aboutsummaryrefslogtreecommitdiff
path: root/absl/container/internal/raw_hash_set_test.cc
diff options
context:
space:
mode:
Diffstat (limited to 'absl/container/internal/raw_hash_set_test.cc')
-rw-r--r--absl/container/internal/raw_hash_set_test.cc34
1 files changed, 29 insertions, 5 deletions
diff --git a/absl/container/internal/raw_hash_set_test.cc b/absl/container/internal/raw_hash_set_test.cc
index 0fba46f..81c4b47 100644
--- a/absl/container/internal/raw_hash_set_test.cc
+++ b/absl/container/internal/raw_hash_set_test.cc
@@ -14,6 +14,7 @@
#include "absl/container/internal/raw_hash_set.h"
+#include <atomic>
#include <cmath>
#include <cstdint>
#include <deque>
@@ -22,6 +23,8 @@
#include <numeric>
#include <random>
#include <string>
+#include <unordered_map>
+#include <unordered_set>
#include "gmock/gmock.h"
#include "gtest/gtest.h"
@@ -48,11 +51,10 @@ struct RawHashSetTestOnlyAccess {
namespace {
-using ::testing::DoubleNear;
using ::testing::ElementsAre;
+using ::testing::Eq;
using ::testing::Ge;
using ::testing::Lt;
-using ::testing::Optional;
using ::testing::Pair;
using ::testing::UnorderedElementsAre;
@@ -75,8 +77,14 @@ TEST(Util, GrowthAndCapacity) {
for (size_t growth = 0; growth < 10000; ++growth) {
SCOPED_TRACE(growth);
size_t capacity = NormalizeCapacity(GrowthToLowerboundCapacity(growth));
- // The capacity is large enough for `growth`
+ // The capacity is large enough for `growth`.
EXPECT_THAT(CapacityToGrowth(capacity), Ge(growth));
+ // For (capacity+1) < kWidth, growth should equal capacity.
+ if (capacity + 1 < Group::kWidth) {
+ EXPECT_THAT(CapacityToGrowth(capacity), Eq(capacity));
+ } else {
+ EXPECT_THAT(CapacityToGrowth(capacity), Lt(capacity));
+ }
if (growth != 0 && capacity > 1) {
// There is no smaller capacity that works.
EXPECT_THAT(CapacityToGrowth(capacity / 2), Lt(growth));
@@ -1875,18 +1883,34 @@ TEST(RawHashSamplerTest, Sample) {
auto& sampler = HashtablezSampler::Global();
size_t start_size = 0;
- start_size += sampler.Iterate([&](const HashtablezInfo&) { ++start_size; });
+ std::unordered_set<const HashtablezInfo*> preexisting_info;
+ start_size += sampler.Iterate([&](const HashtablezInfo& info) {
+ preexisting_info.insert(&info);
+ ++start_size;
+ });
std::vector<IntTable> tables;
for (int i = 0; i < 1000000; ++i) {
tables.emplace_back();
tables.back().insert(1);
+ tables.back().insert(i % 5);
}
size_t end_size = 0;
- end_size += sampler.Iterate([&](const HashtablezInfo&) { ++end_size; });
+ std::unordered_map<size_t, int> observed_checksums;
+ end_size += sampler.Iterate([&](const HashtablezInfo& info) {
+ if (preexisting_info.count(&info) == 0) {
+ observed_checksums[info.hashes_bitwise_xor.load(
+ std::memory_order_relaxed)]++;
+ }
+ ++end_size;
+ });
EXPECT_NEAR((end_size - start_size) / static_cast<double>(tables.size()),
0.01, 0.005);
+ EXPECT_EQ(observed_checksums.size(), 5);
+ for (const auto& [_, count] : observed_checksums) {
+ EXPECT_NEAR((100 * count) / static_cast<double>(tables.size()), 0.2, 0.05);
+ }
}
#endif // ABSL_INTERNAL_HASHTABLEZ_SAMPLE