From a66c05ad4aff70e2ec52d4b6353ff11cbf1f76a9 Mon Sep 17 00:00:00 2001
From: Chad Retz <chad.retz@gmail.com>
Date: Thu, 26 Jul 2018 00:05:18 -0500
Subject: [PATCH] Support large data sections. Fixes #18

---
 .../main/kotlin/asmble/compile/jvm/AsmExt.kt  |  4 ++
 .../asmble/compile/jvm/ByteBufferMem.kt       | 27 +++++++-----
 .../asmble/compile/jvm/LargeDataTest.kt       | 43 +++++++++++++++++++
 3 files changed, 64 insertions(+), 10 deletions(-)
 create mode 100644 compiler/src/test/kotlin/asmble/compile/jvm/LargeDataTest.kt

diff --git a/compiler/src/main/kotlin/asmble/compile/jvm/AsmExt.kt b/compiler/src/main/kotlin/asmble/compile/jvm/AsmExt.kt
index d17385a..3c3586a 100644
--- a/compiler/src/main/kotlin/asmble/compile/jvm/AsmExt.kt
+++ b/compiler/src/main/kotlin/asmble/compile/jvm/AsmExt.kt
@@ -210,3 +210,7 @@ fun ByteArray.asClassNode(): ClassNode {
     ClassReader(this).accept(newNode, 0)
     return newNode
 }
+
+fun ByteArray.chunked(v: Int) = (0 until size step v).asSequence().map {
+    copyOfRange(it, (it + v).takeIf { it < size } ?: size)
+}
\ No newline at end of file
diff --git a/compiler/src/main/kotlin/asmble/compile/jvm/ByteBufferMem.kt b/compiler/src/main/kotlin/asmble/compile/jvm/ByteBufferMem.kt
index ec04248..87cfa02 100644
--- a/compiler/src/main/kotlin/asmble/compile/jvm/ByteBufferMem.kt
+++ b/compiler/src/main/kotlin/asmble/compile/jvm/ByteBufferMem.kt
@@ -46,16 +46,23 @@ open class ByteBufferMem(val direct: Boolean = true) : Mem {
             let(buildOffset).popExpecting(Int::class.ref).
             addInsns(
                 forceFnType<ByteBuffer.(Int) -> Buffer>(ByteBuffer::position).invokeVirtual(),
-                TypeInsnNode(Opcodes.CHECKCAST, memType.asmName),
-                // We're going to do this as an LDC string in ISO-8859 and read it back at runtime
-                LdcInsnNode(bytes.toString(Charsets.ISO_8859_1)),
-                LdcInsnNode("ISO-8859-1"),
-                // Ug, can't do func refs on native types here...
-                MethodInsnNode(Opcodes.INVOKEVIRTUAL, String::class.ref.asmName,
-                    "getBytes", "(Ljava/lang/String;)[B", false),
-                0.const,
-                bytes.size.const,
-                forceFnType<ByteBuffer.(ByteArray, Int, Int) -> ByteBuffer>(ByteBuffer::put).invokeVirtual(),
+                TypeInsnNode(Opcodes.CHECKCAST, memType.asmName)
+            ).addInsns(
+                // We're going to do this as an LDC string in ISO-8859 and read it back at runtime. However,
+                // due to JVM limits, we can't have a string > 65536 chars, so I'll chunk it every 65500 chars.
+                bytes.chunked(65500).flatMap { bytes ->
+                    sequenceOf(
+                        LdcInsnNode(bytes.toString(Charsets.ISO_8859_1)),
+                        LdcInsnNode("ISO-8859-1"),
+                        // Ug, can't do func refs on native types here...
+                        MethodInsnNode(Opcodes.INVOKEVIRTUAL, String::class.ref.asmName,
+                            "getBytes", "(Ljava/lang/String;)[B", false),
+                        0.const,
+                        bytes.size.const,
+                        forceFnType<ByteBuffer.(ByteArray, Int, Int) -> ByteBuffer>(ByteBuffer::put).invokeVirtual()
+                    )
+                }.toList()
+            ).addInsns(
                 InsnNode(Opcodes.POP)
             )
 
diff --git a/compiler/src/test/kotlin/asmble/compile/jvm/LargeDataTest.kt b/compiler/src/test/kotlin/asmble/compile/jvm/LargeDataTest.kt
new file mode 100644
index 0000000..0c189df
--- /dev/null
+++ b/compiler/src/test/kotlin/asmble/compile/jvm/LargeDataTest.kt
@@ -0,0 +1,43 @@
+package asmble.compile.jvm
+
+import asmble.TestBase
+import asmble.ast.Node
+import asmble.run.jvm.ScriptContext
+import asmble.util.get
+import org.junit.Test
+import java.nio.ByteBuffer
+import java.util.*
+import kotlin.test.assertEquals
+
+class LargeDataTest : TestBase() {
+    @Test
+    fun testLargeData() {
+        // This previously failed because string constants can't be longer than 65536 chars
+        val mod = Node.Module(
+            memories = listOf(Node.Type.Memory(
+                limits = Node.ResizableLimits(initial = 2, maximum = 2)
+            )),
+            data = listOf(Node.Data(
+                index = 0,
+                offset = listOf(Node.Instr.I32Const(0)),
+                data = ByteArray(70000) { 'a'.toByte() }
+            ))
+        )
+        val ctx = ClsContext(
+            packageName = "test",
+            className = "Temp" + UUID.randomUUID().toString().replace("-", ""),
+            mod = mod,
+            logger = logger
+        )
+        AstToAsm.fromModule(ctx)
+        val cls = ScriptContext.SimpleClassLoader(javaClass.classLoader, logger).fromBuiltContext(ctx)
+        // Instantiate it, get the memory out, and check it
+        val field = cls.getDeclaredField("memory").apply { isAccessible = true }
+        val buf = field[cls.newInstance()] as ByteBuffer
+        // Grab all + 1 and check values
+        val bytes = ByteArray(70001).also { buf.get(0, it) }
+        bytes.forEachIndexed { index, byte ->
+            assertEquals(if (index == 70000) 0.toByte() else 'a'.toByte(), byte)
+        }
+    }
+}
\ No newline at end of file