You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
There is a virtual_op field in operation_history_object, related code is here. Questions:
what's the supposed use case of this data? I noticed recently it's used in GUI to sort account history, which is introduced as a fix of issue 793.
however, in witness_node, this data is filled with the variable _current_virtual_op in database.hpp with this code, but _current_virtual_op has never been saved to or loaded from disk. So if witness_node is started without --replay-blockchain, _current_virtual_op is reset to 0, then every new operation_history_object saved to database contains a virtual_op value which would be already assigned to an old operation_history_object, thus the logic in above fix is not totally reliable.
Thoughts?
//Edit: logic in the GUI fix is correct, because it compares block_num first, then compares virtual_op.
From @abitmore on March 31, 2016 11:37
There is a
virtual_op
field inoperation_history_object
, related code is here. Questions:_current_virtual_op
indatabase.hpp
with this code, but_current_virtual_op
has never been saved to or loaded from disk. So ifwitness_node
is started without--replay-blockchain
,_current_virtual_op
is reset to 0, then every newoperation_history_object
saved to database contains avirtual_op
value which would be already assigned to an oldoperation_history_object
, thus the logic in above fix is not totally reliable.Thoughts?
//Edit: logic in the GUI fix is correct, because it compares
block_num
first, then comparesvirtual_op
.Copied from original issue: cryptonomex/graphene#644
The text was updated successfully, but these errors were encountered: