NetEq: Fix a bug in expand_rate and speech_expand_rate calculation

webrtc/modules/audio_coding/neteq/statistics_calculator.cc

Index: webrtc/modules/audio_coding/neteq/statistics_calculator.cc
diff --git a/webrtc/modules/audio_coding/neteq/statistics_calculator.cc b/webrtc/modules/audio_coding/neteq/statistics_calculator.cc
index b47ca127dbe7a2107b4315a7bd4eece1ef0d19fa..5214dde97758c33d08c4d9bada6e229b4022c988 100644
--- a/webrtc/modules/audio_coding/neteq/statistics_calculator.cc
+++ b/webrtc/modules/audio_coding/neteq/statistics_calculator.cc
@@ -22,6 +22,17 @@
 
 namespace webrtc {
 
+namespace {
+size_t AddIntToSizeTWithLowerCap(int a, size_t b) {
+  if (a < 0 && static_cast<size_t>(-a) > b) {

[kwiberg-webrtc, 2017/05/04 19:55:04]

if (safe_cmp::Gt(-a, b)) {

Both your and my version will misbehave if a == INT_MIN, though, so maybe

if (safe_cmp::Gt(-int64_t{a}, b)) {

[minyue-webrtc, 2017/05/05 07:14:00]

how about

size_t ret = b + a; // you may cast a if you want
return (a < 0 && ret > b) ? 0 : ret
  

[hlundin-webrtc, 2017/05/05 07:42:14]

Is it conceivable that abs(int_min) is larger than max(size_t) on any platform?
That would break Minyue's proposal. Otherwise, I like it.

I implemented Karl's suggestion for now. Let me know if you think I should
change.

[kwiberg-webrtc, 2017/05/05 08:34:29]

AFAICT, the standard just says that "The type size_t is an
implementation-defined unsigned integer type that is large enough to contain the
size in bytes of any object." But I doubt there's any system whose size_t is
smaller than int. You could static_assert it if you want.
Yes, I like it too---it makes neat use of the fact that signed-to-unsigned
conversion is well-defined (which is not the case for unsigned-to-signed
conversion).
Minyue's version is probably better, since it avoids using 64-bit types on
32-bit platforms.

[hlundin-webrtc, 2017/05/05 11:40:09]

Changed to Minyue's version.

+    // Simply doing a + b would give a negative result, with a negative wrap.
+    // Cap it to zero instead.
+    return 0;
+  }
+  return a + b;

[kwiberg-webrtc, 2017/05/04 19:55:04]

I'm pretty sure a will be promoted to size_t before the addition so that
everything just works, but consider making it explicit:

  return dchecked_cast<size_t>(a) + b;

[hlundin-webrtc, 2017/05/05 07:42:15]

Done.

+}
+}
+
 // Allocating the static const so that it can be passed by reference to
 // RTC_DCHECK.
 const size_t StatisticsCalculator::kLenWaitingTimes;
@@ -148,6 +159,16 @@ void StatisticsCalculator::ExpandedNoiseSamples(size_t num_samples) {
   expanded_noise_samples_ += num_samples;
 }
 
+void StatisticsCalculator::ExpandedVoiceSamplesCorrection(int num_samples) {
+  expanded_speech_samples_ =
+      AddIntToSizeTWithLowerCap(num_samples, expanded_speech_samples_);
+}
+
+void StatisticsCalculator::ExpandedNoiseSamplesCorrection(int num_samples) {
+  expanded_noise_samples_ =
+      AddIntToSizeTWithLowerCap(num_samples, expanded_noise_samples_);
+}
+
 void StatisticsCalculator::PreemptiveExpandedSamples(size_t num_samples) {
   preemptive_samples_ += num_samples;
 }