Fix Cadence's unsafe random

[tarakby]

Closes onflow/cadence#2691

Emulator now supports Cadence random function(s). The randoms are currently deterministic and equal for all blocks.

---

For contributor use:

• Targeted PR against master branch
• Linked to GitHub issue with discussion and accepted design OR link to spec that describes this work
• Code follows the standards mentioned here
• Re-reviewed Files changed in the GitHub PR explorer
• Added appropriate labels

[tarakby]

This causes randomness on emulator to be:

1. deterministic
2. same for all txs/blocks

Feel free to update this to be any behaviour you prefer.

[codecov-commenter]

Codecov Report

Merging #459 (39979a1) into master (d770012) will increase coverage by 0.10%.
The diff coverage is 37.50%.

@@            Coverage Diff             @@
##           master     #459      +/-   ##
==========================================
+ Coverage   54.89%   55.00%   +0.10%     
==========================================
  Files          28       28              
  Lines        3614     3620       +6     
==========================================
+ Hits         1984     1991       +7     
  Misses       1470     1470              
+ Partials      160      159       -1     

Flag	Coverage Δ
unittests	55.00% <37.50%> (+0.10%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

Files Changed	Coverage Δ
server/access/rest.go	0.00% <0.00%> (ø)
emulator/blockchain.go	72.41% <100.00%> (+0.07%)	⬆️

... and 1 file with indirect coverage changes

[bluesign]

does this break mainnet/testnet transaction replay ( simulation ) ? Is there a way we can solve this ? maybe by getting block (transaction ) random seed from archive nodes or similar?

[janezpodhostnik]

does this break mainnet/testnet transaction replay ( simulation ) ? Is there a way we can solve this ? maybe by getting block (transaction ) random seed from archive nodes or similar?

yes, it does break it... but, like you said, we can fix it by exposing the random beacon.
@tarakby is there anyway to access the random beacon for sealed blocks?

[bluesign]

does this break mainnet/testnet transaction replay ( simulation ) ? Is there a way we can solve this ? maybe by getting block (transaction ) random seed from archive nodes or similar?

yes, it does break it... but, like you said, we can fix it by exposing the random beacon. @tarakby is there anyway to access the random beacon for sealed blocks?

maybe we can add to transaction result as random seed.

[tarakby]

@tarakby is there anyway to access the random beacon for sealed blocks?

The issue with the source of randomness is that it is part of the block payload data not the execution state. Archive nodes and other nodes in general are not required to store these data for all the history (most of them only store data since the root block of the current spork only, but not before that).

maybe we can add to transaction result as random seed.

Yes this is a good idea - We can restrict it only to transactions that used Cadence's random function, since most of the transactions do not need it.

I have no idea how to plug that seed into the emulator later when replaying a transaction. Does this PR need further changes to accommodate that?

[bluesign]

I think we can merge this in the mean time, for sure random seed in tx result would take some time

[btspoony]

Hi @turbolent, Approximately when can this PR be included in a certain release of flow-cli?

[turbolent]

@btspoony I only assist with reviews in this project. Please open an issue in this and the CLI repo. Thank you!