The fee calculation (also for Babbage transactions) changed in Conway era / protocol version 9 to include costs proportional to reference script size. Our internal wallet seems to not incorporate that or is slightly off.
Weirdly, this was not captured by our end-to-end test that opens a head in babbage, forks into conway and closes a head again?
Could be related to the cost model; does that change when the eras are forked in our tests?
Context & versions
Hydra commit 209de1dd8c5ae484a45a4db3af043c4a9d271306, but any version should have this issue.
Trying to close the head we had open on
preview
, opened inBabbage
and now to close inConway
.Steps to reproduce
Babbage
Conway
Actual behavior
Failed to post tx:
https://github.com/user-attachments/assets/34cae02e-9207-4834-b76c-c766f7698731
Expected behavior
Hypothesis
The fee calculation (also for Babbage transactions) changed in Conway era / protocol version 9 to include costs proportional to reference script size. Our internal wallet seems to not incorporate that or is slightly off.
Weirdly, this was not captured by our end-to-end test that opens a head in babbage, forks into conway and closes a head again?
Could be related to the cost model; does that change when the eras are forked in our tests?