v2.1: streamline status cache insert (backport of #3365)

[mergify]

Problem

1. There are some redundant operations as part of status cache insert, which is called for every transaction (getting cache entry & setting max slot)
2. Lacking comments to quickly parse what's happening

Summary of Changes

1. Pull out a new function, add_to_slot_delta, in the status cache lib for adding a keyslice to a slot delta entry.
2. Call this new function from fn insert and fn insert_with_slice
3. Don't call fn insert_with_slice from fn insert
4. Add some comments to insert to make it easy to follow

   ---

   This is an automatic backport of pull request streamline status cache insert #3365 done by Mergify.

[t-nelson]

@bw-solana do we still want this one?

[bw-solana]

@bw-solana do we still want this one?

I'll need to take a closer look at #3796 to see what @alessandrod did with the status cache to understand how much we care to take this

[alessandrod]

I'm not planning to backport #3796 so yeah I'd say merge?

(...unless? no lol jk)

[t-nelson]

what does "streamline" mean here? is this purely cosmetic? why didn't insert_with_slice() get similar treatment?

[bw-solana]

what does "streamline" mean here? is this purely cosmetic? why didn't insert_with_slice() get similar treatment?

the streamlining is reducing the amount of work done in insert. Specifically, avoiding redundant hashmap entry fetch and setting of max slot:

        let hash_map =
            self.cache
                .entry(*transaction_blockhash)
                .or_insert((slot, key_index, HashMap::new()));
        hash_map.0 = std::cmp::max(slot, hash_map.0);

While in the area, a few cosmetics improvements were made to do things like destructuring tuples, combining lines, adding comments, etc.

I'm happy to make any changes to make this safer, more reviewable, etc. by decoupling functional/cosmetic changes, BPing functional changes only, or making cosmetic improvements to insert_with_slice as well

[t-nelson]

we agreed to let this one ride on the call, no?

[bw-solana]

we agreed to let this one ride on the call, no?

I don't remember at all - cache got blown away. I'm fine either way