app: address all to-do's #19
Conversation
- we now dedup existing exits to avoid working them twice - reference obol api implementation doesn't return 400 if there are already threshold partial exits
| return nil, errors.Wrap(err, "ejector exits glob") | ||
| } | ||
|
|
||
| ret := map[eth2p0.ValidatorIndex]struct{}{} |
There was a problem hiding this comment.
Can use make to initialize the map instead
| ret := map[eth2p0.ValidatorIndex]struct{}{} | |
| ret := make(map[eth2p0.ValidatorIndex]struct{}) |
There was a problem hiding this comment.
The only difference I see between the two is that make allocs directly on the heap, while doing it as a literal implies that ret will eventually move to the heap.
While I see the optimization in the sense of "less stuff that the runtime must do", I don't think it's a big deal considering the execution pattern of this function (called once at startup).
WDYT?
| if len(ts.partialExits[exit.PublicKey])+1 > len(lock.Operators) { // we're already at threshold | ||
| writeErr(writer, http.StatusBadRequest, "already at threshold for selected validator") | ||
| return | ||
| if len(ts.partialExits[exit.PublicKey])+1 > len(lock.Operators) { // we're already at threshold, ignore |
There was a problem hiding this comment.
what does this condition imply? in specific, what does len(ts.partialExits[exit.PublicKey]) return?
There was a problem hiding this comment.
ts.partialExits maps a validator's public key to the list of its partial exits, so len(ts.partialExits[exit.PublicKey]) returns the amount of partial exits already stored.
This if is equivalent to the question "if we added another partial exit for validator exit.PublicKey, would its size be greater than the operators amount?".
Consider that we must store at must len(lock.Operators) partial exits for each operator.
Co-authored-by: Abhishek Kumar <43061995+xenowits@users.noreply.github.com>
category: refactor
ticket: none