[
https://issues.apache.org/jira/browse/SOLR-242?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#action_12497089
]
Hoss Man commented on SOLR-242:
-------------------------------
this would go against the current model in Solr, that SolrCore (or the now the
dispatcher) looks at as few options as necessary to hand off to a request
handler, and based onteh result then to an outputwriter ... this would make it
necessary for SolrCore to know about properties specific to a single
OutputWriter ... saying that the presence of "tr" means use the
XsltResponseWriter is like saying that if the "json.nl" param is specified, and
a "wt" is not specified, then assume "wt=json" ... it starts crossing a scary
line. there's also no requirement that clients register the
XsltResponseWriter, and custom request handlers might have a different use for
a param named 'tr"
I'm also not sure i understand your last comment...
"Imagine in the future there's a response writer that outputs a different
kind of XML.
That shouldn't preclude the use of a transform on top of that response."
...true, but you would need differnet XSLTs for differnet base XML responses,
and the current XsltResponseWriter forcably uses the XmlResponseWriter to get
it's input .. if someone wanted to style some other hypothetical future xml
format, then the XsltResponseWriter would need to know about that class --
either hardcoded in it, or more likely as an init param .. in which case you'd
wnat to register multiple instances of hte Writer using different names ... how
would you know which one to use given only the "tr" param?
> tr parameter implies XSL, no wt=xslt necessary
> ----------------------------------------------
>
> Key: SOLR-242
> URL: https://issues.apache.org/jira/browse/SOLR-242
> Project: Solr
> Issue Type: Improvement
> Reporter: Brian Whitman
> Priority: Trivial
>
> Perhaps the most trivial issue ever, but &tr=file.xsl should imply that the
> XML from whichever response writer is being used gets parsed by the given
> transform. The wt=xslt is somewhat redundant. And maybe change the tr
> parameter to xslt.
> Imagine in the future there's a response writer that outputs a different kind
> of XML. That shouldn't preclude the use of a transform on top of that
> response.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
