meta-treedb can implement malloc functionality, and much more!
Need a heap that allocates of a fixed size chunk of memory? No problem.
But where treedb comes into its own is when it comes to allocating from a mmap()ed disk file - the memory becomes persistent.
If you follow the tests/examples, you too can store dictionaries, trees, lists or whatever you can think of in disk-based memory, just an open() and mmap() away.
Meta-treedb comes with an AVL tree, doubly-linked-list and variable-entry-sized-array implementations for you to use in your data structures.
So, what distinguishes meta-treedb from treedb? "class" interfaces for treedb collections
These are vtables (virtual function tables) that implement th interface. faster compiles - only the "class" interfaces are visible to client code
This is at the expense of a small performance hit.
* the ability to use implementations that were developed after client code was compiled.