Add extra debugging information to Bech32 decoding corruption tests. #327
Conversation
…aracters. Additionally, make it a precondition that the adjacent characters are actually different to one another.
…cters. Additionally, make it a precondition that the string has actually been successfully corrupted.
…cters. Additionally, make it a precondition that the string has actually been successfully corrupted.
| , " original string: " <> show originalString | ||
| , "corrupted string: " <> show corruptedString ] | ||
| return $ counterexample description $ | ||
| char1 /= char2 ==> |
There was a problem hiding this comment.
Could this precondition be removed? Then change then line below to shouldSatisfy (if char1 == char2 then isRight else isLeft)
There was a problem hiding this comment.
@rvl wrote:
Could this precondition be removed? Then change then line below to
shouldSatisfy (if char1 == char2 then isRight else isLeft)
Actually, we do want this precondition, as we're only interested in the cases where corruption actually does happen. By making corruption a precondition, we guarantee to produce cases that we're interested in (or else QuickCheck will give up), where decoding is presumed to fail.
As it turns out, only a small proportion of tests (those that fail to produce a corrupted string) are discarded here.
Though perhaps you have another reason in mind for removing the precondition?
| , " original string: " <> show originalString | ||
| , " corrupted string: " <> show corruptedString ] | ||
| return $ counterexample description $ | ||
| corruptedString /= originalString ==> |
There was a problem hiding this comment.
This precondition could be moved to the shouldBe easily enough.
There was a problem hiding this comment.
@rvl wrote:
This precondition could be moved to the
shouldBeeasily enough.
I think this probably comes under the same discussion as #327 (comment)
4084: Backporting of proper error kernel-to-API mapping during batch addresses import r=KtorZ a=paweljakubas ## Description <!--- A brief description of this PR and the problem is trying to solve --> This is the last bit of backporting of batch import of addresses through APIs. Here, proper mapping of kernel addresses to API errors is enforced ## Linked issue <!--- Put here the relevant issue from YouTrack --> Part of https://iohk.myjetbrains.com/youtrack/issue/CO-447 , namely cardano-foundation/cardano-wallet#327 Co-authored-by: Pawel Jakubas <[email protected]>
Issue Number
Issue #324
Overview
These tests:
Sample string:
Debugging information is only printed when QuickCheck finds a counterexample.