[jira] [Comment Edited] (IGNITE-6804) Print a warning if HashMap is passed into bulk update operations

["Pavel Tupitsyn (Jira)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/IGNITE-6804?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16952260#comment-16952260 ] 

Pavel Tupitsyn edited comment on IGNITE-6804 at 10/15/19 8:14 PM:
------------------------------------------------------------------

[~ilyak] if you add items to a new empty hashtable, the resulting order is going to be the same every time. And, I guess, this is the most common scenario (e.g. I have a list of Persons, I want to put them all to cache - I have to create and populate a Map).
This also indicates a problem with the API. We waste CPU time populating HashMap or SortedMap, when Ignite only iterates over the result, throwing away all the hash-related stuff.

Anyway, you can keep "will" if you like, but it is better to be technically correct (the best kind of correct).


was (Author: ptupitsyn):
[~ilyak] if you add items to a new empty hashtable, the resulting order is going to be the same every time. And, I guess, this is the most common scenario.
This also indicates a problem with the API. We waste CPU time populating HashMap or SortedMap, when Ignite only iterates over the result, throwing away all the hash-related stuff.

Anyway, you can keep "will" if you like, but it is better to be technically correct (the best kind of correct).

> Print a warning if HashMap is passed into bulk update operations
> ----------------------------------------------------------------
>
>                 Key: IGNITE-6804
>                 URL: https://issues.apache.org/jira/browse/IGNITE-6804
>             Project: Ignite
>          Issue Type: Improvement
>          Components: cache
>            Reporter: Denis A. Magda
>            Assignee: Ilya Kasnacheev
>            Priority: Critical
>              Labels: usability
>          Time Spent: 10m
>  Remaining Estimate: 0h
>
> Ignite newcomers tend to stumble on deadlocks simply because the keys are passed in an unordered HashMap. Propose to do the following:
> * update bulk operations Java docs.
> * print out a warning if not SortedMap (e.g. HashMap, Weak/Identity/Concurrent/Linked HashMap etc) is passed into
> a bulk method (instead of SortedMap) and contains more than 1 element. However, we should make sure that we only print that warning once and not every time the API is called.
> * do not produce warning for explicit optimistic transactions
> More details are here:
> http://apache-ignite-developers.2346864.n4.nabble.com/Re-Ignite-2-0-0-GridUnsafe-unmonitor-td23706.html



--
This message was sent by Atlassian Jira
(v8.3.4#803005)