Re: QueryParsing.SortSpec

2007-11-20 Thread Chris Hostetter
: While we are at it, we should limit consolidate the null sort checking. See: : http://www.nabble.com/Sorting-problem-tf4762114.html#a13627492 : : Currently for "score desc", the parsing utility returns a null SortSpec, then : has to create a new one to attach rows/offset +1 ... there's no rea

Re: QueryParsing.SortSpec

2007-11-20 Thread Ryan McKinley
since 99% of hte use cases are pulling sort, rows, and start from some SOlrParams, that use case should probably be refactored into a utility method as well. While we are at it, we should limit consolidate the null sort checking. See: http://www.nabble.com/Sorting-problem-tf4762114.html#a

Re: QueryParsing.SortSpec

2007-11-20 Thread Chris Hostetter
: I think it does make sense to keep them together. : offset and length only make sense if an ordering is specified. true. : If we change the name, we should also move it to a top-level class : (from a static inner). : Any suggestions? good point .. promoting it to first order is probably more

Re: QueryParsing.SortSpec

2007-11-19 Thread patrick o'leary
I don't see an issue with retaining it, as long as it's got public constructors. The difference between responseBuilder.setSort(customSort);  // if you deprecate SortSpec. and responseBuilder.setSortSpec (new QueryParser.SortSpec(customSort, -1)); isn't going to kill anyone. P Yonik Seeley

Re: QueryParsing.SortSpec

2007-11-19 Thread Yonik Seeley
On Nov 19, 2007 10:26 PM, Chris Hostetter <[EMAIL PROTECTED]> wrote: > Last time i looked at most usages, the getOffset() and getCount() were > totally ignored Not any more... see QueryComponent.process(): results.docList = searcher.getDocList( builder.getQuery(), builder.getFilte

Re: QueryParsing.SortSpec

2007-11-19 Thread Chris Hostetter
: Hmmm, no reason I see. : Unless there are objections, I'll make them public. -0 ... i'd much rather see us deprecate it and make it go away entirely. it's an archaic throwback to the days when every option was going to be encoded into the "q" param instead of having "sort" "rows" and "start"

Re: QueryParsing.SortSpec

2007-11-19 Thread Yonik Seeley
On Nov 19, 2007 9:01 PM, patrick o'leary <[EMAIL PROTECTED]> wrote: > Is there a reason that QueryParsing.SortSpec has private constructors? > It's preventing me from having a custom sort parser for the ResponseBuilder > without a patch for solr. Hmmm, no reaso

QueryParsing.SortSpec

2007-11-19 Thread patrick o'leary
Hey Is there a reason that QueryParsing.SortSpec has private constructors? It's preventing me from having a custom sort parser for the ResponseBuilder without a patch for solr. Thanks P -- Patrick O'Leary You see, wire telegraph is a kind of a very, very long cat. You pull his t