You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@flink.apache.org by "Ufuk Celebi (JIRA)" <ji...@apache.org> on 2016/02/07 21:11:39 UTC

[jira] [Commented] (FLINK-3357) Drop JobId.toShortString()

    [ https://issues.apache.org/jira/browse/FLINK-3357?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15136415#comment-15136415 ] 

Ufuk Celebi commented on FLINK-3357:
------------------------------------

I agree. We had a discussion about this on the ML once with a consensus to remove it as well.

I've assigned this to myself as I introduced it as afor debugging during the network rewrite.

> Drop JobId.toShortString()
> --------------------------
>
>                 Key: FLINK-3357
>                 URL: https://issues.apache.org/jira/browse/FLINK-3357
>             Project: Flink
>          Issue Type: Bug
>          Components: Core
>    Affects Versions: 0.10.1
>            Reporter: Stephan Ewen
>            Assignee: Ufuk Celebi
>             Fix For: 1.0.0
>
>
> There are several places that use JobId.toShortString() instead of the full JobID, for unique directories, etc.
> I think we should drop that method.
>   - If the short String representation (8 bytes) is statistically "unique enough", then it is wrong in the first place that the unique IDs have 16 bytes
>   - If 16 bytes were chosen with a purpose, then using the short string representation in paths, etc, is clearly violating the guarantees the IDs try to give,
> For purely informational logging, it may be a nice util, bit it seems to encourage wrong use.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)