Fix transaction serialization #941
Conversation
| @@ -41,6 +41,18 @@ object CollectionUtil { | |||
| result | |||
| } | |||
|
|
|||
| /** Concatenate all the arrays in the sequence. */ | |||
| def concatArrays[T: ClassTag](seq: Traversable[Array[T]]): Array[T] = { | |||
There was a problem hiding this comment.
Adding here method for one test just is overkill, for tests ++ is okay
|
|
||
| // TODO v6.0: enforce ordering e.g. using the code below (see https://github.com/ScorexFoundation/sigmastate-interpreter/issues/681) | ||
| // if (VersionContext.current.activatedVersion > JitActivationVersion) { |
There was a problem hiding this comment.
Such Versioning does not work here, as old clients will be broken after activation. Remove misleading TODO
| values <- { | ||
| // TODO v6.0: enforce ordering e.g. using the code below (see https://github.com/ScorexFoundation/sigmastate-interpreter/issues/681) | ||
| // if (VersionContext.current.activatedVersion > JitActivationVersion) { | ||
| // ContextExtension(mutable.LinkedHashMap(values: _*)) |
|
@kushti what should I do with failing serialization roundtrip tests? |
I think an issue, to start with |
There is an issue #681 |
I guess vice-versa , so you should refer to the tests from the issue, which is about way different level of visibility. The issue is far from ideal , as talks there are mostly about JSON serialization |
Actually, the problem is not just with binary, but also with Json, and issue highlights these complexity. |
This PR fixes the tx serialization regression bug.