Merge branch 'master' into columnfamilies

Conflicts:
	db/version_set.cc
	tools/db_stress.cc

LICENSE

@@ -2,7 +2,7 @@ BSD License
 
 For rocksdb software
 
-Copyright (c) 2013, Facebook, Inc.
+Copyright (c) 2014, Facebook, Inc.
 All rights reserved.
 ---------------------------------------------------------------------
 

Makefile

@@ -36,7 +36,7 @@ else
 	PLATFORM_CCFLAGS += $(JEMALLOC_INCLUDE) -DHAVE_JEMALLOC
 endif
 
-WARNING_FLAGS = -Wall -Werror
+WARNING_FLAGS = -Wall -Werror -Wno-sign-compare
 CFLAGS += -g $(WARNING_FLAGS) -I. -I./include $(PLATFORM_CCFLAGS) $(OPT)
 CXXFLAGS += -g $(WARNING_FLAGS) -I. -I./include $(PLATFORM_CXXFLAGS) $(OPT) -Woverloaded-virtual
 

db/db_iter.cc

@@ -379,10 +379,10 @@ void DBIter::FindPrevUserEntry() {
   uint64_t num_skipped = 0;
 
   ValueType value_type = kTypeDeletion;
+  bool saved_key_valid = true;
   if (iter_->Valid()) {
     do {
       ParsedInternalKey ikey;
-      bool saved_key_cleared = false;
       if (ParseKey(&ikey) && ikey.sequence <= sequence_) {
         if ((value_type != kTypeDeletion) &&
             user_comparator_->Compare(ikey.user_key, saved_key_) < 0) {
@@ -393,7 +393,7 @@ void DBIter::FindPrevUserEntry() {
         if (value_type == kTypeDeletion) {
           saved_key_.clear();
           ClearSavedValue();
-          saved_key_cleared = true;
+          saved_key_valid = false;
         } else {
           Slice raw_value = iter_->value();
           if (saved_value_.capacity() > raw_value.size() + 1048576) {
@@ -403,13 +403,17 @@ void DBIter::FindPrevUserEntry() {
           SaveKey(ExtractUserKey(iter_->key()), &saved_key_);
           saved_value_.assign(raw_value.data(), raw_value.size());
         }
+      } else {
+        // In the case of ikey.sequence > sequence_, we might have already
+        // iterated to a different user key.
+        saved_key_valid = false;
       }
       num_skipped++;
       // If we have sequentially iterated via numerous keys and still not
       // found the prev user-key, then it is better to seek so that we can
       // avoid too many key comparisons. We seek to the first occurence of
       // our current key by looking for max sequence number.
-      if (!saved_key_cleared && num_skipped > max_skip_) {
+      if (saved_key_valid && num_skipped > max_skip_) {
         num_skipped = 0;
         std::string last_key;
         AppendInternalKey(&last_key,

db/db_test.cc

@@ -1504,6 +1504,75 @@ TEST(DBTest, IterSeekBeforePrev) {
   delete iter;
 }
 
+TEST(DBTest, IterNextWithNewerSeq) {
+  ASSERT_OK(Put("0", "0"));
+  dbfull()->Flush(FlushOptions());
+  ASSERT_OK(Put("a", "b"));
+  ASSERT_OK(Put("c", "d"));
+  ASSERT_OK(Put("d", "e"));
+  auto iter = db_->NewIterator(ReadOptions());
+
+  // Create a key that needs to be skipped for Seq too new
+  for (uint64_t i = 0; i < last_options_.max_sequential_skip_in_iterations + 1;
+       i++) {
+    ASSERT_OK(Put("b", "f"));
+  }
+
+  iter->Seek(Slice("a"));
+  ASSERT_EQ(IterStatus(iter), "a->b");
+  iter->Next();
+  ASSERT_EQ(IterStatus(iter), "c->d");
+  delete iter;
+}
+
+TEST(DBTest, IterPrevWithNewerSeq) {
+  ASSERT_OK(Put("0", "0"));
+  dbfull()->Flush(FlushOptions());
+  ASSERT_OK(Put("a", "b"));
+  ASSERT_OK(Put("c", "d"));
+  ASSERT_OK(Put("d", "e"));
+  auto iter = db_->NewIterator(ReadOptions());
+
+  // Create a key that needs to be skipped for Seq too new
+  for (uint64_t i = 0; i < last_options_.max_sequential_skip_in_iterations + 1;
+       i++) {
+    ASSERT_OK(Put("b", "f"));
+  }
+
+  iter->Seek(Slice("d"));
+  ASSERT_EQ(IterStatus(iter), "d->e");
+  iter->Prev();
+  ASSERT_EQ(IterStatus(iter), "c->d");
+  iter->Prev();
+  ASSERT_EQ(IterStatus(iter), "a->b");
+
+  iter->Prev();
+  delete iter;
+}
+
+TEST(DBTest, IterPrevWithNewerSeq2) {
+  ASSERT_OK(Put("0", "0"));
+  dbfull()->Flush(FlushOptions());
+  ASSERT_OK(Put("a", "b"));
+  ASSERT_OK(Put("c", "d"));
+  ASSERT_OK(Put("d", "e"));
+  auto iter = db_->NewIterator(ReadOptions());
+  iter->Seek(Slice("c"));
+  ASSERT_EQ(IterStatus(iter), "c->d");
+
+  // Create a key that needs to be skipped for Seq too new
+  for (uint64_t i = 0; i < last_options_.max_sequential_skip_in_iterations + 1;
+      i++) {
+    ASSERT_OK(Put("b", "f"));
+  }
+
+  iter->Prev();
+  ASSERT_EQ(IterStatus(iter), "a->b");
+
+  iter->Prev();
+  delete iter;
+}
+
 TEST(DBTest, IterEmpty) {
   do {
     CreateAndReopenWithCF({"pikachu"});

db/log_test.cc

@@ -451,7 +451,7 @@ TEST(LogTest, TruncatedTrailingRecordIsIgnored) {
   ShrinkSize(4);   // Drop all payload as well as a header byte
   ASSERT_EQ("EOF", Read());
   // Truncated last record is ignored, not treated as an error
-  ASSERT_EQ(0, DroppedBytes());
+  ASSERT_EQ(0U, DroppedBytes());
   ASSERT_EQ("", ReportMessage());
 }
 
@@ -470,7 +470,7 @@ TEST(LogTest, BadLengthAtEndIsIgnored) {
   Write("foo");
   ShrinkSize(1);
   ASSERT_EQ("EOF", Read());
-  ASSERT_EQ(0, DroppedBytes());
+  ASSERT_EQ(0U, DroppedBytes());
   ASSERT_EQ("", ReportMessage());
 }
 
@@ -528,7 +528,7 @@ TEST(LogTest, MissingLastIsIgnored) {
   ShrinkSize(14);
   ASSERT_EQ("EOF", Read());
   ASSERT_EQ("", ReportMessage());
-  ASSERT_EQ(0, DroppedBytes());
+  ASSERT_EQ(0U, DroppedBytes());
 }
 
 TEST(LogTest, PartialLastIsIgnored) {
@@ -537,7 +537,7 @@ TEST(LogTest, PartialLastIsIgnored) {
   ShrinkSize(1);
   ASSERT_EQ("EOF", Read());
   ASSERT_EQ("", ReportMessage());
-  ASSERT_EQ(0, DroppedBytes());
+  ASSERT_EQ(0U, DroppedBytes());
 }
 
 TEST(LogTest, ErrorJoinsRecords) {

tools/db_stress.cc

@@ -972,10 +972,10 @@ class StressTest {
       prefixes[i].resize(FLAGS_prefix_size);
       prefix_slices[i] = Slice(prefixes[i]);
       readoptionscopy[i] = readoptions;
-      readoptionscopy[i].prefix = &prefix_slices[i];
+      readoptionscopy[i].prefix_seek = true;
       readoptionscopy[i].snapshot = snapshot;
       iters[i] = db_->NewIterator(readoptionscopy[i], column_family);
-      iters[i]->SeekToFirst();
+      iters[i]->Seek(prefix_slices[i]);
     }
 
     int count = 0;
@@ -1157,11 +1157,11 @@ class StressTest {
         // prefix
         if (!FLAGS_test_batches_snapshots) {
           Slice prefix = Slice(key.data(), FLAGS_prefix_size);
-          read_opts.prefix = &prefix;
+          read_opts.prefix_seek = true;
           Iterator* iter = db_->NewIterator(read_opts, column_family);
           int64_t count = 0;
-          for (iter->SeekToFirst(); iter->Valid(); iter->Next()) {
+          for (iter->Seek(prefix);
-            assert(iter->key().starts_with(prefix));
+               iter->Valid() && iter->key().starts_with(prefix); iter->Next()) {
             ++count;
           }
           assert(count <=
@@ -1175,7 +1175,6 @@ class StressTest {
         } else {
           MultiPrefixScan(thread, read_opts, column_family, key);
         }
-        read_opts.prefix = nullptr;
       } else if (prefixBound <= prob_op && prob_op < writeBound) {
         // OPERATION write
         uint32_t value_base = thread->rand.Next();

util/env_test.cc

@@ -172,8 +172,8 @@ TEST(EnvPosixTest, TwoPools) {
   env_->SetBackgroundThreads(kLowPoolSize);
   env_->SetBackgroundThreads(kHighPoolSize, Env::Priority::HIGH);
 
-  ASSERT_EQ(0, env_->GetThreadPoolQueueLen(Env::Priority::LOW));
+  ASSERT_EQ(0U, env_->GetThreadPoolQueueLen(Env::Priority::LOW));
-  ASSERT_EQ(0, env_->GetThreadPoolQueueLen(Env::Priority::HIGH));
+  ASSERT_EQ(0U, env_->GetThreadPoolQueueLen(Env::Priority::HIGH));
 
   // schedule same number of jobs in each pool
   for (int i = 0; i < kJobs; i++) {
@@ -182,10 +182,11 @@ TEST(EnvPosixTest, TwoPools) {
   }
   // Wait a short while for the jobs to be dispatched.
   Env::Default()->SleepForMicroseconds(kDelayMicros);
-  ASSERT_EQ(kJobs - kLowPoolSize, env_->GetThreadPoolQueueLen());
+  ASSERT_EQ((unsigned int)(kJobs - kLowPoolSize),
-  ASSERT_EQ(kJobs - kLowPoolSize,
+            env_->GetThreadPoolQueueLen());
+  ASSERT_EQ((unsigned int)(kJobs - kLowPoolSize),
             env_->GetThreadPoolQueueLen(Env::Priority::LOW));
-  ASSERT_EQ(kJobs - kHighPoolSize,
+  ASSERT_EQ((unsigned int)(kJobs - kHighPoolSize),
             env_->GetThreadPoolQueueLen(Env::Priority::HIGH));
 
   // wait for all jobs to finish
@@ -194,8 +195,8 @@ TEST(EnvPosixTest, TwoPools) {
     env_->SleepForMicroseconds(kDelayMicros);
   }
 
-  ASSERT_EQ(0, env_->GetThreadPoolQueueLen(Env::Priority::LOW));
+  ASSERT_EQ(0U, env_->GetThreadPoolQueueLen(Env::Priority::LOW));
-  ASSERT_EQ(0, env_->GetThreadPoolQueueLen(Env::Priority::HIGH));
+  ASSERT_EQ(0U, env_->GetThreadPoolQueueLen(Env::Priority::HIGH));
 }
 
 bool IsSingleVarint(const std::string& s) {
@@ -296,18 +297,18 @@ TEST(EnvPosixTest, AllocateTest) {
 
   struct stat f_stat;
   stat(fname.c_str(), &f_stat);
-  ASSERT_EQ(data.size(), f_stat.st_size);
+  ASSERT_EQ((unsigned int)data.size(), f_stat.st_size);
   // verify that blocks are preallocated
-  ASSERT_EQ(kPreallocateSize / kBlockSize, f_stat.st_blocks);
+  ASSERT_EQ((unsigned int)(kPreallocateSize / kBlockSize), f_stat.st_blocks);
 
   // close the file, should deallocate the blocks
   wfile.reset();
 
   stat(fname.c_str(), &f_stat);
-  ASSERT_EQ(data.size(), f_stat.st_size);
+  ASSERT_EQ((unsigned int)data.size(), f_stat.st_size);
   // verify that preallocated blocks were deallocated on file close
   size_t data_blocks_pages = ((data.size() + kPageSize - 1) / kPageSize);
-  ASSERT_EQ(data_blocks_pages * kPageSize / kBlockSize, f_stat.st_blocks);
+  ASSERT_EQ((unsigned int)(data_blocks_pages * kPageSize / kBlockSize), f_stat.st_blocks);
 }
 #endif