Fix #457 utils.MaxSlice and utils.MinSlice

[romainbou]

Addressing #457

• Fix utils.MaxSlice and utils.MinSlice returning 0
• Fix EqualSlice if slices have difference lengths
• Add tests for slices utils

[Pro7ech]

EqualSlice is not constant time anymore after these changes

[romainbou]

EqualSlice is not constant time anymore after these changes

Ah yes, but before it would also panic early depending on the size of b.

It's essentially the implementation of the standard library. (available since Go 1.21, we could use it once we drop older Go version support) https://cs.opensource.google/go/go/+/refs/tags/go1.22.3:src/slices/slices.go;l=18

The only critical use is here:

lattigo/mhe/keygen_evk.go

Line 131 in c031b14

if !utils.EqualSlice(shareOut.BaseTwoDecompositionVectorSize(), crp.BaseTwoDecompositionVectorSize()) {

Do you think it's really a concern?

[Pro7ech]

EqualSlice is not constant time anymore after these changes

Ah yes, but before it would also panic early depending on the size of b.

It's essentially the implementation of the standard library. (available since Go 1.21, we could use it once we drop older Go version support) https://cs.opensource.google/go/go/+/refs/tags/go1.22.3:src/slices/slices.go;l=18

The only critical use is here:

lattigo/mhe/keygen_evk.go

Line 131 in c031b14

if !utils.EqualSlice(shareOut.BaseTwoDecompositionVectorSize(), crp.BaseTwoDecompositionVectorSize()) {

Do you think it's really a concern?

To me it is not a concern since it is not used in critical code and but its behavior should then be documented (this is a cryptographic library). Since Go 1.21+ provides native support for this kind of utility, shouldn't it then be removed/replaced by it?