submit_tx swallows error after upgrading from 0.25.0 to 0.28.0

[jonathangenlambda]

We have upgraded pallas from 0.25.0 to 0.28.0 and now when a TX was rejected by the local node the node_client.submission().submit_tx function seems to not return a corresponding error and seems to treats it as Ok. In the node however we clearly see that the TX got dropped from mempool (in our testcase due to UTxO already spent). This seems to be a regression from 0.25.0 to 0.28.0 (sorry, we didnt do the versions in between).

For additional info: for 0.25.0 we were using our own fork, with very minor changes, which have been added to 0.28.0 by now as far as we can see. These are the changes of our own fork: main...thalerjonathan:pallas:main