Benchmarked memory models for Toy_Array, read more

The results can be found in
'tests/benchmarks/array_allocation/results.md'

The results are disappointing, as 'malloc()' is simply faster in every
possible situation compared to my custom arena allocator.

source/toy_bucket.h

@@ -2,10 +2,10 @@
 
 #include "toy_common.h"
 
-//NOTE: this structure has restrictions on it's usage:
+//NOTE: this is an 'arena allocator', and has restrictions on it's usage:
 // - It can only expand until it is freed
-// - It cannot be copied around within RAM
-// - It cannot allocate more memory than it has capacity
+// - It cannot be copied or moved around in memory
+// - It cannot allocate more memory than it has 'capacity'
 // If each of these rules are followed, this is actually more efficient than other options
 
 //a custom allocator