Introduce accessors for currently public data members (threads and thread_index) (#1208)

* [benchmark] Introduce accessors for currently public data members `threads` and `thread_index`

Also deprecate the direct access to these fields.

Motivations:

Our internal library provides accessors for those fields because the styleguide disalows accessing classes' data members directly (even if they're const).
There has been a discussion to simply move internal library to make its fields public similarly to the OSS version here, however, the concern is that these kinds of direct access would prevent many types of future design changes (eg how/whether the values would be stored in the data member)

I think the concensus in the end is that we'd change the external library for this case.
AFAIK, there are three important third_party users that we'd need to migrate: tcmalloc, abseil and tensorflow.
Please let me know if I'm missing anyone else.

* [benchmark] Introduce accessors for currently public data members `threads` and `thread_index`

Also deprecate the direct access to these fields.

Motivations:

Our internal library provides accessors for those fields because the styleguide disalows accessing classes' data members directly (even if they're const).
There has been a discussion to simply move internal library to make its fields public similarly to the OSS version here, however, the concern is that these kinds of direct access would prevent many types of future design changes (eg how/whether the values would be stored in the data member)

I think the concensus in the end is that we'd change the external library for this case.
AFAIK, there are three important third_party users that we'd need to migrate: tcmalloc, abseil and tensorflow.
Please let me know if I'm missing anyone else.

* [benchmark] Introduce accessors for currently public data members `threads` and `thread_index`

Also deprecate direct access to `.thread_index` and make threads a private field

Motivations:

Our internal library provides accessors for those fields because the styleguide disalows accessing classes' data members directly (even if they're const).
There has been a discussion to simply move internal library to make its fields public similarly to the OSS version here, however, the concern is that these kinds of direct access would prevent many types of future design changes (eg how/whether the values would be stored in the data member)

I think the concensus in the end is that we'd change the external library for this case.
AFAIK, there are three important third_party users that we'd need to migrate: tcmalloc, abseil and tensorflow.
Please let me know if I'm missing anyone else.

* [benchmark] Introduce accessors for currently public data members `threads` and `thread_index`

Also deprecate direct access to `.thread_index` and make threads a private field

Motivations:

Our internal library provides accessors for those fields because the styleguide disalows accessing classes' data members directly (even if they're const).
There has been a discussion to simply move internal library to make its fields public similarly to the OSS version here, however, the concern is that these kinds of direct access would prevent many types of future design changes (eg how/whether the values would be stored in the data member)

I think the concensus in the end is that we'd change the external library for this case.
AFAIK, there are three important third_party users that we'd need to migrate: tcmalloc, abseil and tensorflow.
Please let me know if I'm missing anyone else.

* [benchmark] Introduce accessors for currently public data members `threads` and `thread_index`

Also deprecate direct access to `.thread_index` and make threads a private field

Motivations:

Our internal library provides accessors for those fields because the styleguide disalows accessing classes' data members directly (even if they're const).
There has been a discussion to simply move internal library to make its fields public similarly to the OSS version here, however, the concern is that these kinds of direct access would prevent many types of future design changes (eg how/whether the values would be stored in the data member)

I think the concensus in the end is that we'd change the external library for this case.
AFAIK, there are three important third_party users that we'd need to migrate: tcmalloc, abseil and tensorflow.
Please let me know if I'm missing anyone else.

* [benchmark] Introduce accessors for currently public data members `threads` and `thread_index`

Also deprecate direct access to `.thread_index` and make threads a private field

Motivations:

Our internal library provides accessors for those fields because the styleguide disalows accessing classes' data members directly (even if they're const).
There has been a discussion to simply move internal library to make its fields public similarly to the OSS version here, however, the concern is that these kinds of direct access would prevent many types of future design changes (eg how/whether the values would be stored in the data member)

I think the concensus in the end is that we'd change the external library for this case.
AFAIK, there are three important third_party users that we'd need to migrate: tcmalloc, abseil and tensorflow.
Please let me know if I'm missing anyone else.

* [benchmark] Introduce accessors for currently public data members `threads` and `thread_index`

Also deprecate direct access to `.thread_index` and make threads a private field

Motivations:

Our internal library provides accessors for those fields because the styleguide disalows accessing classes' data members directly (even if they're const).
There has been a discussion to simply move internal library to make its fields public similarly to the OSS version here, however, the concern is that these kinds of direct access would prevent many types of future design changes (eg how/whether the values would be stored in the data member)

I think the concensus in the end is that we'd change the external library for this case.
AFAIK, there are three important third_party users that we'd need to migrate: tcmalloc, abseil and tensorflow.
Please let me know if I'm missing anyone else.

* [benchmark] Introduce accessors for currently public data members `threads` and `thread_index`

Also deprecate direct access to `.thread_index` and make threads a private field

Motivations:

Our internal library provides accessors for those fields because the styleguide disalows accessing classes' data members directly (even if they're const).
There has been a discussion to simply move internal library to make its fields public similarly to the OSS version here, however, the concern is that these kinds of direct access would prevent many types of future design changes (eg how/whether the values would be stored in the data member)

I think the concensus in the end is that we'd change the external library for this case.
AFAIK, there are three important third_party users that we'd need to migrate: tcmalloc, abseil and tensorflow.
Please let me know if I'm missing anyone else.

* [benchmark] Introduce accessors for currently public data members `threads` and `thread_index`

Also deprecate direct access to `.thread_index` and make threads a private field

Motivations:

Our internal library provides accessors for those fields because the styleguide disalows accessing classes' data members directly (even if they're const).
There has been a discussion to simply move internal library to make its fields public similarly to the OSS version here, however, the concern is that these kinds of direct access would prevent many types of future design changes (eg how/whether the values would be stored in the data member)

I think the concensus in the end is that we'd change the external library for this case.
AFAIK, there are three important third_party users that we'd need to migrate: tcmalloc, abseil and tensorflow.
Please let me know if I'm missing anyone else.

* [benchmark] Introduce accessors for currently public data members `threads` and `thread_index`

Also deprecate direct access to `.thread_index` and make threads a private field

Motivations:

Our internal library provides accessors for those fields because the styleguide disalows accessing classes' data members directly (even if they're const).
There has been a discussion to simply move internal library to make its fields public similarly to the OSS version here, however, the concern is that these kinds of direct access would prevent many types of future design changes (eg how/whether the values would be stored in the data member)

I think the concensus in the end is that we'd change the external library for this case.
AFAIK, there are three important third_party users that we'd need to migrate: tcmalloc, abseil and tensorflow.
Please let me know if I'm missing anyone else.

3 files changed, 16 insertions, 16 deletions

diff --git a/test/benchmark_test.cc b/test/benchmark_test.cc
index 3cd4f55..c63fc31 100644
--- a/test/benchmark_test.cc
+++ b/test/benchmark_test.cc
@@ -126,7 +126,7 @@ static void BM_StringCompare(benchmark::State& state) {
 BENCHMARK(BM_StringCompare)->Range(1, 1 << 20);
 
 static void BM_SetupTeardown(benchmark::State& state) {
-  if (state.thread_index == 0) {
+  if (state.thread_index() == 0) {
     // No need to lock test_vector_mu here as this is running single-threaded.
     test_vector = new std::vector<int>();
   }
@@ -139,7 +139,7 @@ static void BM_SetupTeardown(benchmark::State& state) {
       test_vector->pop_back();
     ++i;
   }
-  if (state.thread_index == 0) {
+  if (state.thread_index() == 0) {
     delete test_vector;
   }
 }
@@ -156,11 +156,11 @@ BENCHMARK(BM_LongTest)->Range(1 << 16, 1 << 28);
 
 static void BM_ParallelMemset(benchmark::State& state) {
   int64_t size = state.range(0) / static_cast<int64_t>(sizeof(int));
-  int thread_size = static_cast<int>(size) / state.threads;
-  int from = thread_size * state.thread_index;
+  int thread_size = static_cast<int>(size) / state.threads();
+  int from = thread_size * state.thread_index();
   int to = from + thread_size;
 
-  if (state.thread_index == 0) {
+  if (state.thread_index() == 0) {
     test_vector = new std::vector<int>(static_cast<size_t>(size));
   }
 
@@ -172,7 +172,7 @@ static void BM_ParallelMemset(benchmark::State& state) {
     }
   }
 
-  if (state.thread_index == 0) {
+  if (state.thread_index() == 0) {
     delete test_vector;
   }
 }
@@ -223,14 +223,14 @@ BENCHMARK_CAPTURE(BM_non_template_args, basic_test, 0, 0);
 static void BM_DenseThreadRanges(benchmark::State& st) {
   switch (st.range(0)) {
     case 1:
-      assert(st.threads == 1 || st.threads == 2 || st.threads == 3);
+      assert(st.threads() == 1 || st.threads() == 2 || st.threads() == 3);
       break;
     case 2:
-      assert(st.threads == 1 || st.threads == 3 || st.threads == 4);
+      assert(st.threads() == 1 || st.threads() == 3 || st.threads() == 4);
       break;
     case 3:
-      assert(st.threads == 5 || st.threads == 8 || st.threads == 11 ||
-             st.threads == 14);
+      assert(st.threads() == 5 || st.threads() == 8 || st.threads() == 11 ||
+             st.threads() == 14);
       break;
     default:
       assert(false && "Invalid test case number");
diff --git a/test/fixture_test.cc b/test/fixture_test.cc
index eba0a42..452dd87 100644
--- a/test/fixture_test.cc
+++ b/test/fixture_test.cc
@@ -9,14 +9,14 @@
 class FIXTURE_BECHMARK_NAME : public ::benchmark::Fixture {
  public:
   void SetUp(const ::benchmark::State& state) BENCHMARK_OVERRIDE {
-    if (state.thread_index == 0) {
+    if (state.thread_index() == 0) {
       assert(data.get() == nullptr);
       data.reset(new int(42));
     }
   }
 
   void TearDown(const ::benchmark::State& state) BENCHMARK_OVERRIDE {
-    if (state.thread_index == 0) {
+    if (state.thread_index() == 0) {
       assert(data.get() != nullptr);
       data.reset();
     }
@@ -35,7 +35,7 @@ BENCHMARK_F(FIXTURE_BECHMARK_NAME, Foo)(benchmark::State &st) {
 }
 
 BENCHMARK_DEFINE_F(FIXTURE_BECHMARK_NAME, Bar)(benchmark::State& st) {
-  if (st.thread_index == 0) {
+  if (st.thread_index() == 0) {
     assert(data.get() != nullptr);
     assert(*data == 42);
   }
diff --git a/test/skip_with_error_test.cc b/test/skip_with_error_test.cc
index 1156bc0..4d89671 100644
--- a/test/skip_with_error_test.cc
+++ b/test/skip_with_error_test.cc
@@ -97,7 +97,7 @@ ADD_CASES("BM_error_before_running_range_for", {{"", true, "error message"}});
 void BM_error_during_running(benchmark::State& state) {
   int first_iter = true;
   while (state.KeepRunning()) {
-    if (state.range(0) == 1 && state.thread_index <= (state.threads / 2)) {
+    if (state.range(0) == 1 && state.thread_index() <= (state.threads() / 2)) {
       assert(first_iter);
       first_iter = false;
       state.SkipWithError("error message");
@@ -142,7 +142,7 @@ void BM_error_after_running(benchmark::State& state) {
   for (auto _ : state) {
     benchmark::DoNotOptimize(state.iterations());
   }
-  if (state.thread_index <= (state.threads / 2))
+  if (state.thread_index() <= (state.threads() / 2))
     state.SkipWithError("error message");
 }
 BENCHMARK(BM_error_after_running)->ThreadRange(1, 8);
@@ -154,7 +154,7 @@ ADD_CASES("BM_error_after_running", {{"/threads:1", true, "error message"},
 void BM_error_while_paused(benchmark::State& state) {
   bool first_iter = true;
   while (state.KeepRunning()) {
-    if (state.range(0) == 1 && state.thread_index <= (state.threads / 2)) {
+    if (state.range(0) == 1 && state.thread_index() <= (state.threads() / 2)) {
       assert(first_iter);
       first_iter = false;
       state.PauseTiming();