Is this a bug at check-account in stolon/src/jepsen/stolon/ledger.clj ?

[HaxiSnake]

Hi, I'm learning jepsen on postgres, and I notice a check rule in ledger.clj which confuses me.

`

(defn check-account
"Takes an [account, ops] tuple, and checks to make sure it has a non-negative
balance."
[[account ops]]
; When ops are indeterminate, we don't know if they took place or not. We
; have to take the most charitable interpretation: assume deposits succeed,
; withdrawals fail.
(let [deposits (filter (comp pos? second :value) ops)
withdrawals (filter (comp neg? second :value) ops)
balance (->> (concat deposits (filter op/ok? withdrawals))
(map (comp second :value))
(reduce + 0))]
(when (or (neg? balance) (pos? balance))
{:account account
:balance balance})))

`

This test is aimed at to check each banlance of the account is non-negative, why return result when banlance is pos?

`

(defn checker
[]
(reify checker/Checker
(check [_ test history opts]
(let [errs (->> history
(filter (comp #{:ok :info} :type))
(group-by (comp first :value))
(keep check-account))]
{:valid? (not (seq errs))
:errors errs}))))

`

this may lead to that it will be recognized as errors when checking, Am I reading this correctly？Thanks for your attention.

[aphyr]

Entirely possible this is wrong--just going through the namespace again, this looks like it may have been a work-in-progress I was sketching out but never finished.

[HaxiSnake]

ok, thanks for your reply! But is this ledger case will be finished?

[aphyr]

I have no plans to revisit that particular code right now.On Nov 16, 2022 01:55, Dong ***@***.***> wrote: ok, thanks for your reply! But is this ledger case will be finished? —Reply to this email directly, view it on GitHub, or unsubscribe.You are receiving this because you commented.Message ID: ***@***.***>