Support variable block sizes in the LRU stack

[thedavidchu]

• I don't understand what's going on in the splay tree code well enough to rewrite it to incorporate this.
  • At first, the reason was that I'm trying to infer the weight of nodes from the structure rather than just storing this weight explicitly.
  • I fixed this but it is now broken in that state. Run meson test debug_tree_test to see the brokenness

The (anonymized) Meson test output is here:

Log of Meson test suite run on 2024-10-23T16:56:19.390647

==================================== 1/1 =====================================
test:         debug_tree_test
start time:   16:56:19
duration:     0.00s
result:       exit status 1
command:      ASAN_OPTIONS=halt_on_error=1:abort_on_error=1:print_summary=1 MALLOC_PERTURB_=79 UBSAN_OPTIONS=halt_on_error=1:abort_on_error=1:print_summary=1:print_stacktrace=1 
----------------------------------- stdout -----------------------------------
Key: 92, Size: 92
---
  Key: 92, Size: 31
Key: 31, Size: 62
---
[ERROR] [2024-10-23 16:56:19] [ ../src/lib/tree/basic_tree.c:349 ] [errno 0: Success] mismatching cardinalities: 31 vs 61
[ERROR] ../src/lib/tree/basic_tree.c:399
[ERROR] [2024-10-23 16:56:19] [ ../test/lib_test/tree_test/debug_tree_test.c:36 ] [errno 0: Success] validation failed
==============================================================================


Summary of Failures:

1/1 debug_tree_test FAIL            0.00s   exit status 1

Ok:                 0   
Expected Fail:      0   
Fail:               1   
Unexpected Pass:    0   
Skipped:            0   
Timeout:            0   

The tree should look like:

  Key: 92, Size: 92
Key: 31, Size: 123

• I think this will negatively impact performance because it will either require larger memory sizes (to include the size of each specific object) or will require me to access more nodes (to find the weight of the current node)