Memory footprint optimization: duplicate arrays

[mpollmeier]

Looking at the "duplicate objects" inspection from yourkit, we currently have many duplicates of very large arrays.
In this instance it's a cpg of the linux4 kernel, created by c2cpg and imported into joern.
The entire graph requires 20g of heap, out of which roughly 1g is wasted on duplicates of very large arrays: