Debug help for failed Plutus scripts #2430
Conversation
This adds the hex encoding of the plutus inputs to the predicate failure associated with failed scripts. It also adds a function 'debugPlutus' which can be used to re-run a Plutus script in verbose mode.
JaredCorduan
changed the title
|
@michaelpj I've removed your review since I've taken it out of draft and changed it a lot. Don't feel obligated to review it again (unless you want to), I can get a review from the usual suspects. |
| data TagMismatchDescription | ||
| = PassedUnexpectedly | ||
| | FailedUnexpectedly [FailureDescription] | ||
| deriving (Show, Eq, Generic, NoThunks) |
There was a problem hiding this comment.
Deriving NoThunks here is a bit weird, since it's not obvious that we will have no thunks. Are we forcing the descriptions during construction?
| data ScriptResult = Passes | Fails ![String] | ||
| data FailureDescription | ||
| = OnePhaseFailure Text | ||
| | PlutusFailure Text ByteString |
There was a problem hiding this comment.
do you mean add some comments? that seems a good idea to me, thanks
| dec 1 = SumD PlutusFailure <! From <! From | ||
| dec n = Invalid n | ||
|
|
||
| data ScriptResult = Passes | Fails ![FailureDescription] |
There was a problem hiding this comment.
This bang will only force the head of the list...
| "The third data argument, does not decode to a context\n" ++ show info | ||
| ] | ||
| db = B64.encode . serialize' $ PlutusDebug cm eu scriptbytestring ds PlutusV1 | ||
| Just info2 -> Fails [PlutusFailure line db] |
There was a problem hiding this comment.
line and db will be thunks here...
| [ "\nThe 2 arg plutus script (" ++ name ++ ") fails.", | ||
| show e, | ||
| "The redeemer is: " ++ show redeemer, | ||
| "The second data argument, does not decode to a context\n" ++ show info |
There was a problem hiding this comment.
Is the textual information here now redundant with the debug information?
There was a problem hiding this comment.
yep, it is. maybe the human readable part should be whittled down, good point, thanks.
| -- A three data argument script. | ||
| let ss :: Script era | ||
| ss = PlutusScript scriptbytestring | ||
| name :: String | ||
| name = show ss | ||
| in case P.fromData info of | ||
| Nothing -> Fails [line] | ||
| Nothing -> Fails [PlutusFailure line db] |
There was a problem hiding this comment.
also I forget, but could we put the structured type in here and just print it specially later?
There was a problem hiding this comment.
I think the train has left the station, sorry. But as for you question, I don't know, we would have to write our own show for it to produce the base64 in the logs messages, which is maybe a weird/not-always-what-you-want show instance?
There was a problem hiding this comment.
You're using Show for log messages? I'm not totally sure that's a good idea, but if that's the situation then okay.
There was a problem hiding this comment.
I'm not actually sure, to be honest.
This adds all the necessary information for re-runing a failed plutus script inside the
ValidationTagMismatchpredicate failure.Additionally, I've replaced the
[Text]field inValidationTagMismatchwith descriptive types for all the different scenarios. Previously, theEqinstance forValidationTagMismatchwas ignoring this field, I suspect to make testing easier. I've made theEqinstance care about all the fields, and gave the tests a way to ignore the inner details of the plutus error message.