@ohos-ports/lmdb 2.5.3-beta.0

package/src/compression.cpp

@@ -0,0 +1,222 @@
+#include "lz4.h"
+#include "lmdb-js.h"
+#include <atomic>
+
+using namespace Napi;
+
+thread_local LZ4_stream_t* Compression::stream = nullptr;
+Compression::Compression(const CallbackInfo& info) : ObjectWrap<Compression>(info) {
+	unsigned int compressionThreshold = 1000;
+	char* dictionary = nullptr;
+	size_t dictSize = 0;
+	if (info[0].IsObject()) {
+		auto dictionaryOption = info[0].As<Object>().Get("dictionary");
+		if (!dictionaryOption.IsUndefined()) {
+			if (!dictionaryOption.IsTypedArray()) {
+				throwError(info.Env(), "Dictionary must be a buffer");
+				return;
+			}
+			napi_get_typedarray_info(info.Env(), dictionaryOption, nullptr, &dictSize, (void**) &dictionary, nullptr, nullptr);
+			dictSize = (dictSize >> 3) << 3; // make sure it is word-aligned
+		}
+		auto thresholdOption = info[0].As<Object>().Get("threshold");
+		if (thresholdOption.IsNumber()) {
+			compressionThreshold = thresholdOption.As<Number>();
+		}
+	}
+	this->dictionary = this->compressDictionary = dictionary;
+	this->dictionarySize = dictSize;
+	this->decompressTarget = dictionary + dictSize;
+	this->decompressSize = 0;
+	this->acceleration = 1;
+	this->compressionThreshold = compressionThreshold;
+	info.This().As<Object>().Set("address", Number::New(info.Env(), (double) (size_t) this));
+}
+
+Napi::Value Compression::setBuffer(const CallbackInfo& info) {
+	napi_get_typedarray_info(info.Env(), info[0], nullptr, nullptr, (void**) &this->decompressTarget, nullptr, nullptr);
+	this->decompressSize = info[1].As<Number>();
+	napi_get_typedarray_info(info.Env(), info[2], nullptr, nullptr, (void**) &this->dictionary, nullptr, nullptr);
+	this->dictionarySize = info[3].As<Number>();
+	return info.Env().Undefined();
+}
+void Compression::decompress(MDB_val& data, bool &isValid, bool canAllocate) {
+	uint32_t uncompressedLength;
+	int compressionHeaderSize;
+	uint32_t compressedLength = data.mv_size;
+	unsigned char* charData = (unsigned char*) data.mv_data;
+
+	if (charData[0] == 254) {
+		uncompressedLength = ((uint32_t)charData[1] << 16) | ((uint32_t)charData[2] << 8) | (uint32_t)charData[3];
+		compressionHeaderSize = 4;
+	}
+	else if (charData[0] == 255) {
+		uncompressedLength = ((uint32_t)charData[4] << 24) | ((uint32_t)charData[5] << 16) | ((uint32_t)charData[6] << 8) | (uint32_t)charData[7];
+		compressionHeaderSize = 8;
+	}
+	else {
+		fprintf(stderr, "Unknown status byte %u\n", charData[0]);
+		//if (canAllocate)
+		//	Nan::ThrowError("Unknown status byte");
+		isValid = false;
+		return;
+	}
+	data.mv_data = decompressTarget;
+	data.mv_size = uncompressedLength;
+	//TODO: For larger blocks with known encoding, it might make sense to allocate space for it and use an ExternalString
+	//fprintf(stdout, "compressed size %u uncompressedLength %u, first byte %u\n", data.mv_size, uncompressedLength, charData[compressionHeaderSize]);
+	if (uncompressedLength > decompressSize) {
+		isValid = false;
+		return;
+	}
+	int written = LZ4_decompress_safe_usingDict(
+		(char*)charData + compressionHeaderSize, decompressTarget,
+		compressedLength - compressionHeaderSize, decompressSize,
+		dictionary, dictionarySize);
+	//fprintf(stdout, "first uncompressed byte %X %X %X %X %X %X\n", uncompressedData[0], uncompressedData[1], uncompressedData[2], uncompressedData[3], uncompressedData[4], uncompressedData[5]);
+	if (written < 0) {
+		fprintf(stderr, "Failed to decompress data %u %u bytes:\n", compressionHeaderSize, uncompressedLength);
+		for (uint32_t i = 0; i < compressedLength; i++) {
+			fprintf(stderr, "%u ", charData[i]);
+		}
+		//if (canAllocate)
+		//	Nan::ThrowError("Failed to decompress data");
+		isValid = false;
+		return;
+	}
+	isValid = true;
+}
+
+int Compression::compressInstruction(EnvWrap* env, double* compressionAddress) {
+	MDB_val value;
+	value.mv_data = (void*)((size_t) * (compressionAddress - 1));
+	value.mv_size = *(((uint32_t*)compressionAddress) - 3);
+	argtokey_callback_t compressedData = compress(&value, nullptr);
+	if (compressedData) {
+		*(((uint32_t*)compressionAddress) - 3) = value.mv_size;
+		*((size_t*)(compressionAddress - 1)) = (size_t)value.mv_data;
+		int64_t status = std::atomic_exchange((std::atomic<int64_t>*) compressionAddress, (int64_t) 0);
+		if (status == 1 && env) {
+			pthread_mutex_lock(env->writingLock);
+			pthread_cond_signal(env->writingCond);
+			pthread_mutex_unlock(env->writingLock);
+			//fprintf(stderr, "sent compression completion signal\n");
+		}
+		//fprintf(stdout, "compressed to %p %u %u %p\n", value.mv_data, value.mv_size, status, env);
+		return 0;
+	} else {
+		fprintf(stdout, "failed to compress\n");
+		return 1;
+	}
+}
+
+argtokey_callback_t Compression::compress(MDB_val* value, void (*freeValue)(MDB_val&)) {
+	size_t dataLength = value->mv_size;
+	char* data = (char*)value->mv_data;
+	if (value->mv_size < compressionThreshold && !(value->mv_size > 0 && ((uint8_t*)data)[0] >= 250))
+		return freeValue; // don't compress if less than threshold (but we must compress if the first byte is the compression indicator)
+	bool longSize = dataLength >= 0x1000000;
+	int prefixSize = (longSize ? 8 : 4);
+	int maxCompressedSize = LZ4_COMPRESSBOUND(dataLength);
+	char* compressed = new char[maxCompressedSize + prefixSize];
+	//fprintf(stdout, "compressing %u\n", dataLength);
+	if (!stream)
+		stream = LZ4_createStream();
+	LZ4_loadDict(stream, compressDictionary, dictionarySize);
+	int compressedSize = LZ4_compress_fast_continue(stream, data, compressed + prefixSize, dataLength, maxCompressedSize, acceleration);
+	if (compressedSize > 0) {
+		if (freeValue)
+			freeValue(*value);
+		uint8_t* compressedData = (uint8_t*)compressed;
+		if (longSize) {
+			compressedData[0] = 255;
+			compressedData[2] = (uint8_t)(dataLength >> 40u);
+			compressedData[3] = (uint8_t)(dataLength >> 32u);
+			compressedData[4] = (uint8_t)(dataLength >> 24u);
+			compressedData[5] = (uint8_t)(dataLength >> 16u);
+			compressedData[6] = (uint8_t)(dataLength >> 8u);
+			compressedData[7] = (uint8_t)dataLength;
+		}
+		else {
+			compressedData[0] = 254;
+			compressedData[1] = (uint8_t)(dataLength >> 16u);
+			compressedData[2] = (uint8_t)(dataLength >> 8u);
+			compressedData[3] = (uint8_t)dataLength;
+		}
+		value->mv_size = compressedSize + prefixSize;
+		value->mv_data = compressed;
+		return ([](MDB_val &value) -> void {
+			delete[] (char*)value.mv_data;
+		});
+	}
+	else {
+		delete[] compressed;
+		return nullptr;
+	}
+}
+
+class CompressionWorker : public AsyncWorker {
+  public:
+	CompressionWorker(EnvWrap* env, double* compressionAddress, const Function& callback)
+	  : AsyncWorker(callback), env(env), compressionAddress(compressionAddress) {}
+
+
+	void Execute() {
+		uint64_t compressionPointer;
+		compressionPointer = std::atomic_exchange((std::atomic<int64_t>*) compressionAddress, (int64_t) 2);
+		if (compressionPointer > 1) {
+			Compression* compression = (Compression*)(size_t) * ((double*)&compressionPointer);
+			compression->compressInstruction(env, compressionAddress);
+		}
+	}
+	void OnOK() {
+		// don't actually call the callback, no need
+	}
+
+  private:
+	EnvWrap* env;
+	double* compressionAddress;
+};
+
+NAPI_FUNCTION(EnvWrap::compress) {
+	ARGS(3)
+    GET_INT64_ARG(0);
+    EnvWrap* ew = (EnvWrap*) i64;	
+    napi_get_value_int64(env, args[1], &i64);
+    double* compressionAddress = (double*) i64;
+	CompressionWorker* worker = new CompressionWorker(ew, (double*) compressionAddress, Function(env, args[2]));
+	worker->Queue();
+	RETURN_UNDEFINED;
+}
+
+void Compression::setupExports(Napi::Env env, Object exports) {
+	Function CompressionClass = DefineClass(env, "Compression", {
+		Compression::InstanceMethod("setBuffer", &Compression::setBuffer),
+	});
+	exports.Set("Compression", CompressionClass);
+//	compressionTpl->InstanceTemplate()->SetInternalFieldCount(1);
+}
+
+
+// This file contains code from the node-lmdb project
+// Copyright (c) 2013-2017 Timur Kristóf
+// Copyright (c) 2021 Kristopher Tate
+// Licensed to you under the terms of the MIT license
+//
+// Permission is hereby granted, free of charge, to any person obtaining a copy
+// of this software and associated documentation files (the "Software"), to deal
+// in the Software without restriction, including without limitation the rights
+// to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+// copies of the Software, and to permit persons to whom the Software is
+// furnished to do so, subject to the following conditions:
+
+// The above copyright notice and this permission notice shall be included in
+// all copies or substantial portions of the Software.
+
+// THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+// IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+// FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+// AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+// LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+// OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+// THE SOFTWARE.

package/src/cursor.cpp

@@ -0,0 +1,344 @@
+#include "lmdb-js.h"
+#include <string.h>
+
+using namespace Napi;
+
+CursorWrap::CursorWrap(const CallbackInfo& info) : Napi::ObjectWrap<CursorWrap>(info) {
+	this->keyType = LmdbKeyType::StringKey;
+	this->freeKey = nullptr;
+	this->endKey.mv_size = 0; // indicates no end key (yet)
+	if (info.Length() < 1) {
+		throwError(info.Env(), "Wrong number of arguments");
+		return;
+	}
+
+	DbiWrap *dw;
+	napi_unwrap(info.Env(), info[0], (void**)&dw);
+
+	// Open the cursor
+	MDB_cursor *cursor;
+	MDB_txn *txn = dw->ew->getReadTxn();
+	int rc = mdb_cursor_open(txn, dw->dbi, &cursor);
+	if (rc != 0) {
+		throwLmdbError(info.Env(), rc);
+		return;
+	}
+	info.This().As<Object>().Set("address", Number::New(info.Env(), (size_t) this));
+	this->cursor = cursor;
+	this->dw = dw;
+	this->txn = txn;
+	this->keyType = keyType;
+}
+
+CursorWrap::~CursorWrap() {
+	if (this->cursor) {
+		// Don't close cursor here, it is possible that the environment may already be closed, which causes it to crash
+		//mdb_cursor_close(this->cursor);
+	}
+	if (this->freeKey) {
+		this->freeKey(this->key);
+	}
+}
+
+Value CursorWrap::close(const CallbackInfo& info) {
+	if (!this->cursor) {
+	  return throwError(info.Env(), "cursor.close: Attempt to close a closed cursor!");
+	}
+	mdb_cursor_close(this->cursor);
+	this->cursor = nullptr;
+	return info.Env().Undefined();
+}
+
+Value CursorWrap::del(const CallbackInfo& info) {
+	int flags = 0;
+
+	if (info.Length() == 1) {
+		if (!info[0].IsObject()) {
+			return throwError(info.Env(), "cursor.del: Invalid options argument. It should be an object.");
+		}
+		
+		auto options = info[0].As<Object>();
+		setFlagFromValue(&flags, MDB_NODUPDATA, "noDupData", false, options);
+	}
+
+	int rc = mdb_cursor_del(this->cursor, flags);
+	if (rc != 0) {
+		return throwLmdbError(info.Env(), rc);
+	}
+	return info.Env().Undefined();
+}
+int CursorWrap::returnEntry(int lastRC, MDB_val &key, MDB_val &data) {
+	if (lastRC) {
+		if (lastRC == MDB_NOTFOUND)
+			return 0;
+		else {
+			return lastRC > 0 ? -lastRC : lastRC;
+		}
+	}
+	if (endKey.mv_size > 0) {
+		int comparison;
+		if (flags & 0x800)
+			comparison = mdb_dcmp(txn, dw->dbi, &endKey, &data);
+		else
+			comparison = mdb_cmp(txn, dw->dbi, &endKey, &key);
+		if ((flags & 0x400) ? comparison >= 0 : (comparison <= 0)) {
+			return 0;
+		}
+	}
+	char* keyBuffer = dw->ew->keyBuffer;
+	if (flags & 0x100) {
+		bool result = getVersionAndUncompress(data, dw);
+		if (result)
+			result = valToBinaryFast(data, dw);
+		*((size_t*)keyBuffer) = data.mv_size;
+	}
+	if (!(flags & 0x800)) {
+		memcpy(keyBuffer + 32, key.mv_data, key.mv_size);
+		*(keyBuffer + 32 + key.mv_size) = 0; // make sure it is null terminated for the sake of better ordered-binary performance
+	}
+
+	return key.mv_size;
+}
+
+const int START_ADDRESS_POSITION = 4064;
+int32_t CursorWrap::doPosition(uint32_t offset, uint32_t keySize, uint64_t endKeyAddress) {
+	//char* keyBuffer = dw->ew->keyBuffer;
+	MDB_val key, data;
+	int rc;
+	if (flags & 0x2000) { // TODO: check the txn_id to determine if we need to renew
+		rc = mdb_cursor_renew(txn = dw->ew->getReadTxn(), cursor);
+		if (rc) {
+			if (rc > 0)
+				rc = -rc;
+			return rc;
+		}
+	}
+	if (endKeyAddress) {
+		uint32_t* keyBuffer = (uint32_t*) endKeyAddress;
+		endKey.mv_size = *keyBuffer;
+		endKey.mv_data = (char*)(keyBuffer + 1);
+	} else
+		endKey.mv_size = 0;
+	iteratingOp = (flags & 0x400) ?
+		(flags & 0x100) ?
+			(flags & 0x800) ? MDB_PREV_DUP : MDB_PREV :
+			MDB_PREV_NODUP :
+		(flags & 0x100) ?
+			(flags & 0x800) ? MDB_NEXT_DUP : MDB_NEXT :
+			MDB_NEXT_NODUP;
+	key.mv_size = keySize;
+	key.mv_data = dw->ew->keyBuffer;
+	if (keySize == 0) {
+		rc = mdb_cursor_get(cursor, &key, &data, flags & 0x400 ? MDB_LAST : MDB_FIRST);  
+	} else {
+		if (flags & 0x800) { // only values for this key
+			// take the next part of the key buffer as a pointer to starting data
+			uint32_t* startValueBuffer = (uint32_t*)(size_t)(*(double*)(dw->ew->keyBuffer + START_ADDRESS_POSITION));
+			data.mv_size = endKeyAddress ? *((uint32_t*)startValueBuffer) : 0;
+			data.mv_data = startValueBuffer + 1;
+			if (flags & 0x400) {// reverse through values
+				MDB_val startValue = data; // save it for comparison
+				rc = mdb_cursor_get(cursor, &key, &data, data.mv_size ? MDB_GET_BOTH_RANGE : MDB_SET_KEY);
+				if (rc) {
+					if (startValue.mv_size) {
+						// value specified, but not found, so find key and go to last item
+						rc = mdb_cursor_get(cursor, &key, &data, MDB_SET_KEY);
+						if (!rc)
+							rc = mdb_cursor_get(cursor, &key, &data, MDB_LAST_DUP);
+					} // else just couldn't find the key
+				} else { // found entry
+					if (startValue.mv_size == 0) // no value specified, so go to last value
+						rc = mdb_cursor_get(cursor, &key, &data, MDB_LAST_DUP);
+					else if (mdb_dcmp(txn, dw->dbi, &startValue, &data)) // the range found the next value *after* the start
+						rc = mdb_cursor_get(cursor, &key, &data, MDB_PREV_DUP);
+				}
+			} else // forward, just do a get by range
+				rc = mdb_cursor_get(cursor, &key, &data, data.mv_size ?
+					(flags & 0x4000) ? MDB_GET_BOTH : MDB_GET_BOTH_RANGE : MDB_SET_KEY);
+
+			if (rc == MDB_NOTFOUND)
+				return 0;
+			if (flags & 0x1000 && (!endKeyAddress || (flags & 0x4000))) {
+				size_t count;
+				rc = mdb_cursor_count(cursor, &count);
+				if (rc)
+					return rc > 0 ? -rc : rc;
+				return count;
+			}
+		} else {
+			if (flags & 0x400) {// reverse
+				MDB_val firstKey = key; // save it for comparison
+				rc = mdb_cursor_get(cursor, &key, &data, MDB_SET_RANGE);
+				if (rc)
+					rc = mdb_cursor_get(cursor, &key, &data, MDB_LAST);
+				else if (mdb_cmp(txn, dw->dbi, &firstKey, &key)) // the range found the next entry *after* the start
+					rc = mdb_cursor_get(cursor, &key, &data, MDB_PREV);
+			} else // forward, just do a get by range
+				rc = mdb_cursor_get(cursor, &key, &data, (flags & 0x4000) ? MDB_SET_KEY : MDB_SET_RANGE);
+		}
+	}
+	while (offset-- > 0 && !rc) {
+		rc = mdb_cursor_get(cursor, &key, &data, iteratingOp);
+	}
+	if (flags & 0x1000) {
+		uint32_t count = 0;
+		bool useCursorCount = false;
+		// if we are in a dupsort database, and we are iterating over all entries, we can just count all the values for each key
+		if (dw->flags & MDB_DUPSORT) {
+			if (iteratingOp == MDB_PREV) {
+				iteratingOp = MDB_PREV_NODUP;
+				useCursorCount = true;
+			}
+			if (iteratingOp == MDB_NEXT) {
+				iteratingOp = MDB_NEXT_NODUP;
+				useCursorCount = true;
+			}
+		}
+
+		while (!rc) {
+			if (endKey.mv_size > 0) {
+				int comparison;
+				if (flags & 0x800)
+					comparison = mdb_dcmp(txn, dw->dbi, &endKey, &data);
+				else
+					comparison = mdb_cmp(txn, dw->dbi, &endKey, &key);
+				if ((flags & 0x400) ? comparison >= 0 : (comparison <=0)) {
+					return count;
+				}
+			}
+			if (useCursorCount) {
+				size_t countForKey;
+				rc = mdb_cursor_count(cursor, &countForKey);
+				if (rc) {
+					if (rc > 0)
+						rc = -rc;
+					return rc;
+				}
+				count += countForKey;
+			} else
+				count++;
+			rc = mdb_cursor_get(cursor, &key, &data, iteratingOp);
+		}
+		return count;
+	}
+	// TODO: Handle count?
+	return returnEntry(rc, key, data);
+}
+NAPI_FUNCTION(position) {
+	ARGS(5)
+    GET_INT64_ARG(0);
+    CursorWrap* cw = (CursorWrap*) i64;
+	GET_UINT32_ARG(cw->flags, 1);
+	uint32_t offset;
+	GET_UINT32_ARG(offset, 2);
+	uint32_t keySize;
+	GET_UINT32_ARG(keySize, 3);
+    napi_get_value_int64(env, args[4], &i64);
+    int64_t endKeyAddress = i64;
+	int32_t result = cw->doPosition(offset, keySize, endKeyAddress);
+	RETURN_INT32(result);
+}
+int32_t positionFFI(double cwPointer, uint32_t flags, uint32_t offset, uint32_t keySize, uint64_t endKeyAddress) {
+	CursorWrap* cw = (CursorWrap*) (size_t) cwPointer;
+	DbiWrap* dw = cw->dw;
+	dw->getFast = true;
+	cw->flags = flags;
+	return cw->doPosition(offset, keySize, endKeyAddress);
+}
+
+NAPI_FUNCTION(iterate) {
+	ARGS(1)
+    GET_INT64_ARG(0);
+    CursorWrap* cw = (CursorWrap*) i64;
+	MDB_val key, data;
+	int rc = mdb_cursor_get(cw->cursor, &key, &data, cw->iteratingOp);
+	RETURN_INT32(cw->returnEntry(rc, key, data));
+}
+
+int32_t iterateFFI(double cwPointer) {
+	CursorWrap* cw = (CursorWrap*) (size_t) cwPointer;
+	DbiWrap* dw = cw->dw;
+	dw->getFast = true;
+	MDB_val key, data;
+	int rc = mdb_cursor_get(cw->cursor, &key, &data, cw->iteratingOp);
+	return cw->returnEntry(rc, key, data);
+}
+
+
+NAPI_FUNCTION(getCurrentValue) {
+	ARGS(1)
+    GET_INT64_ARG(0);
+    CursorWrap* cw = (CursorWrap*) i64;
+	MDB_val key, data;
+	int rc = mdb_cursor_get(cw->cursor, &key, &data, MDB_GET_CURRENT);
+	RETURN_INT32(cw->returnEntry(rc, key, data));
+}
+
+napi_finalize noopCursor = [](napi_env, void *, void *) {
+	// Data belongs to LMDB, we shouldn't free it here
+};
+NAPI_FUNCTION(getCurrentShared) {
+	ARGS(1)
+    GET_INT64_ARG(0);
+    CursorWrap* cw = (CursorWrap*) i64;
+	MDB_val key, data;
+	int rc = mdb_cursor_get(cw->cursor, &key, &data, MDB_GET_CURRENT);
+	if (rc)
+		RETURN_INT32(cw->returnEntry(rc, key, data));
+	getVersionAndUncompress(data, cw->dw);
+	napi_create_external_buffer(env, data.mv_size,
+		(char*) data.mv_data, noopCursor, nullptr, &returnValue);
+	return returnValue;
+}
+
+NAPI_FUNCTION(renew) {
+	ARGS(1)
+    GET_INT64_ARG(0);
+    CursorWrap* cw = (CursorWrap*) i64;
+	mdb_cursor_renew(cw->txn = cw->dw->ew->getReadTxn(), cw->cursor);
+	RETURN_UNDEFINED;
+}
+
+void CursorWrap::setupExports(Napi::Env env, Object exports) {
+	// CursorWrap: Prepare constructor template
+	Function CursorClass = DefineClass(env, "Cursor", {
+	// CursorWrap: Add functions to the prototype
+		CursorWrap::InstanceMethod("close", &CursorWrap::close),
+		CursorWrap::InstanceMethod("del", &CursorWrap::del),
+	});
+	EXPORT_NAPI_FUNCTION("position", position);
+	EXPORT_NAPI_FUNCTION("iterate", iterate);
+	EXPORT_NAPI_FUNCTION("getCurrentValue", getCurrentValue);
+	EXPORT_NAPI_FUNCTION("getCurrentShared", getCurrentShared);
+	EXPORT_NAPI_FUNCTION("renew", renew);
+	EXPORT_FUNCTION_ADDRESS("positionPtr", positionFFI);
+	EXPORT_FUNCTION_ADDRESS("iteratePtr", iterateFFI);
+
+	exports.Set("Cursor", CursorClass);
+
+//	cursorTpl->InstanceTemplate()->SetInternalFieldCount(1);
+}
+
+// This file contains code from the node-lmdb project
+// Copyright (c) 2013-2017 Timur Kristóf
+// Copyright (c) 2021 Kristopher Tate
+// Licensed to you under the terms of the MIT license
+//
+// Permission is hereby granted, free of charge, to any person obtaining a copy
+// of this software and associated documentation files (the "Software"), to deal
+// in the Software without restriction, including without limitation the rights
+// to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+// copies of the Software, and to permit persons to whom the Software is
+// furnished to do so, subject to the following conditions:
+
+// The above copyright notice and this permission notice shall be included in
+// all copies or substantial portions of the Software.
+
+// THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+// IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+// FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+// AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+// LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+// OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+// THE SOFTWARE.
+