[
https://issues.apache.org/jira/browse/METRON-1815?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16675302#comment-16675302
]
ASF GitHub Bot commented on METRON-1815:
----------------------------------------
Github user merrimanr commented on the issue:
https://github.com/apache/metron/pull/1249
Following along here. I'm in agreement with @ottobackwards's point about
not having to depend on the module that contains all the parsers that ship with
Metron. So having a separate `metron-parsers` and `metron-parsers-common`
makes sense to me.
I'm not understanding what would go in `metron-parsing-common` though.
Maybe @ottobackwards is correct and there is shared code between storm/spark.
But wouldn't that belong in `metron-parsers`? I feel like the right approach
will come into focus after we take an initial pass at splitting things up. I
propose we start with:
```
└── metron-parsing
├── metron-parsers
├── metron-parsers-common
└── metron-parsing-storm
```
> Separate metron-parsers into metron-parsers-common and metron-parsers-storm
> ---------------------------------------------------------------------------
>
> Key: METRON-1815
> URL: https://issues.apache.org/jira/browse/METRON-1815
> Project: Metron
> Issue Type: Improvement
> Reporter: Justin Leet
> Assignee: Justin Leet
> Priority: Major
>
> In order to expose our parsers to 3rd party components (e.g. the discussions
> on NiFi and potentially other platforms like Spark), we should
> separate the storm-bits into its own project. The metron-parsers-common
> project should contain only parser-oriented code, whereas the
> metron-parsers-storm project should contain the storm specific code
> (e.g. the parser bolt).
>
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
