Added fuzzer tests #500
Added fuzzer tests #500
Conversation
matteosz
force-pushed
the
matteo-fuzz-testing1
branch
from
5b3f446 to
06e5dc1
Compare
matteosz
force-pushed
the
matteo-fuzz-testing1
branch
from
06e5dc1 to
7da5f54
Compare
AnomalRoil
force-pushed
the
matteo-fuzz-testing1
branch
from
7da5f54 to
234bbc5
Compare
sign/tbls/tbls_test.go
Outdated
| suite := bn256.NewSuite() | ||
| n := 10 | ||
| t := n/2 + 1 | ||
| th := n/2 + 1 | ||
|
|
||
| secret := suite.G1().Scalar().Pick(suite.RandomStream()) | ||
| priPoly := share.NewPriPoly(suite.G2(), t, secret, suite.RandomStream()) | ||
| priPoly := share.NewPriPoly(suite.G2(), th, secret, suite.RandomStream()) | ||
| pubPoly := priPoly.Commit(suite.G2().Point().Base()) | ||
| sigShares := make([][]byte, 0) |
There was a problem hiding this comment.
I wonder if this setup phase really needs to be repeated when fuzzing or whether fuzzing should occur on the message only, WDYT?
There was a problem hiding this comment.
I was thinking that fuzzing on the parameter n would have been more interesting than using just; I'd say that the cost of repeating these operations is not that expensive and leads some value in increased exploration. But if you feel it's relevant to fuzz only on n I can change it
There was a problem hiding this comment.
In general, fuzzing should be repeatable. Given the same "fuzzing corpus of input messages" I should get the same outputs/results in all runs.
Could we change this code so that there is no randomness outside of the fuzzer msg?
Maybe seed the RandomStream using msg.
Add a more prominent disclaimer about target audience of this library.
matteosz
force-pushed
the
matteo-fuzz-testing1
branch
from
234bbc5 to
f496513
Compare
matteosz
force-pushed
the
matteo-fuzz-testing1
branch
from
c85c0ef to
8f0aa7c
Compare
matteosz
force-pushed
the
matteo-fuzz-testing1
branch
from
b717b41 to
5a66cf2
Compare
|
Please retry analysis of this Pull-Request directly on SonarCloud |
|
This PR adds some fuzzers to sign/tbls and sign/bls. It also reimplements the bls tests which were commented out in the drandmerge fork