Store empty pending block instead of deleting the key

[joshklop]

To decrease the number of block not found errors when simulating transactions against the pending block.

[joshklop]

While Juno is syncing:

• A user can simulate a transaction against the pending state and potentially get a confusing result
• starknet_getBlockWithTxs with the pending param will return the empty pending block

Are these behavior changes acceptable?

[omerfirmak]

• starknet_getBlockWithTxs with the pending param will return the empty pending block

I think that is fine

• A user can simulate a transaction against the pending state and potentially get a confusing result

This I am not sure of

[codecov]

Codecov Report

Patch coverage: 77.19% and project coverage change: -0.01% ⚠️

Comparison is base (935eb23) 73.65% compared to head (0ea311d) 73.65%.

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #1241      +/-   ##
==========================================
- Coverage   73.65%   73.65%   -0.01%     
==========================================
  Files          70       70              
  Lines        7561     7594      +33     
==========================================
+ Hits         5569     5593      +24     
- Misses       1533     1540       +7     
- Partials      459      461       +2     

Files Changed	Coverage Δ
blockchain/blockchain.go	76.80% <77.19%> (+0.10%)	⬆️

... and 2 files with indirect coverage changes

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.