Fix store malfeasant ATX

[fasmat]

Motivation

In specific cases storing an ATX will fail if it is malicious.

Description

atx.AddGettingNonce might return nil, nil for an invalid initial ATX. Specifically an ATX that isn't the first of an identity but references 0x000... as its previous ATX. Such an ATX might not contain a VRF nonce so it will fail to store with the error no nonce when persisting the malfeasance proof.

Additionally I made the check for a double 0x000... previous ATX more efficient by just fetching the initial ATX of the identity rather than going through the complete chain to find the collision.

Test Plan

Added a new test for the fix

TODO

• Explain motivation or link existing issue(s)
• Test changes and document test plan
• Update documentation as needed
• Update changelog as needed

[codecov]

Codecov Report

Attention: Patch coverage is 66.66667% with 8 lines in your changes are missing coverage. Please review.

Project coverage is 80.4%. Comparing base (c1f9fb7) to head (17e54f3).
Report is 4 commits behind head on v1.5.

Files	Patch %	Lines
activation/handler.go	66.6%	4 Missing and 3 partials ⚠️
malfeasance/wire/malfeasance.go	0.0%	1 Missing ⚠️

Additional details and impacted files

@@          Coverage Diff           @@
##            v1.5   #5929    +/-   ##
======================================
  Coverage   80.4%   80.4%            
======================================
  Files        285     286     +1     
  Lines      29579   29736   +157     
======================================
+ Hits       23804   23937   +133     
- Misses      4168    4176     +8     
- Partials    1607    1623    +16     

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

[poszu]

Specifically an ATX that isn't the first of an identity but references 0x000... as its previous ATX. Such an ATX might not contain a VRF nonce (...).

I don't understand, an ATX with previous=EMPTY that doesn't have a VRF nonce is syntactically invalid, isn't it?

[fasmat]

Specifically an ATX that isn't the first of an identity but references 0x000... as its previous ATX. Such an ATX might not contain a VRF nonce (...).

I don't understand, an ATX with previous=EMPTY that doesn't have a VRF nonce is syntactically invalid, isn't it?

You're right, the problem as a different one: blob already stored in atx_blobs. I reverted this part of the PR.