Differentiating results between local and CI runs

[elenadimitrova]

Using slither@0.6.12 we have successfully ignored a number of detector-raised warnings for the argent-contracts repo however the following informational messages are being raised in Circle CI:

INFO:Detectors:
setName(bytes32,string) should be declared external:
	- ArgentENSResolver.setName(bytes32,string) (infrastructure/ens/ArgentENSResolver.sol#59-62)
	- ENSResolver.setName(bytes32,string) (infrastructure/ens/ENSResolver.sol#28)
addr(bytes32) should be declared external:
	- ArgentENSResolver.addr(bytes32) (infrastructure/ens/ArgentENSResolver.sol#69-71)
	- ENSResolver.addr(bytes32) (infrastructure/ens/ENSResolver.sol#25)
Reference: https://github.com/crytic/slither/wiki/Detector-Documentation#public-function-that-could-be-declared-as-external
INFO:Slither:. analyzed (69 contracts with 28 detectors), 2 result(s) found

while not when run locally:

> argent-contracts@1.0.0 security:slither /Users/Elena/Source/argent-contracts
> rm -rf build && slither . --filter-paths lib --solc-disable-warnings --exclude-low --exclude-informational --exclude=naming-convention,unused-state-variables,solc-version,assembly-usage,low-level-calls
....
INFO:Slither:. analyzed (69 contracts with 28 detectors), 0 result(s) found

[elenadimitrova]

Providing a more fresh example from our last night's build:

The following command npm run security:slither:infrastructure_0.5 produces in CI:

INFO:Detectors:
setName(bytes32,string) should be declared external:
	- ArgentENSResolver.setName(bytes32,string) (infrastructure_0.5/ens/ArgentENSResolver.sol#59-62)
	- ENSResolver.setName(bytes32,string) (infrastructure_0.5/ens/ENSResolver.sol#28)
Reference: https://github.com/crytic/slither/wiki/Detector-Documentation#public-function-that-could-be-declared-external
INFO:Slither:. analyzed (31 contracts with 28 detectors), 1 result(s) found

while locally we get no errors:

INFO:Slither:. analyzed (31 contracts with 28 detectors), 0 result(s) found

[elenadimitrova]

On top of this we get the same build pass one day and fail the next 🤷‍♀️
As well as fail here then pass the next day

[montyly]

Thanks @elenadimitrova. We are going to investigate why this happens