in try_consume_batch. In case producers constantly produce
new data, try_consume_batch will be called over and over again
leading to stack overflow.

mpsc_queue: extend queue to support arbitrary sized values

It needs to be further debugged

- rename try_consume to try_consume_batch
- rename read_accessor to batch_type
- clear cachelines only after user callback executes:
  this allows iterating over the data arbitrary number of times
  (including 0)
- call user callback within a transaction to simplify logic

- remove recover() method (try_consume can now be used)
- store written offset on pmem so that we can recover
  consumer/producer offsets on restart
- call user callback twice in try_consume if wraparound
  happens

And remove size parameter from mpsc_queue ctor - rely on pmem_log_type size.

Fix windows CI build

we want to enable builds, but current queue implementation
is problematic on Windows - it will be fixed along other queue changes.

Radix: concurrent iteration

Iterator operator++/--:
 Removing nodes/leaves which are pointed to be iterators might
 result in operator++/operator-- failures (since nodes/leaves are
 detached from the tree find_leaf/next_leaf functions will not
 find a proper entry).

 Fix this by adding a fallback path which just searches from the top
 (using lower_bound/upper_bound).

internal_bound:
 Do not traverse the tree twice. Instead save the path from first
 traversal to a vector and just iterate over the vector. Before
 returning iterator to the user, we can check if the path changed
 (if any node on the path was modified). If so, the operation needs
 to be retried.

Revert "gha: temporarily disable codecov"

This reverts commit 7a41d10f.

Extend run-build.sh error msg if no build_steps provided

Mpsc queue

* Make implementation header only
* Opaque implementation in namespace
* Use std::atomic instead of custom implementation
* Modify ringbuf test to work with c++ changes

Importi multi-producer single-consumer (MPSC) ring buffer from https://github.com/rmind/ringbuf

Implement defer free for radix_tree

and enable overwrite tests (since runtime_initialize_mt is
available).

which enables single-writer multi-reader concurrency. After calling
runtime_initialize_mt() each assignment (done through assing_val or
insert_or_assign) does not modify the value in-place but rather
allocates new node. Also, erase does not free nodes/leaves immediately
but adds them to a garbage list which can be processed using
garbage_collect.

gha: temporarily disable codecov

gha: temporarily disable windows builds