[
https://issues.apache.org/jira/browse/IGNITE-15666?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17447457#comment-17447457
]
Vyacheslav Koptilin edited comment on IGNITE-15666 at 12/3/21, 12:20 AM:
-------------------------------------------------------------------------
Hello [~NSAmelchev], [~PetrovMikhail],
It seems to me, this fix is arguable and contradicts the specification JCache
API.
In accordance with specification *12.4. Statistics Effects of Cache Operations*
the `remove(K key)` method should update statistics _if the method returns
true._ Yep, it explicitly mentions the "Number of Removals", however, I think
the average time is assumed as well.
Let's consider the following situation:
- cache.put(key, value); // removals = 0, averageTime = 0
- cache.remove(key); // removals = 1, averageTime = 100ms
- cache.remove(notExistingKey1); // removals = 1, averageTime = 200ms (in your
pull-request, this metric is always updated)
- cache.remove(notExistingKey1); // removals = 1, averageTime = 300ms
(--{-}//{-}--)
In my understanding, this behavior does not seem correct.
Yes, I see that your pull request passed the TCK (as the previous version of
code as well), however, it speaks about the quality of the TCK itself :)
was (Author: slava.koptilin):
Hell [~NSAmelchev], [~PetrovMikhail],
It seems to me, this fix is arguable and contradicts the specification JCache
API.
In accordance with specification *12.4. Statistics Effects of Cache Operations*
the `remove(K key)` method should update statistics _if the method returns
true._ Yep, it explicitly mentions the "Number of Removals", however, I think
the average time is assumed as well.
Let's consider the following situation:
- cache.put(key, value); // removals = 0, averageTime = 0
- cache.remove(key); // removals = 1, averageTime = 100ms
- cache.remove(notExistingKey1); // removals = 1, averageTime = 200ms (in your
pull-request, this metric is always updated)
- cache.remove(notExistingKey1); // removals = 1, averageTime = 300ms
(---//---)
In my understanding, this behavior does not seem correct.
Yes, I see that your pull request passed the TCK (as the previous version of
code as well), however, it speaks about the quality of the TCK itself :)
> The remove metric value is different for sync and async methods
> ---------------------------------------------------------------
>
> Key: IGNITE-15666
> URL: https://issues.apache.org/jira/browse/IGNITE-15666
> Project: Ignite
> Issue Type: Bug
> Reporter: Amelchev Nikita
> Assignee: Amelchev Nikita
> Priority: Major
> Fix For: 2.12
>
> Attachments: screenshot-1.png
>
> Time Spent: 0.5h
> Remaining Estimate: 0h
>
> The remove metric value is different for sync and async methods.
> The following metrics are updated only if the key was exist for the sync
> version:
> {noformat}
> RemoveTimeTotal
> RemoveTime
> {noformat}
--
This message was sent by Atlassian Jira
(v8.20.1#820001)
