fix(sync): address some minor issues in block sync

[bowenwang1996]

Address a couple issues related to block sync:

• blocks_received is incorrectly calculated. It should not increase when we do not receive any new blocks, but currently if some heights are skipped and we accept a block that was previously orphaned, it does increase.
• The timeout for progress is too short. This is not a problem per se. However, every time block sync is due we recompute the hashes of blocks that we need to sync. When the epoch is long and we just start block sync, this could mean going through 40k blocks every 1-2s, which is fairly suboptimal. The proper fix would be to implement caching for the requested hashes and do not recompute every time block sync is due, but given that we plan to rewrite block sync in the future, I don't see the need to change it now.

Test plan

• test_blocks_received.

[gitpod-io]

[codecov]

Codecov Report

Merging #3104 into master will decrease coverage by 0.52%.
The diff coverage is 94.28%.

@@            Coverage Diff             @@
##           master    #3104      +/-   ##
==========================================
- Coverage   87.94%   87.42%   -0.53%     
==========================================
  Files         212      212              
  Lines       42976    41247    -1729     
==========================================
- Hits        37795    36059    -1736     
- Misses       5181     5188       +7     

Impacted Files	Coverage Δ
chain/chain/tests/sync_chain.rs	89.74% <90.90%> (+1.50%)	⬆️
chain/chain/src/chain.rs	88.62% <100.00%> (-0.69%)	⬇️
chain/client/src/sync.rs	87.15% <100.00%> (+0.48%)	⬆️
runtime/runtime/src/actions.rs	79.02% <0.00%> (-5.80%)	⬇️
chain/client/src/test_utils.rs	87.45% <0.00%> (-4.07%)	⬇️
chain/client/src/types.rs	63.75% <0.00%> (-3.75%)	⬇️
runtime/runtime/src/lib.rs	87.90% <0.00%> (-3.62%)	⬇️
chain/client/src/info.rs	68.99% <0.00%> (-3.11%)	⬇️
chain/chain/src/types.rs	92.25% <0.00%> (-2.36%)	⬇️
neard/src/runtime.rs	86.34% <0.00%> (-2.18%)	⬇️
... and 22 more

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 9ca3117...66f6c1c. Read the comment docs.

[bowenwang1996]

superseded by #3241