[ 
https://issues.apache.org/jira/browse/SOLR-2798?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15096748#comment-15096748
 ] 

Hoss Man commented on SOLR-2798:
--------------------------------

My off the cuff personal opinions w/o reviewing the code and all usages:

* the local param parsing should produce an instance of SolrParams
** given the way the "int" return value is used, i guess that would mean the 
"target" param should be changed to take in a ModifiableSolrParams instance.
* the simplest way to go about this, would probably be:
*# refactor the guts of the current method to implement this new API (correctly)
*# add back a deprecated method with the old signature that delegates to the 
new API and uses the result to populate it's Map<String,String> in a back 
compat way
*# start converting all existing usages of the (now deprecated) old method API 
to the one that uses ModifiableSolrParams

...the key question being: how hard will that last step be?  hopefully not too 
bad.

----

On a tangential note: Demian, i would like to sincerely thank you for sticking 
with this issue --  both your previous offer to help, and your more recent 
proposed solution -- even though you never got any responses.

> Local Param parsing does not support multivalued params
> -------------------------------------------------------
>
>                 Key: SOLR-2798
>                 URL: https://issues.apache.org/jira/browse/SOLR-2798
>             Project: Solr
>          Issue Type: Bug
>            Reporter: Hoss Man
>            Assignee: Anshum Gupta
>
> As noted by Demian on the solr-user mailing list, Local Param parsing seems 
> to use a "last one wins" approach when parsing multivalued params.
> In this example, the value of "111" is completely ignored:
> {code}
> http://localhost:8983/solr/select?debug=query&q={!dismax%20bq=111%20bq=222}foo
> {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]

Reply via email to