Make Courgette as much as 5x faster.

courgette/memory_allocator.h

Index: courgette/memory_allocator.h
diff --git a/courgette/memory_allocator.h b/courgette/memory_allocator.h
index 5386dad0d7d9ead4b26bd66d083eefb9f2ceb0d1..ada7f40b8438b42c14d4de0c952d552f727677d9 100644
--- a/courgette/memory_allocator.h
+++ b/courgette/memory_allocator.h
@@ -333,6 +333,12 @@ class NoThrowBuffer {
     if (size < kStartSize)
       size = kStartSize;
 
+    // Use a size 1% higher than requested. In practice, this makes Courgette as
+    // much as 5x faster on typical Chrome update payloads as a lot of future
+    // reserve() calls will become no-ops instead of costly resizes that copy
+    // all the data. Note that doing this here instead of outside the function
+    // is more efficient, since it's after the no-op early return checks above.
+    size *= 1.01;

[gab, 2014/09/12 04:12:19]

This will have no impact on reserves below 100, should we also increment |size|
by a fixed amount before applying the multiplier?

Or how about we round |size| up to the nearest multiple of 2? This would most
likely provide nice memory alignment properties while also guaranteeing an
amortized linear time buffer size increase if reserve is called with sequential
indices (which is terrible usage of reserve() and should be fixed in itself).

[gab, 2014/09/12 04:16:06]

Oops s/multiple of 2/power of 2/ of course.

     T* new_buffer = alloc_.allocate(size);
     if (!new_buffer) {
       clear();