Lucene50NormsConsumer(
SegmentWriteState state,
String dataCodec,
String dataExtension,
String metaCodec,
String metaExtension)
throws IOException {
boolean success = false;
try {
String dataName =
IndexFileNames.segmentFileName(
state.segmentInfo.name, state.segmentSuffix, dataExtension);
data = state.directory.createOutput(dataName, state.context);
CodecUtil.writeIndexHeader(
data, dataCodec, VERSION_CURRENT, state.segmentInfo.getId(), state.segmentSuffix);
String metaName =
IndexFileNames.segmentFileName(
state.segmentInfo.name, state.segmentSuffix, metaExtension);
meta = state.directory.createOutput(metaName, state.context);
CodecUtil.writeIndexHeader(
meta, metaCodec, VERSION_CURRENT, state.segmentInfo.getId(), state.segmentSuffix);
success = true;
} finally {
if (!success) {
IOUtils.closeWhileHandlingException(this);
}
}
}
public SepPostingsReader(
Directory dir,
FieldInfos fieldInfos,
SegmentInfo segmentInfo,
IOContext context,
IntStreamFactory intFactory,
String segmentSuffix)
throws IOException {
boolean success = false;
try {
final String docFileName =
IndexFileNames.segmentFileName(
segmentInfo.name, segmentSuffix, SepPostingsWriter.DOC_EXTENSION);
docIn = intFactory.openInput(dir, docFileName, context);
skipIn =
dir.openInput(
IndexFileNames.segmentFileName(
segmentInfo.name, segmentSuffix, SepPostingsWriter.SKIP_EXTENSION),
context);
if (fieldInfos.hasFreq()) {
freqIn =
intFactory.openInput(
dir,
IndexFileNames.segmentFileName(
segmentInfo.name, segmentSuffix, SepPostingsWriter.FREQ_EXTENSION),
context);
} else {
freqIn = null;
}
if (fieldInfos.hasProx()) {
posIn =
intFactory.openInput(
dir,
IndexFileNames.segmentFileName(
segmentInfo.name, segmentSuffix, SepPostingsWriter.POS_EXTENSION),
context);
payloadIn =
dir.openInput(
IndexFileNames.segmentFileName(
segmentInfo.name, segmentSuffix, SepPostingsWriter.PAYLOAD_EXTENSION),
context);
} else {
posIn = null;
payloadIn = null;
}
success = true;
} finally {
if (!success) {
close();
}
}
}
public CompletionFieldsConsumer(SegmentWriteState state) throws IOException {
this.delegatesFieldsConsumer = delegatePostingsFormat.fieldsConsumer(state);
String suggestFSTFile =
IndexFileNames.segmentFileName(state.segmentInfo.name, state.segmentSuffix, EXTENSION);
IndexOutput output = null;
boolean success = false;
try {
output = state.directory.createOutput(suggestFSTFile, state.context);
CodecUtil.writeIndexHeader(
output,
CODEC_NAME,
SUGGEST_VERSION_CURRENT,
state.segmentInfo.getId(),
state.segmentSuffix);
/*
* we write the delegate postings format name so we can load it
* without getting an instance in the ctor
*/
output.writeString(delegatePostingsFormat.getName());
output.writeString(writeProvider.getName());
this.suggestFieldsConsumer = writeProvider.consumer(output);
success = true;
} finally {
if (!success) {
IOUtils.closeWhileHandlingException(output);
}
}
}
/** Full constructor */
public Lucene60PointsWriter(
SegmentWriteState writeState, int maxPointsInLeafNode, double maxMBSortInHeap)
throws IOException {
assert writeState.fieldInfos.hasPointValues();
this.writeState = writeState;
this.maxPointsInLeafNode = maxPointsInLeafNode;
this.maxMBSortInHeap = maxMBSortInHeap;
String dataFileName =
IndexFileNames.segmentFileName(
writeState.segmentInfo.name,
writeState.segmentSuffix,
Lucene60PointsFormat.DATA_EXTENSION);
dataOut = writeState.directory.createOutput(dataFileName, writeState.context);
boolean success = false;
try {
CodecUtil.writeIndexHeader(
dataOut,
Lucene60PointsFormat.DATA_CODEC_NAME,
Lucene60PointsFormat.DATA_VERSION_CURRENT,
writeState.segmentInfo.getId(),
writeState.segmentSuffix);
success = true;
} finally {
if (success == false) {
IOUtils.closeWhileHandlingException(dataOut);
}
}
}
public BloomFilteredFieldsProducer(SegmentReadState state) throws IOException {
String bloomFileName =
IndexFileNames.segmentFileName(
state.segmentInfo.name, state.segmentSuffix, BLOOM_EXTENSION);
IndexInput bloomIn = null;
boolean success = false;
try {
bloomIn = state.directory.openInput(bloomFileName, state.context);
CodecUtil.checkHeader(bloomIn, BLOOM_CODEC_NAME, BLOOM_CODEC_VERSION, BLOOM_CODEC_VERSION);
// // Load the hash function used in the BloomFilter
// hashFunction = HashFunction.forName(bloomIn.readString());
// Load the delegate postings format
PostingsFormat delegatePostingsFormat = PostingsFormat.forName(bloomIn.readString());
this.delegateFieldsProducer = delegatePostingsFormat.fieldsProducer(state);
int numBlooms = bloomIn.readInt();
for (int i = 0; i < numBlooms; i++) {
int fieldNum = bloomIn.readInt();
FuzzySet bloom = FuzzySet.deserialize(bloomIn);
FieldInfo fieldInfo = state.fieldInfos.fieldInfo(fieldNum);
bloomsByFieldName.put(fieldInfo.name, bloom);
}
IOUtils.close(bloomIn);
success = true;
} finally {
if (!success) {
IOUtils.closeWhileHandlingException(bloomIn, delegateFieldsProducer);
}
}
}
@Override
public DocValuesConsumer normsConsumer(SegmentWriteState state) throws IOException {
String filename =
IndexFileNames.segmentFileName(
state.segmentInfo.name, "nrm", IndexFileNames.COMPOUND_FILE_EXTENSION);
return new Lucene40DocValuesWriter(
state, filename, Lucene40FieldInfosReader.LEGACY_NORM_TYPE_KEY);
}
public CompletionFieldsProducer(SegmentReadState state) throws IOException {
String suggestFSTFile =
IndexFileNames.segmentFileName(state.segmentInfo.name, state.segmentSuffix, EXTENSION);
IndexInput input = state.directory.openInput(suggestFSTFile, state.context);
if (state.segmentInfo.getVersion().onOrAfter(Version.LUCENE_6_2_0)) {
// Lucene 6.2.0+ requires all index files to use index header, but prior to that we used an
// ordinary codec header:
version =
CodecUtil.checkIndexHeader(
input,
CODEC_NAME,
SUGGEST_CODEC_VERSION,
SUGGEST_VERSION_CURRENT,
state.segmentInfo.getId(),
state.segmentSuffix);
} else {
version =
CodecUtil.checkHeader(
input, CODEC_NAME, SUGGEST_CODEC_VERSION, SUGGEST_VERSION_CURRENT);
}
FieldsProducer delegateProducer = null;
boolean success = false;
try {
PostingsFormat delegatePostingsFormat = PostingsFormat.forName(input.readString());
String providerName = input.readString();
CompletionLookupProvider completionLookupProvider = providers.get(providerName);
if (completionLookupProvider == null) {
throw new IllegalStateException(
"no provider with name [" + providerName + "] registered");
}
// TODO: we could clone the ReadState and make it always forward IOContext.MERGE to prevent
// unecessary heap usage?
delegateProducer = delegatePostingsFormat.fieldsProducer(state);
/*
* If we are merging we don't load the FSTs at all such that we
* don't consume so much memory during merge
*/
if (state.context.context != Context.MERGE) {
// TODO: maybe we can do this in a fully lazy fashion based on some configuration
// eventually we should have some kind of curciut breaker that prevents us from going OOM
// here
// with some configuration
this.lookupFactory = completionLookupProvider.load(input);
} else {
this.lookupFactory = null;
}
this.delegateProducer = delegateProducer;
success = true;
} finally {
if (!success) {
IOUtils.closeWhileHandlingException(delegateProducer, input);
} else {
IOUtils.close(input);
}
}
}
@Override
public void write(Directory dir, SegmentInfo si, IOContext context) throws IOException {
String dataFile = IndexFileNames.segmentFileName(si.name, "", DATA_EXTENSION);
int numFiles = si.files().size();
String names[] = si.files().toArray(new String[numFiles]);
Arrays.sort(names);
long startOffsets[] = new long[numFiles];
long endOffsets[] = new long[numFiles];
BytesRefBuilder scratch = new BytesRefBuilder();
try (IndexOutput out = dir.createOutput(dataFile, context)) {
for (int i = 0; i < names.length; i++) {
// write header for file
SimpleTextUtil.write(out, HEADER);
SimpleTextUtil.write(out, names[i], scratch);
SimpleTextUtil.writeNewline(out);
// write bytes for file
startOffsets[i] = out.getFilePointer();
try (IndexInput in = dir.openInput(names[i], IOContext.READONCE)) {
out.copyBytes(in, in.length());
}
endOffsets[i] = out.getFilePointer();
}
long tocPos = out.getFilePointer();
// write CFS table
SimpleTextUtil.write(out, TABLE);
SimpleTextUtil.write(out, Integer.toString(numFiles), scratch);
SimpleTextUtil.writeNewline(out);
for (int i = 0; i < names.length; i++) {
SimpleTextUtil.write(out, TABLENAME);
SimpleTextUtil.write(out, names[i], scratch);
SimpleTextUtil.writeNewline(out);
SimpleTextUtil.write(out, TABLESTART);
SimpleTextUtil.write(out, Long.toString(startOffsets[i]), scratch);
SimpleTextUtil.writeNewline(out);
SimpleTextUtil.write(out, TABLEEND);
SimpleTextUtil.write(out, Long.toString(endOffsets[i]), scratch);
SimpleTextUtil.writeNewline(out);
}
DecimalFormat df =
new DecimalFormat(OFFSETPATTERN, DecimalFormatSymbols.getInstance(Locale.ROOT));
SimpleTextUtil.write(out, TABLEPOS);
SimpleTextUtil.write(out, df.format(tocPos), scratch);
SimpleTextUtil.writeNewline(out);
}
}
@Override
public void write(
Directory directory,
SegmentInfo segmentInfo,
String segmentSuffix,
FieldInfos infos,
IOContext context)
throws IOException {
final String fileName =
IndexFileNames.segmentFileName(
segmentInfo.name, segmentSuffix, Lucene46FieldInfosFormat.EXTENSION);
try (IndexOutput output = directory.createOutput(fileName, context)) {
CodecUtil.writeHeader(
output, Lucene46FieldInfosFormat.CODEC_NAME, Lucene46FieldInfosFormat.FORMAT_CURRENT);
output.writeVInt(infos.size());
for (FieldInfo fi : infos) {
IndexOptions indexOptions = fi.getIndexOptions();
byte bits = 0x0;
if (fi.hasVectors()) bits |= Lucene46FieldInfosFormat.STORE_TERMVECTOR;
if (fi.omitsNorms()) bits |= Lucene46FieldInfosFormat.OMIT_NORMS;
if (fi.hasPayloads()) bits |= Lucene46FieldInfosFormat.STORE_PAYLOADS;
if (fi.getIndexOptions() != IndexOptions.NONE) {
bits |= Lucene46FieldInfosFormat.IS_INDEXED;
assert indexOptions.compareTo(IndexOptions.DOCS_AND_FREQS_AND_POSITIONS) >= 0
|| !fi.hasPayloads();
if (indexOptions == IndexOptions.DOCS) {
bits |= Lucene46FieldInfosFormat.OMIT_TERM_FREQ_AND_POSITIONS;
} else if (indexOptions == IndexOptions.DOCS_AND_FREQS_AND_POSITIONS_AND_OFFSETS) {
bits |= Lucene46FieldInfosFormat.STORE_OFFSETS_IN_POSTINGS;
} else if (indexOptions == IndexOptions.DOCS_AND_FREQS) {
bits |= Lucene46FieldInfosFormat.OMIT_POSITIONS;
}
}
output.writeString(fi.name);
output.writeVInt(fi.number);
output.writeByte(bits);
// pack the DV types in one byte
final byte dv = docValuesByte(fi.getDocValuesType());
final byte nrm = docValuesByte(fi.hasNorms() ? DocValuesType.NUMERIC : DocValuesType.NONE);
assert (dv & (~0xF)) == 0 && (nrm & (~0x0F)) == 0;
byte val = (byte) (0xff & ((nrm << 4) | dv));
output.writeByte(val);
output.writeLong(fi.getDocValuesGen());
output.writeStringStringMap(fi.attributes());
}
CodecUtil.writeFooter(output);
}
}
public VariableGapTermsIndexWriter(SegmentWriteState state, IndexTermSelector policy)
throws IOException {
final String indexFileName =
IndexFileNames.segmentFileName(
state.segmentName, state.segmentSuffix, TERMS_INDEX_EXTENSION);
out = state.directory.createOutput(indexFileName, state.context);
boolean success = false;
try {
fieldInfos = state.fieldInfos;
this.policy = policy;
writeHeader(out);
success = true;
} finally {
if (!success) {
IOUtils.closeWhileHandlingException(out);
}
}
}
public SimpleTextTermVectorsReader(Directory directory, SegmentInfo si, IOContext context)
throws IOException {
boolean success = false;
try {
in =
directory.openInput(
IndexFileNames.segmentFileName(si.name, "", VECTORS_EXTENSION), context);
success = true;
} finally {
if (!success) {
try {
close();
} catch (Throwable t) {
} // ensure we throw our original exception
}
}
readIndex(si.maxDoc());
}
public FSTTermsReader(SegmentReadState state, PostingsReaderBase postingsReader)
throws IOException {
final String termsFileName =
IndexFileNames.segmentFileName(
state.segmentInfo.name, state.segmentSuffix, FSTTermsWriter.TERMS_EXTENSION);
this.postingsReader = postingsReader;
final IndexInput in = state.directory.openInput(termsFileName, state.context);
boolean success = false;
try {
version = readHeader(in);
if (version >= FSTTermsWriter.TERMS_VERSION_CHECKSUM) {
CodecUtil.checksumEntireFile(in);
}
this.postingsReader.init(in);
seekDir(in);
final FieldInfos fieldInfos = state.fieldInfos;
final int numFields = in.readVInt();
for (int i = 0; i < numFields; i++) {
int fieldNumber = in.readVInt();
FieldInfo fieldInfo = fieldInfos.fieldInfo(fieldNumber);
long numTerms = in.readVLong();
long sumTotalTermFreq =
fieldInfo.getIndexOptions() == IndexOptions.DOCS_ONLY ? -1 : in.readVLong();
long sumDocFreq = in.readVLong();
int docCount = in.readVInt();
int longsSize = in.readVInt();
TermsReader current =
new TermsReader(
fieldInfo, in, numTerms, sumTotalTermFreq, sumDocFreq, docCount, longsSize);
TermsReader previous = fields.put(fieldInfo.name, current);
checkFieldSummary(state.segmentInfo, in, current, previous);
}
success = true;
} finally {
if (success) {
IOUtils.close(in);
} else {
IOUtils.closeWhileHandlingException(in);
}
}
}
public SimpleTextStoredFieldsReader(
Directory directory, SegmentInfo si, FieldInfos fn, IOContext context) throws IOException {
this.fieldInfos = fn;
boolean success = false;
try {
in =
directory.openInput(
IndexFileNames.segmentFileName(
si.name, "", SimpleTextStoredFieldsWriter.FIELDS_EXTENSION),
context);
success = true;
} finally {
if (!success) {
try {
close();
} catch (Throwable t) {
} // ensure we throw our original exception
}
}
readIndex(si.maxDoc());
}
@Override
public void files(SegmentInfo segmentInfo, String segmentSuffix, Set<String> files)
throws IOException {
final String seedFileName =
IndexFileNames.segmentFileName(segmentInfo.name, segmentSuffix, SEED_EXT);
files.add(seedFileName);
SepPostingsReader.files(segmentInfo, segmentSuffix, files);
Lucene40PostingsReader.files(segmentInfo, segmentSuffix, files);
BlockTermsReader.files(segmentInfo, segmentSuffix, files);
BlockTreeTermsReader.files(segmentInfo, segmentSuffix, files);
FixedGapTermsIndexReader.files(segmentInfo, segmentSuffix, files);
VariableGapTermsIndexReader.files(segmentInfo, segmentSuffix, files);
// hackish!
Iterator<String> it = files.iterator();
while (it.hasNext()) {
final String file = it.next();
if (!segmentInfo.dir.fileExists(file)) {
it.remove();
}
}
// System.out.println("MockRandom.files return " + files);
}
@Override
public IndexOutput createTempOutput(String prefix, String suffix, IOContext context)
throws IOException {
ensureOpen();
// Make the file first...
RAMFile file = newRAMFile();
// ... then try to find a unique name for it:
while (true) {
String name =
IndexFileNames.segmentFileName(
prefix,
suffix
+ "_"
+ Long.toString(nextTempFileCounter.getAndIncrement(), Character.MAX_RADIX),
"tmp");
if (fileMap.putIfAbsent(name, file) == null) {
return new RAMOutputStream(name, file, true);
}
}
}
@Override
public void close() throws IOException {
delegateFieldsConsumer.close();
// Now we are done accumulating values for these fields
List<Entry<FieldInfo, FuzzySet>> nonSaturatedBlooms =
new ArrayList<Map.Entry<FieldInfo, FuzzySet>>();
for (Entry<FieldInfo, FuzzySet> entry : bloomFilters.entrySet()) {
FuzzySet bloomFilter = entry.getValue();
if (!bloomFilterFactory.isSaturated(bloomFilter, entry.getKey())) {
nonSaturatedBlooms.add(entry);
}
}
String bloomFileName =
IndexFileNames.segmentFileName(
state.segmentInfo.name, state.segmentSuffix, BLOOM_EXTENSION);
IndexOutput bloomOutput = null;
try {
bloomOutput = state.directory.createOutput(bloomFileName, state.context);
CodecUtil.writeHeader(bloomOutput, BLOOM_CODEC_NAME, BLOOM_CODEC_VERSION);
// remember the name of the postings format we will delegate to
bloomOutput.writeString(delegatePostingsFormat.getName());
// First field in the output file is the number of fields+blooms saved
bloomOutput.writeInt(nonSaturatedBlooms.size());
for (Entry<FieldInfo, FuzzySet> entry : nonSaturatedBlooms) {
FieldInfo fieldInfo = entry.getKey();
FuzzySet bloomFilter = entry.getValue();
bloomOutput.writeInt(fieldInfo.number);
saveAppropriatelySizedBloomFilter(bloomOutput, bloomFilter, fieldInfo);
}
} finally {
IOUtils.close(bloomOutput);
}
// We are done with large bitsets so no need to keep them hanging around
bloomFilters.clear();
}
@Override
public void finish() throws IOException {
if (finished) {
throw new IllegalStateException("already finished");
}
finished = true;
CodecUtil.writeFooter(dataOut);
String indexFileName =
IndexFileNames.segmentFileName(
writeState.segmentInfo.name,
writeState.segmentSuffix,
Lucene60PointsFormat.INDEX_EXTENSION);
// Write index file
try (IndexOutput indexOut =
writeState.directory.createOutput(indexFileName, writeState.context)) {
CodecUtil.writeIndexHeader(
indexOut,
Lucene60PointsFormat.META_CODEC_NAME,
Lucene60PointsFormat.INDEX_VERSION_CURRENT,
writeState.segmentInfo.getId(),
writeState.segmentSuffix);
int count = indexFPs.size();
indexOut.writeVInt(count);
for (Map.Entry<String, Long> ent : indexFPs.entrySet()) {
FieldInfo fieldInfo = writeState.fieldInfos.fieldInfo(ent.getKey());
if (fieldInfo == null) {
throw new IllegalStateException(
"wrote field=\"" + ent.getKey() + "\" but that field doesn't exist in FieldInfos");
}
indexOut.writeVInt(fieldInfo.number);
indexOut.writeVLong(ent.getValue());
}
CodecUtil.writeFooter(indexOut);
}
}
/**
* prunes the list of index files such that only the latest del generation files are contained.
*/
private void pruneOldDeleteGenerations(Set<Path> files) {
final TreeSet<Path> delFiles = new TreeSet<>();
for (Path file : files) {
if (file.getFileName().toString().endsWith(".liv")) {
delFiles.add(file);
}
}
Path last = null;
for (Path current : delFiles) {
if (last != null) {
final String newSegmentName =
IndexFileNames.parseSegmentName(current.getFileName().toString());
final String oldSegmentName =
IndexFileNames.parseSegmentName(last.getFileName().toString());
if (newSegmentName.equals(oldSegmentName)) {
int oldGen =
Integer.parseInt(
IndexFileNames.stripExtension(
IndexFileNames.stripSegmentName(last.getFileName().toString()))
.replace("_", ""),
Character.MAX_RADIX);
int newGen =
Integer.parseInt(
IndexFileNames.stripExtension(
IndexFileNames.stripSegmentName(current.getFileName().toString()))
.replace("_", ""),
Character.MAX_RADIX);
if (newGen > oldGen) {
files.remove(last);
} else {
files.remove(current);
continue;
}
}
}
last = current;
}
}
@Override
public Directory getCompoundReader(Directory dir, SegmentInfo si, IOContext context)
throws IOException {
String dataFile = IndexFileNames.segmentFileName(si.name, "", DATA_EXTENSION);
final IndexInput in = dir.openInput(dataFile, context);
BytesRefBuilder scratch = new BytesRefBuilder();
// first get to TOC:
DecimalFormat df =
new DecimalFormat(OFFSETPATTERN, DecimalFormatSymbols.getInstance(Locale.ROOT));
long pos = in.length() - TABLEPOS.length - OFFSETPATTERN.length() - 1;
in.seek(pos);
SimpleTextUtil.readLine(in, scratch);
assert StringHelper.startsWith(scratch.get(), TABLEPOS);
long tablePos = -1;
try {
tablePos = df.parse(stripPrefix(scratch, TABLEPOS)).longValue();
} catch (ParseException e) {
throw new CorruptIndexException(
"can't parse CFS trailer, got: " + scratch.get().utf8ToString(), in);
}
// seek to TOC and read it
in.seek(tablePos);
SimpleTextUtil.readLine(in, scratch);
assert StringHelper.startsWith(scratch.get(), TABLE);
int numEntries = Integer.parseInt(stripPrefix(scratch, TABLE));
final String fileNames[] = new String[numEntries];
final long startOffsets[] = new long[numEntries];
final long endOffsets[] = new long[numEntries];
for (int i = 0; i < numEntries; i++) {
SimpleTextUtil.readLine(in, scratch);
assert StringHelper.startsWith(scratch.get(), TABLENAME);
fileNames[i] = si.name + IndexFileNames.stripSegmentName(stripPrefix(scratch, TABLENAME));
if (i > 0) {
// files must be unique and in sorted order
assert fileNames[i].compareTo(fileNames[i - 1]) > 0;
}
SimpleTextUtil.readLine(in, scratch);
assert StringHelper.startsWith(scratch.get(), TABLESTART);
startOffsets[i] = Long.parseLong(stripPrefix(scratch, TABLESTART));
SimpleTextUtil.readLine(in, scratch);
assert StringHelper.startsWith(scratch.get(), TABLEEND);
endOffsets[i] = Long.parseLong(stripPrefix(scratch, TABLEEND));
}
return new Directory() {
private int getIndex(String name) throws IOException {
int index = Arrays.binarySearch(fileNames, name);
if (index < 0) {
throw new FileNotFoundException(
"No sub-file found (fileName="
+ name
+ " files: "
+ Arrays.toString(fileNames)
+ ")");
}
return index;
}
@Override
public String[] listAll() throws IOException {
ensureOpen();
return fileNames.clone();
}
@Override
public long fileLength(String name) throws IOException {
ensureOpen();
int index = getIndex(name);
return endOffsets[index] - startOffsets[index];
}
@Override
public IndexInput openInput(String name, IOContext context) throws IOException {
ensureOpen();
int index = getIndex(name);
return in.slice(name, startOffsets[index], endOffsets[index] - startOffsets[index]);
}
@Override
public void close() throws IOException {
in.close();
}
// write methods: disabled
@Override
public IndexOutput createOutput(String name, IOContext context) {
throw new UnsupportedOperationException();
}
@Override
public void sync(Collection<String> names) {
throw new UnsupportedOperationException();
}
@Override
public void deleteFile(String name) {
throw new UnsupportedOperationException();
}
@Override
public void renameFile(String source, String dest) {
throw new UnsupportedOperationException();
}
@Override
public Lock makeLock(String name) {
throw new UnsupportedOperationException();
}
};
}
Lucene42DocValuesProducer(
SegmentReadState state,
String dataCodec,
String dataExtension,
String metaCodec,
String metaExtension)
throws IOException {
maxDoc = state.segmentInfo.maxDoc();
merging = false;
String metaName =
IndexFileNames.segmentFileName(state.segmentInfo.name, state.segmentSuffix, metaExtension);
// read in the entries from the metadata file.
ChecksumIndexInput in = state.directory.openChecksumInput(metaName, state.context);
boolean success = false;
ramBytesUsed = new AtomicLong(RamUsageEstimator.shallowSizeOfInstance(getClass()));
try {
version = CodecUtil.checkHeader(in, metaCodec, VERSION_START, VERSION_CURRENT);
numerics = new HashMap<>();
binaries = new HashMap<>();
fsts = new HashMap<>();
numEntries = readFields(in, state.fieldInfos);
if (version >= VERSION_CHECKSUM) {
CodecUtil.checkFooter(in);
} else {
CodecUtil.checkEOF(in);
}
success = true;
} finally {
if (success) {
IOUtils.close(in);
} else {
IOUtils.closeWhileHandlingException(in);
}
}
String dataName =
IndexFileNames.segmentFileName(state.segmentInfo.name, state.segmentSuffix, dataExtension);
this.data = state.directory.openInput(dataName, state.context);
success = false;
try {
final int version2 = CodecUtil.checkHeader(data, dataCodec, VERSION_START, VERSION_CURRENT);
if (version != version2) {
throw new CorruptIndexException(
"Format versions mismatch: meta=" + version + ", data=" + version2, data);
}
if (version >= VERSION_CHECKSUM) {
// NOTE: data file is too costly to verify checksum against all the bytes on open,
// but for now we at least verify proper structure of the checksum footer: which looks
// for FOOTER_MAGIC + algorithmID. This is cheap and can detect some forms of corruption
// such as file truncation.
CodecUtil.retrieveChecksum(data);
}
success = true;
} finally {
if (!success) {
IOUtils.closeWhileHandlingException(this.data);
}
}
}
@Override
public FieldsProducer fieldsProducer(SegmentReadState state) throws IOException {
final String seedFileName =
IndexFileNames.segmentFileName(state.segmentInfo.name, state.segmentSuffix, SEED_EXT);
final IndexInput in = state.dir.openInput(seedFileName, state.context);
final long seed = in.readLong();
if (LuceneTestCase.VERBOSE) {
System.out.println(
"MockRandomCodec: reading from seg="
+ state.segmentInfo.name
+ " formatID="
+ state.segmentSuffix
+ " seed="
+ seed);
}
in.close();
final Random random = new Random(seed);
int readBufferSize = _TestUtil.nextInt(random, 1, 4096);
if (LuceneTestCase.VERBOSE) {
System.out.println("MockRandomCodec: readBufferSize=" + readBufferSize);
}
PostingsReaderBase postingsReader;
if (random.nextBoolean()) {
if (LuceneTestCase.VERBOSE) {
System.out.println("MockRandomCodec: reading Sep postings");
}
postingsReader =
new SepPostingsReader(
state.dir,
state.segmentInfo,
state.context,
new MockIntStreamFactory(random),
state.segmentSuffix);
} else {
if (LuceneTestCase.VERBOSE) {
System.out.println("MockRandomCodec: reading Standard postings");
}
postingsReader =
new Lucene40PostingsReader(
state.dir, state.segmentInfo, state.context, state.segmentSuffix);
}
if (random.nextBoolean()) {
final int totTFCutoff = _TestUtil.nextInt(random, 1, 20);
if (LuceneTestCase.VERBOSE) {
System.out.println(
"MockRandomCodec: reading pulsing postings with totTFCutoff=" + totTFCutoff);
}
postingsReader = new PulsingPostingsReader(postingsReader);
}
final FieldsProducer fields;
if (random.nextBoolean()) {
// Use BlockTree terms dict
if (LuceneTestCase.VERBOSE) {
System.out.println("MockRandomCodec: reading BlockTree terms dict");
}
boolean success = false;
try {
fields =
new BlockTreeTermsReader(
state.dir,
state.fieldInfos,
state.segmentInfo.name,
postingsReader,
state.context,
state.segmentSuffix,
state.termsIndexDivisor);
success = true;
} finally {
if (!success) {
postingsReader.close();
}
}
} else {
if (LuceneTestCase.VERBOSE) {
System.out.println("MockRandomCodec: reading Block terms dict");
}
final TermsIndexReaderBase indexReader;
boolean success = false;
try {
final boolean doFixedGap = random.nextBoolean();
// randomness diverges from writer, here:
if (state.termsIndexDivisor != -1) {
state.termsIndexDivisor = _TestUtil.nextInt(random, 1, 10);
}
if (doFixedGap) {
// if termsIndexDivisor is set to -1, we should not touch it. It means a
// test explicitly instructed not to load the terms index.
if (LuceneTestCase.VERBOSE) {
System.out.println(
"MockRandomCodec: fixed-gap terms index (divisor=" + state.termsIndexDivisor + ")");
}
indexReader =
new FixedGapTermsIndexReader(
state.dir,
state.fieldInfos,
state.segmentInfo.name,
state.termsIndexDivisor,
BytesRef.getUTF8SortedAsUnicodeComparator(),
state.segmentSuffix,
state.context);
} else {
final int n2 = random.nextInt(3);
if (n2 == 1) {
random.nextInt();
} else if (n2 == 2) {
random.nextLong();
}
if (LuceneTestCase.VERBOSE) {
System.out.println(
"MockRandomCodec: variable-gap terms index (divisor="
+ state.termsIndexDivisor
+ ")");
}
indexReader =
new VariableGapTermsIndexReader(
state.dir,
state.fieldInfos,
state.segmentInfo.name,
state.termsIndexDivisor,
state.segmentSuffix,
state.context);
}
success = true;
} finally {
if (!success) {
postingsReader.close();
}
}
final int termsCacheSize = _TestUtil.nextInt(random, 1, 1024);
success = false;
try {
fields =
new BlockTermsReader(
indexReader,
state.dir,
state.fieldInfos,
state.segmentInfo.name,
postingsReader,
state.context,
termsCacheSize,
state.segmentSuffix);
success = true;
} finally {
if (!success) {
try {
postingsReader.close();
} finally {
indexReader.close();
}
}
}
}
return fields;
}
/**
* Create a new writer. The number of items (terms or sub-blocks) per block will aim to be between
* minItemsPerBlock and maxItemsPerBlock, though in some cases the blocks may be smaller than the
* min.
*/
public BlockTreeTermsWriter(
SegmentWriteState state,
PostingsWriterBase postingsWriter,
int minItemsInBlock,
int maxItemsInBlock)
throws IOException {
if (minItemsInBlock <= 1) {
throw new IllegalArgumentException("minItemsInBlock must be >= 2; got " + minItemsInBlock);
}
if (maxItemsInBlock <= 0) {
throw new IllegalArgumentException("maxItemsInBlock must be >= 1; got " + maxItemsInBlock);
}
if (minItemsInBlock > maxItemsInBlock) {
throw new IllegalArgumentException(
"maxItemsInBlock must be >= minItemsInBlock; got maxItemsInBlock="
+ maxItemsInBlock
+ " minItemsInBlock="
+ minItemsInBlock);
}
if (2 * (minItemsInBlock - 1) > maxItemsInBlock) {
throw new IllegalArgumentException(
"maxItemsInBlock must be at least 2*(minItemsInBlock-1); got maxItemsInBlock="
+ maxItemsInBlock
+ " minItemsInBlock="
+ minItemsInBlock);
}
final String termsFileName =
IndexFileNames.segmentFileName(
state.segmentInfo.name, state.segmentSuffix, TERMS_EXTENSION);
out = state.directory.createOutput(termsFileName, state.context);
boolean success = false;
IndexOutput indexOut = null;
try {
fieldInfos = state.fieldInfos;
this.minItemsInBlock = minItemsInBlock;
this.maxItemsInBlock = maxItemsInBlock;
writeHeader(out);
// DEBUG = state.segmentName.equals("_4a");
final String termsIndexFileName =
IndexFileNames.segmentFileName(
state.segmentInfo.name, state.segmentSuffix, TERMS_INDEX_EXTENSION);
indexOut = state.directory.createOutput(termsIndexFileName, state.context);
writeIndexHeader(indexOut);
currentField = null;
this.postingsWriter = postingsWriter;
// segment = state.segmentName;
// System.out.println("BTW.init seg=" + state.segmentName);
postingsWriter.start(out); // have consumer write its format/header
success = true;
} finally {
if (!success) {
IOUtils.closeWhileHandlingException(out, indexOut);
}
}
this.indexOut = indexOut;
}
@Override
public FieldsConsumer fieldsConsumer(SegmentWriteState state) throws IOException {
// we pull this before the seed intentionally: because its not consumed at runtime
// (the skipInterval is written into postings header)
int skipInterval = _TestUtil.nextInt(seedRandom, 2, 10);
if (LuceneTestCase.VERBOSE) {
System.out.println("MockRandomCodec: skipInterval=" + skipInterval);
}
final long seed = seedRandom.nextLong();
if (LuceneTestCase.VERBOSE) {
System.out.println(
"MockRandomCodec: writing to seg="
+ state.segmentName
+ " formatID="
+ state.segmentSuffix
+ " seed="
+ seed);
}
final String seedFileName =
IndexFileNames.segmentFileName(state.segmentName, state.segmentSuffix, SEED_EXT);
final IndexOutput out = state.directory.createOutput(seedFileName, state.context);
try {
out.writeLong(seed);
} finally {
out.close();
}
final Random random = new Random(seed);
random.nextInt(); // consume a random for buffersize
PostingsWriterBase postingsWriter;
if (random.nextBoolean()) {
postingsWriter = new SepPostingsWriter(state, new MockIntStreamFactory(random), skipInterval);
} else {
if (LuceneTestCase.VERBOSE) {
System.out.println("MockRandomCodec: writing Standard postings");
}
postingsWriter = new Lucene40PostingsWriter(state, skipInterval);
}
if (random.nextBoolean()) {
final int totTFCutoff = _TestUtil.nextInt(random, 1, 20);
if (LuceneTestCase.VERBOSE) {
System.out.println(
"MockRandomCodec: writing pulsing postings with totTFCutoff=" + totTFCutoff);
}
postingsWriter = new PulsingPostingsWriter(totTFCutoff, postingsWriter);
}
final FieldsConsumer fields;
if (random.nextBoolean()) {
// Use BlockTree terms dict
if (LuceneTestCase.VERBOSE) {
System.out.println("MockRandomCodec: writing BlockTree terms dict");
}
// TODO: would be nice to allow 1 but this is very
// slow to write
final int minTermsInBlock = _TestUtil.nextInt(random, 2, 100);
final int maxTermsInBlock = Math.max(2, (minTermsInBlock - 1) * 2 + random.nextInt(100));
boolean success = false;
try {
fields = new BlockTreeTermsWriter(state, postingsWriter, minTermsInBlock, maxTermsInBlock);
success = true;
} finally {
if (!success) {
postingsWriter.close();
}
}
} else {
if (LuceneTestCase.VERBOSE) {
System.out.println("MockRandomCodec: writing Block terms dict");
}
boolean success = false;
final TermsIndexWriterBase indexWriter;
try {
if (random.nextBoolean()) {
state.termIndexInterval = _TestUtil.nextInt(random, 1, 100);
if (LuceneTestCase.VERBOSE) {
System.out.println(
"MockRandomCodec: fixed-gap terms index (tii=" + state.termIndexInterval + ")");
}
indexWriter = new FixedGapTermsIndexWriter(state);
} else {
final VariableGapTermsIndexWriter.IndexTermSelector selector;
final int n2 = random.nextInt(3);
if (n2 == 0) {
final int tii = _TestUtil.nextInt(random, 1, 100);
selector = new VariableGapTermsIndexWriter.EveryNTermSelector(tii);
if (LuceneTestCase.VERBOSE) {
System.out.println("MockRandomCodec: variable-gap terms index (tii=" + tii + ")");
}
} else if (n2 == 1) {
final int docFreqThresh = _TestUtil.nextInt(random, 2, 100);
final int tii = _TestUtil.nextInt(random, 1, 100);
selector =
new VariableGapTermsIndexWriter.EveryNOrDocFreqTermSelector(docFreqThresh, tii);
} else {
final long seed2 = random.nextLong();
final int gap = _TestUtil.nextInt(random, 2, 40);
if (LuceneTestCase.VERBOSE) {
System.out.println("MockRandomCodec: random-gap terms index (max gap=" + gap + ")");
}
selector =
new VariableGapTermsIndexWriter.IndexTermSelector() {
final Random rand = new Random(seed2);
@Override
public boolean isIndexTerm(BytesRef term, TermStats stats) {
return rand.nextInt(gap) == gap / 2;
}
@Override
public void newField(FieldInfo fieldInfo) {}
};
}
indexWriter = new VariableGapTermsIndexWriter(state, selector);
}
success = true;
} finally {
if (!success) {
postingsWriter.close();
}
}
success = false;
try {
fields = new BlockTermsWriter(indexWriter, state, postingsWriter);
success = true;
} finally {
if (!success) {
try {
postingsWriter.close();
} finally {
indexWriter.close();
}
}
}
}
return fields;
}
@Override
public FieldInfos read(
Directory directory, SegmentInfo segmentInfo, String segmentSuffix, IOContext context)
throws IOException {
final String fileName =
IndexFileNames.segmentFileName(
segmentInfo.name, segmentSuffix, Lucene46FieldInfosFormat.EXTENSION);
try (ChecksumIndexInput input = directory.openChecksumInput(fileName, context)) {
int codecVersion =
CodecUtil.checkHeader(
input,
Lucene46FieldInfosFormat.CODEC_NAME,
Lucene46FieldInfosFormat.FORMAT_START,
Lucene46FieldInfosFormat.FORMAT_CURRENT);
final int size = input.readVInt(); // read in the size
FieldInfo infos[] = new FieldInfo[size];
for (int i = 0; i < size; i++) {
String name = input.readString();
final int fieldNumber = input.readVInt();
if (fieldNumber < 0) {
throw new CorruptIndexException(
"invalid field number for field: " + name + ", fieldNumber=" + fieldNumber, input);
}
byte bits = input.readByte();
boolean isIndexed = (bits & Lucene46FieldInfosFormat.IS_INDEXED) != 0;
boolean storeTermVector = (bits & Lucene46FieldInfosFormat.STORE_TERMVECTOR) != 0;
boolean omitNorms = (bits & Lucene46FieldInfosFormat.OMIT_NORMS) != 0;
boolean storePayloads = (bits & Lucene46FieldInfosFormat.STORE_PAYLOADS) != 0;
final IndexOptions indexOptions;
if (!isIndexed) {
indexOptions = IndexOptions.NONE;
} else if ((bits & Lucene46FieldInfosFormat.OMIT_TERM_FREQ_AND_POSITIONS) != 0) {
indexOptions = IndexOptions.DOCS;
} else if ((bits & Lucene46FieldInfosFormat.OMIT_POSITIONS) != 0) {
indexOptions = IndexOptions.DOCS_AND_FREQS;
} else if ((bits & Lucene46FieldInfosFormat.STORE_OFFSETS_IN_POSTINGS) != 0) {
indexOptions = IndexOptions.DOCS_AND_FREQS_AND_POSITIONS_AND_OFFSETS;
} else {
indexOptions = IndexOptions.DOCS_AND_FREQS_AND_POSITIONS;
}
// DV Types are packed in one byte
byte val = input.readByte();
final DocValuesType docValuesType = getDocValuesType(input, (byte) (val & 0x0F));
final DocValuesType normsType = getDocValuesType(input, (byte) ((val >>> 4) & 0x0F));
final long dvGen = input.readLong();
final Map<String, String> attributes = input.readStringStringMap();
if (isIndexed && omitNorms == false && normsType == DocValuesType.NONE) {
// Undead norms! Lucene42NormsProducer will check this and bring norms back from the
// dead:
UndeadNormsProducer.setUndead(attributes);
}
infos[i] =
new FieldInfo(
name,
fieldNumber,
storeTermVector,
omitNorms,
storePayloads,
indexOptions,
docValuesType,
dvGen,
Collections.unmodifiableMap(attributes));
infos[i].checkConsistency();
}
if (codecVersion >= Lucene46FieldInfosFormat.FORMAT_CHECKSUM) {
CodecUtil.checkFooter(input);
} else {
CodecUtil.checkEOF(input);
}
return new FieldInfos(infos);
}
}
/** Sole constructor. */
public BlockTreeTermsReader(PostingsReaderBase postingsReader, SegmentReadState state)
throws IOException {
boolean success = false;
IndexInput indexIn = null;
this.postingsReader = postingsReader;
this.segment = state.segmentInfo.name;
String termsName =
IndexFileNames.segmentFileName(segment, state.segmentSuffix, TERMS_EXTENSION);
try {
termsIn = state.directory.openInput(termsName, state.context);
version =
CodecUtil.checkIndexHeader(
termsIn,
TERMS_CODEC_NAME,
VERSION_START,
VERSION_CURRENT,
state.segmentInfo.getId(),
state.segmentSuffix);
String indexName =
IndexFileNames.segmentFileName(segment, state.segmentSuffix, TERMS_INDEX_EXTENSION);
indexIn = state.directory.openInput(indexName, state.context);
CodecUtil.checkIndexHeader(
indexIn,
TERMS_INDEX_CODEC_NAME,
version,
version,
state.segmentInfo.getId(),
state.segmentSuffix);
CodecUtil.checksumEntireFile(indexIn);
// Have PostingsReader init itself
postingsReader.init(termsIn, state);
// NOTE: data file is too costly to verify checksum against all the bytes on open,
// but for now we at least verify proper structure of the checksum footer: which looks
// for FOOTER_MAGIC + algorithmID. This is cheap and can detect some forms of corruption
// such as file truncation.
CodecUtil.retrieveChecksum(termsIn);
// Read per-field details
seekDir(termsIn, dirOffset);
seekDir(indexIn, indexDirOffset);
final int numFields = termsIn.readVInt();
if (numFields < 0) {
throw new CorruptIndexException("invalid numFields: " + numFields, termsIn);
}
for (int i = 0; i < numFields; ++i) {
final int field = termsIn.readVInt();
final long numTerms = termsIn.readVLong();
if (numTerms <= 0) {
throw new CorruptIndexException("Illegal numTerms for field number: " + field, termsIn);
}
final int numBytes = termsIn.readVInt();
if (numBytes < 0) {
throw new CorruptIndexException(
"invalid rootCode for field number: " + field + ", numBytes=" + numBytes, termsIn);
}
final BytesRef rootCode = new BytesRef(new byte[numBytes]);
termsIn.readBytes(rootCode.bytes, 0, numBytes);
rootCode.length = numBytes;
final FieldInfo fieldInfo = state.fieldInfos.fieldInfo(field);
if (fieldInfo == null) {
throw new CorruptIndexException("invalid field number: " + field, termsIn);
}
final long sumTotalTermFreq =
fieldInfo.getIndexOptions() == IndexOptions.DOCS ? -1 : termsIn.readVLong();
final long sumDocFreq = termsIn.readVLong();
final int docCount = termsIn.readVInt();
final int longsSize = termsIn.readVInt();
if (longsSize < 0) {
throw new CorruptIndexException(
"invalid longsSize for field: " + fieldInfo.name + ", longsSize=" + longsSize,
termsIn);
}
BytesRef minTerm = readBytesRef(termsIn);
BytesRef maxTerm = readBytesRef(termsIn);
if (docCount < 0
|| docCount > state.segmentInfo.getDocCount()) { // #docs with field must be <= #docs
throw new CorruptIndexException(
"invalid docCount: " + docCount + " maxDoc: " + state.segmentInfo.getDocCount(),
termsIn);
}
if (sumDocFreq < docCount) { // #postings must be >= #docs with field
throw new CorruptIndexException(
"invalid sumDocFreq: " + sumDocFreq + " docCount: " + docCount, termsIn);
}
if (sumTotalTermFreq != -1
&& sumTotalTermFreq < sumDocFreq) { // #positions must be >= #postings
throw new CorruptIndexException(
"invalid sumTotalTermFreq: " + sumTotalTermFreq + " sumDocFreq: " + sumDocFreq,
termsIn);
}
final long indexStartFP = indexIn.readVLong();
FieldReader previous =
fields.put(
fieldInfo.name,
new FieldReader(
this,
fieldInfo,
numTerms,
rootCode,
sumTotalTermFreq,
sumDocFreq,
docCount,
indexStartFP,
longsSize,
indexIn,
minTerm,
maxTerm));
if (previous != null) {
throw new CorruptIndexException("duplicate field: " + fieldInfo.name, termsIn);
}
}
indexIn.close();
success = true;
} finally {
if (!success) {
// this.close() will close in:
IOUtils.closeWhileHandlingException(indexIn, this);
}
}
}
public BlockTermsReader(
TermsIndexReaderBase indexReader, PostingsReaderBase postingsReader, SegmentReadState state)
throws IOException {
this.postingsReader = postingsReader;
String filename =
IndexFileNames.segmentFileName(
state.segmentInfo.name, state.segmentSuffix, BlockTermsWriter.TERMS_EXTENSION);
in = state.directory.openInput(filename, state.context);
boolean success = false;
try {
CodecUtil.checkIndexHeader(
in,
BlockTermsWriter.CODEC_NAME,
BlockTermsWriter.VERSION_START,
BlockTermsWriter.VERSION_CURRENT,
state.segmentInfo.getId(),
state.segmentSuffix);
// Have PostingsReader init itself
postingsReader.init(in, state);
// NOTE: data file is too costly to verify checksum against all the bytes on open,
// but for now we at least verify proper structure of the checksum footer: which looks
// for FOOTER_MAGIC + algorithmID. This is cheap and can detect some forms of corruption
// such as file truncation.
CodecUtil.retrieveChecksum(in);
// Read per-field details
seekDir(in);
final int numFields = in.readVInt();
if (numFields < 0) {
throw new CorruptIndexException("invalid number of fields: " + numFields, in);
}
for (int i = 0; i < numFields; i++) {
final int field = in.readVInt();
final long numTerms = in.readVLong();
assert numTerms >= 0;
final long termsStartPointer = in.readVLong();
final FieldInfo fieldInfo = state.fieldInfos.fieldInfo(field);
final long sumTotalTermFreq =
fieldInfo.getIndexOptions() == IndexOptions.DOCS ? -1 : in.readVLong();
final long sumDocFreq = in.readVLong();
final int docCount = in.readVInt();
final int longsSize = in.readVInt();
if (docCount < 0
|| docCount > state.segmentInfo.maxDoc()) { // #docs with field must be <= #docs
throw new CorruptIndexException(
"invalid docCount: " + docCount + " maxDoc: " + state.segmentInfo.maxDoc(), in);
}
if (sumDocFreq < docCount) { // #postings must be >= #docs with field
throw new CorruptIndexException(
"invalid sumDocFreq: " + sumDocFreq + " docCount: " + docCount, in);
}
if (sumTotalTermFreq != -1
&& sumTotalTermFreq < sumDocFreq) { // #positions must be >= #postings
throw new CorruptIndexException(
"invalid sumTotalTermFreq: " + sumTotalTermFreq + " sumDocFreq: " + sumDocFreq, in);
}
FieldReader previous =
fields.put(
fieldInfo.name,
new FieldReader(
fieldInfo,
numTerms,
termsStartPointer,
sumTotalTermFreq,
sumDocFreq,
docCount,
longsSize));
if (previous != null) {
throw new CorruptIndexException("duplicate fields: " + fieldInfo.name, in);
}
}
success = true;
} finally {
if (!success) {
in.close();
}
}
this.indexReader = indexReader;
}
Lucene70NormsProducer(
SegmentReadState state,
String dataCodec,
String dataExtension,
String metaCodec,
String metaExtension)
throws IOException {
maxDoc = state.segmentInfo.maxDoc();
String metaName =
IndexFileNames.segmentFileName(state.segmentInfo.name, state.segmentSuffix, metaExtension);
int version = -1;
// read in the entries from the metadata file.
try (ChecksumIndexInput in = state.directory.openChecksumInput(metaName, state.context)) {
Throwable priorE = null;
try {
version =
CodecUtil.checkIndexHeader(
in,
metaCodec,
VERSION_START,
VERSION_CURRENT,
state.segmentInfo.getId(),
state.segmentSuffix);
readFields(in, state.fieldInfos);
} catch (Throwable exception) {
priorE = exception;
} finally {
CodecUtil.checkFooter(in, priorE);
}
}
String dataName =
IndexFileNames.segmentFileName(state.segmentInfo.name, state.segmentSuffix, dataExtension);
data = state.directory.openInput(dataName, state.context);
boolean success = false;
try {
final int version2 =
CodecUtil.checkIndexHeader(
data,
dataCodec,
VERSION_START,
VERSION_CURRENT,
state.segmentInfo.getId(),
state.segmentSuffix);
if (version != version2) {
throw new CorruptIndexException(
"Format versions mismatch: meta=" + version + ",data=" + version2, data);
}
// NOTE: data file is too costly to verify checksum against all the bytes on open,
// but for now we at least verify proper structure of the checksum footer: which looks
// for FOOTER_MAGIC + algorithmID. This is cheap and can detect some forms of corruption
// such as file truncation.
CodecUtil.retrieveChecksum(data);
success = true;
} finally {
if (!success) {
IOUtils.closeWhileHandlingException(this.data);
}
}
}
