Inline the only callers of TypedArrayBase::{setRange,zeroRange}

Source/modules/webaudio/AudioBuffer.cpp

Index: Source/modules/webaudio/AudioBuffer.cpp
diff --git a/Source/modules/webaudio/AudioBuffer.cpp b/Source/modules/webaudio/AudioBuffer.cpp
index 54612e86bec61f079a18752d9fe652fee8b38a2a..a3462590825de07f0a8a177e07bdca7ac40b3f3a 100644
--- a/Source/modules/webaudio/AudioBuffer.cpp
+++ b/Source/modules/webaudio/AudioBuffer.cpp
@@ -176,7 +176,9 @@ AudioBuffer::AudioBuffer(AudioBus* bus)
             return;
 
         channelDataArray->setNeuterable(false);
-        channelDataArray->setRange(bus->channel(i)->data(), m_length, 0);
+        const float* src = bus->channel(i)->data();
+        float* dst = channelDataArray->data();
+        memmove(dst, src, m_length * sizeof(*dst));

[adamk, 2015/06/12 17:59:25]

For WebAudio reviewers: this memmove is from the original, generic code. I
suspect this could be memcpy here, though I also don't know how important perf
is here.

[Raymond Toy, 2015/06/12 18:05:50]

If the original did this, I'm happy with leaving it as is.

Also, the original had some checks on the parameters.  I think these checks
should be here. (Unless we know the checks will always pass. I didn't look at
that.)

[adamk, 2015/06/12 18:14:28]

The destination of this memmove was just created with size m_length, so this is
safe. I could make it look safer by calling
channelDataArray->view()->byteLength() (which will do the same math internally
that I'm doing here). Would you prefer that?

         m_channels.append(channelDataArray);
     }
 }
@@ -317,8 +319,10 @@ void AudioBuffer::copyToChannel(DOMFloat32Array* source, long channelNumber, uns
 void AudioBuffer::zero()
 {
     for (unsigned i = 0; i < m_channels.size(); ++i) {
-        if (getChannelData(i))
-            getChannelData(i)->zeroRange(0, length());
+        if (DOMFloat32Array* array = getChannelData(i)) {
+            float* data = array->data();
+            memset(data, 0, length() * sizeof(*data));

[Raymond Toy, 2015/06/12 18:05:50]

The original had some checks on the parameters. Are we guaranteed we're safe
here?

[adamk, 2015/06/12 18:14:28]

Same as above, all the channel data arrays have the same length as m_len gth.
But I can make this look safer by calling byteLength() if you'd prefer.

+        }
     }
 }