[ 
https://issues.apache.org/jira/browse/SOLR-8475?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15076606#comment-15076606
 ] 

Shai Erera commented on SOLR-8475:
----------------------------------

bq. Why was "log" changed to "LOG"?

I think that static fields should be uppercase. I've done the same in 
SOLR-8454. Is it a problem?

bq. Or that your first refactor patch caused a bug?

The refactoring had a bug, and I fixed it in the second patch.

bq. I assume this patch should be for trunk only?

No, I intend to commit it to 5x too, as I've noted in the "Fix Versions". Why 
would it be trunk only?

bq. I notice a lot of changes like this:

Hmm .. I'm not sure why. I always run "ant eclipse" as that applies our 
formatting rules too, but for some reason, when I hit ENTER inside a file, new 
lines are automatically indented. I looked inside 
{{dev-tooles/eclipse/dot.settings/org.eclipse.jdt.core/prefs: 
org.eclipse.jdt.core.formatter.indent_empty_lines=true}} -- this could explain 
it? I changed it to false and now it doesn't do that anymore.

I agree about the re-wrapped lines. I found it annoying too. But again, I 
thought this is just the standard of our formatting rules. E.g. if I change 
{{org.eclipse.jdt.core.formatter.join_wrapped_lines}} to false, it doesn't do 
that.

I changed both of these and will include them in a new patch, which also 
reverts the {{.set()}} wrapping and leading whitespace in new lines.

> Some refactoring to SolrIndexSearcher
> -------------------------------------
>
>                 Key: SOLR-8475
>                 URL: https://issues.apache.org/jira/browse/SOLR-8475
>             Project: Solr
>          Issue Type: Improvement
>          Components: search
>            Reporter: Shai Erera
>            Assignee: Shai Erera
>            Priority: Minor
>             Fix For: 5.5, Trunk
>
>         Attachments: SOLR-8475.patch, SOLR-8475.patch, SOLR-8475.patch
>
>
> While reviewing {{SolrIndexSearcher}}, I started to correct a thing here and 
> there, and eventually it led to these changes:
> * Moving {{QueryCommand}} and {{QueryResult}} to their own classes.
> * Moving FilterImpl into a private static class (was package-private and 
> defined in the same .java file, but separate class).
> * Some code formatting, imports organizing and minor log changes.
> * Removed fieldNames (handled the TODO in the code)
> * Got rid of usage of deprecated classes such as {{LegacyNumericUtils}} and 
> {{Legacy-*-Field}}.
> I wish we'd cut down the size of this file much more (it's 2500 lines now), 
> but I've decided to stop here so that the patch is manageable. I would like 
> to explore further refactorings afterwards, e.g. extracting cache management 
> code to an outer class (but keep {{SolrIndexSearcher}}'s API the same, if 
> possible).
> If you have additional ideas of more cleanups / simplifications, I'd be glad 
> to do them.



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

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org
For additional commands, e-mail: dev-h...@lucene.apache.org

Reply via email to