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
Right now, if you write a test where the RPC messages are susceptible to data nondeterminism, when we try to inject a fault for those RPCs, we will not actually inject the fault. To solve the problem, the developer needs to set the dataNondeterminism = true option in the test annotation which alters the search algorithm to deal with this.
However, developers using Filibuster may be unaware that their code contains data nondeterminism. Therefore, we should warn the developer if we tried to inject faults that we couldn't. Right now, in the test reporting there is some code that finds places where the RPCs we tried to inject faults on doesn't match the RPCs we observed in the exception. We should use this information to warn the developer about this and recommend that the data nondeterminism option is used.
The text was updated successfully, but these errors were encountered:
Right now, if you write a test where the RPC messages are susceptible to data nondeterminism, when we try to inject a fault for those RPCs, we will not actually inject the fault. To solve the problem, the developer needs to set the
dataNondeterminism = true
option in the test annotation which alters the search algorithm to deal with this.However, developers using Filibuster may be unaware that their code contains data nondeterminism. Therefore, we should warn the developer if we tried to inject faults that we couldn't. Right now, in the test reporting there is some code that finds places where the RPCs we tried to inject faults on doesn't match the RPCs we observed in the exception. We should use this information to warn the developer about this and recommend that the data nondeterminism option is used.
The text was updated successfully, but these errors were encountered: