 import com.sun.source.util.Trees;
 import com.sun.source.util.TreePathScanner;
diff --git a/semanticdb-kotlinc/src/main/kotlin/com/sourcegraph/semanticdb_kotlinc/SemanticdbSymbols.kt b/semanticdb-kotlinc/src/main/kotlin/com/sourcegraph/semanticdb_kotlinc/SemanticdbSymbols.kt
@@ -1,24 +1,30 @@
 package com.sourcegraph.semanticdb_kotlinc
 
+import com.sourcegraph.semanticdb.SemanticdbSymbols as SharedSymbols
+
 @JvmInline
 value class Symbol(private val symbol: String) {
     companion object {
-        val NONE = Symbol("")
-        val ROOT_PACKAGE = Symbol("_root_/")
+        val NONE = Symbol(SharedSymbols.NONE)
+        val ROOT_PACKAGE = Symbol(SharedSymbols.ROOT_PACKAGE)
 
+        // Note: this intentionally diverges from `SharedSymbols.global` when
+        // `desc == NONE` — Java returns `NONE`, Kotlin returns the owner.
+        // SymbolsCache relies on this behavior; do not delegate without first
+        // updating those call sites.
         fun createGlobal(owner: Symbol, desc: SemanticdbSymbolDescriptor): Symbol =
             when {
                 desc == SemanticdbSymbolDescriptor.NONE -> owner
                 owner != ROOT_PACKAGE -> Symbol(owner.symbol + desc.encode().symbol)
                 else -> desc.encode()
             }
 
-        fun createLocal(i: Int) = Symbol("local$i")
+        fun createLocal(i: Int) = Symbol(SharedSymbols.local(i))
     }
 
-    fun isGlobal() = !isLocal()
+    fun isGlobal() = SharedSymbols.isGlobal(symbol)
 
-    fun isLocal() = symbol.startsWith("local")
+    fun isLocal() = SharedSymbols.isLocal(symbol)
 
     override fun toString(): String = symbol
 }
@@ -28,22 +34,13 @@ fun String.symbol(): Symbol = Symbol(this)
 data class SemanticdbSymbolDescriptor(
     val kind: Kind,
     val name: String,
+    // Default differs from `SharedSymbols.Descriptor` (which is "") because
+    // Kotlin call sites — getters/setters in particular — rely on the no-arg
+    // overload producing `name().` rather than `name.` for METHOD kinds.
     val disambiguator: String = "()"
 ) {
     companion object {
         val NONE = SemanticdbSymbolDescriptor(Kind.NONE, "")
-
-        private fun encodeName(name: String): String {
-            if (name.isEmpty()) return "``"
-            val isStartOk = Character.isJavaIdentifierStart(name[0])
-            var isPartsOk = true
-            var i = 1
-            while (isPartsOk && i < name.length) {
-                isPartsOk = Character.isJavaIdentifierPart(name[i])
-                i++
-            }
-            return if (isStartOk && isPartsOk) name else "`$name`"
-        }
     }
 
     enum class Kind {
@@ -53,18 +50,21 @@ data class SemanticdbSymbolDescriptor(
         TYPE,
         PACKAGE,
         PARAMETER,
-        TYPE_PARAMETER
+        TYPE_PARAMETER;
+
+        internal fun toSharedKind(): SharedSymbols.Descriptor.Kind =
+            when (this) {
+                NONE -> SharedSymbols.Descriptor.Kind.None
+                TERM -> SharedSymbols.Descriptor.Kind.Term
+                METHOD -> SharedSymbols.Descriptor.Kind.Method
+                TYPE -> SharedSymbols.Descriptor.Kind.Type
+                PACKAGE -> SharedSymbols.Descriptor.Kind.Package
+                PARAMETER -> SharedSymbols.Descriptor.Kind.Parameter
+                TYPE_PARAMETER -> SharedSymbols.Descriptor.Kind.TypeParameter
+            }
     }
 
-    fun encode() =
-        Symbol(
-            when (kind) {
-                Kind.NONE -> ""
-                Kind.TERM -> "${encodeName(name)}."
-                Kind.METHOD -> "${encodeName(name)}${disambiguator}."
-                Kind.TYPE -> "${encodeName(name)}#"
-                Kind.PACKAGE -> "${encodeName(name)}/"
-                Kind.PARAMETER -> "(${encodeName(name)})"
-                Kind.TYPE_PARAMETER -> "[${encodeName(name)}]"
-            })
+    fun encode(): Symbol =
+        if (kind == Kind.NONE) Symbol(SharedSymbols.NONE)
+        else Symbol(SharedSymbols.Descriptor(kind.toSharedKind(), name, disambiguator).encode())
 }
diff --git a/semanticdb-shared/BUILD b/semanticdb-shared/BUILD
@@ -0,0 +1,10 @@
+load("@rules_java//java:defs.bzl", "java_library")
+
+package(
+    default_visibility = ["//visibility:public"],
+)
+
+java_library(
+    name = "semanticdb-shared",
+    srcs = glob(["src/main/java/**/*.java"]),
+)
diff --git a/semanticdb-shared/src/main/java/com/sourcegraph/semanticdb/SemanticdbSymbols.java b/semanticdb-shared/src/main/java/com/sourcegraph/semanticdb/SemanticdbSymbols.java
@@ -1,4 +1,4 @@
-package com.sourcegraph.semanticdb_javac;
+package com.sourcegraph.semanticdb;
 
 import java.util.Objects;
 
diff --git a/tests/unit/src/test/scala/tests/SymbolDescriptorSuite.scala b/tests/unit/src/test/scala/tests/SymbolDescriptorSuite.scala
@@ -3,7 +3,7 @@ package tests
 import scala.meta.internal.semanticdb.Scala._
 
 import com.sourcegraph.scip_semanticdb.SymbolDescriptor
-import com.sourcegraph.semanticdb_javac.SemanticdbSymbols.Descriptor.Kind
+import com.sourcegraph.semanticdb.SemanticdbSymbols.Descriptor.Kind
 import munit.FunSuite
 import munit.TestOptions
 

Original file line number	Diff line number	Diff line change
`@@ -33,5 +33,6 @@ java_library(`
`33`	`33`	`deps = [`
`34`	`34`	`"//semanticdb-java",`
`35`	`35`	`"//semanticdb-java/src/main/protobuf:semanticdb_java_proto",`
	`36`	`+ "//semanticdb-shared",`
`36`	`37`	`],`
`37`	`38`	`)`