[
https://issues.apache.org/jira/browse/LUCENE-7277?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Dawid Weiss updated LUCENE-7277:
--------------------------------
Attachment: LUCENE-7277.patch
Updated patch. Tests pass.
This is a rather large patch already, so I didn't want to make it any larger,
leaving some hashCode implementation the way they were implemented (although
some of them look rather odd).
I think we should put this in as soon as possible, it really is a nice cleanup
over various forms and shapes of the hashCode/equals implementation encountered
in the code.
As a separate issue, I'd ban the use of getClass().hashCode() entirely (it is
still used in a number of places) as it changes from execution to execution and
can cause problems with debugging.
> Make Query.hashCode and Query.equals abstract
> ---------------------------------------------
>
> Key: LUCENE-7277
> URL: https://issues.apache.org/jira/browse/LUCENE-7277
> Project: Lucene - Core
> Issue Type: Improvement
> Reporter: Dawid Weiss
> Assignee: Dawid Weiss
> Priority: Trivial
> Attachments: LUCENE-7277-20160518.patch, LUCENE-7277.patch,
> LUCENE-7277.patch
>
>
> Custom subclasses of the Query class have the default implementation of
> hashCode/equals that make all instances of the subclass equal. If somebody
> doesn't know this it can be pretty tricky to debug with IndexSearcher's query
> cache on.
> Is there any rationale for declaring it this way instead of making those
> methods abstract (and enforcing their proper implementation in a subclass)?
> {code}
> public int hashCode() {
> return getClass().hashCode();
> }
> public boolean equals(Object obj) {
> if (obj == null)
> return false;
> return getClass() == obj.getClass();
> }
> {code}
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
