You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
So, basically there are differences in the output of the go 1.18 runtime vs 1.19, specifically with regard to crypto libs.
In the world of validator sets, this can break down into apphash errors.
Since some validators will use go 1.18, and some go 1.19, it makes sense to set go 1.19 in go.mod because then go1.19 is required to build and you end up with only a single go runtime, and you don't run into the same errors that would otherwise occur with mixed runtimes. Setting to go 1.18 still allows for validators to mix and... they will, so that's the reasoning behind using go 1.19 out of the gate, despite the fact that the cosmos-sdk v0.46.x uses go 1.18. (all validators will need to use go 1.19)
The text was updated successfully, but these errors were encountered:
So, basically there are differences in the output of the go 1.18 runtime vs 1.19, specifically with regard to crypto libs.
In the world of validator sets, this can break down into apphash errors.
Since some validators will use go 1.18, and some go 1.19, it makes sense to set go 1.19 in go.mod because then go1.19 is required to build and you end up with only a single go runtime, and you don't run into the same errors that would otherwise occur with mixed runtimes. Setting to go 1.18 still allows for validators to mix and... they will, so that's the reasoning behind using go 1.19 out of the gate, despite the fact that the cosmos-sdk v0.46.x uses go 1.18. (all validators will need to use go 1.19)
The text was updated successfully, but these errors were encountered: