session-frame-corrupt
[wizzard] frame <n> is not a wizard state. The recording is truncated or corrupt, and a replay woulddraw a state the engine never produced. Record the run again, or restore the recording from where itwas saved. …/errors/session-frame-corrupt[wizzard] visited does not contain the current step "<step>". Breadcrumbs are coloured from visited, sothis frame and every one after it would be drawn wrong. Record the run again, or restore therecording from where it was saved. …/errors/session-frame-corruptReturned by checkSession from @wizzard-packages/core/session, once per frame, with path naming the frame and the field. The site’s inspector reads it to decide whether a recording may be replayed, and the devtools recorder runs it on an export whose redact hook changed the frames.
Every frame of a recording is a WizardState the engine committed. This one is not, in one of two
ways, and path says which:
path |
What is wrong |
|---|---|
frames[<n>] |
the frame is not a wizard state at all - truncated, or not an object |
frames[<n>].visited |
the current step is missing from visited |
Neither can come out of a running wizard, so the recording was damaged after it was made: cut
short on its way to disk, edited by hand, or rewritten by a redact hook that removed more than
data. A replay of such a frame would draw a state that never happened - breadcrumbs are coloured
from visited, so a gap there is wrong in this frame and every one after it.
Record the run again, or restore the recording from where it was saved. If a redact hook produced
it, the hook has to leave stack and visited as it found them.