AVM: Unify inner transaction ID calculation

[jasonpaulos]

Summary

Unifies the inner transaction ID calculation with a new consensus parameter, UnifyInnerTxIDs.

Supersedes #3806

Test Plan

Unit tests added for new and current behavior.

[codecov-commenter]

Codecov Report

Merging #3927 (b41bfb9) into master (12ded27) will decrease coverage by 0.00%.
The diff coverage is 100.00%.

@@            Coverage Diff             @@
##           master    #3927      +/-   ##
==========================================
- Coverage   50.09%   50.09%   -0.01%     
==========================================
  Files         394      394              
  Lines       68456    68489      +33     
==========================================
+ Hits        34294    34309      +15     
- Misses      30465    30477      +12     
- Partials     3697     3703       +6     

Impacted Files	Coverage Δ
config/consensus.go	85.58% <100.00%> (+0.04%)	⬆️
data/transactions/logic/eval.go	90.00% <100.00%> (+0.12%)	⬆️
ledger/blockqueue.go	82.18% <0.00%> (-2.88%)	⬇️
network/wsPeer.go	69.16% <0.00%> (-2.50%)	⬇️
catchup/peerSelector.go	98.95% <0.00%> (-1.05%)	⬇️
data/abi/abi_type.go	87.67% <0.00%> (-0.95%)	⬇️
cmd/tealdbg/debugger.go	72.69% <0.00%> (-0.81%)	⬇️
catchup/service.go	70.12% <0.00%> (+0.49%)	⬆️

---

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 12ded27...b41bfb9. Read the comment docs.

[jannotti]

I think this looks good. One thing that should maybe be noted somewhere - I had previously hoped that if this hits mainnet and there were still no examples of any code grabbing an inner txn id, we'd be able to remove this code. But even if that doesn't happen, group ids have been calculated, and appear in blocks using these ids as their input, so we're stuck forever, right?

[jasonpaulos]

I think this looks good. One thing that should maybe be noted somewhere - I had previously hoped that if this hits mainnet and there were still no examples of any code grabbing an inner txn id, we'd be able to remove this code. But even if that doesn't happen, group ids have been calculated, and appear in blocks using these ids as their input, so we're stuck forever, right?

That's right, the new consensus param affects GroupID calculations, so we unfortunately have to keep at least some of the existing code around. But we could in theory fix the caching issue and different respones from itxn TxID vs an inner's txn TxID as long as these haven't been used.