[hibernate-dev] [HSearch] Upgrade to Lucene 3.0

[Sanne Grinovero]

absolutely, that's why I'm proposing to try loading what the user said
first, and only in case of failure try our patched version.
So the day Solr will have an updated package, our mad loading logic
will become harmless as I suppose the loading will work fine at first
try, and you'll get the nice Solr updated features again.
Still to implement this, you'll have to fork all of them to delegate
to implementations having proper workarounds for the new Lucene API.

Sanne

2010/9/9 Hardy Ferentschik <hibernate at ferentschik.de>:
> On Thu, 09 Sep 2010 14:07:57 +0200, Gustavo Fernandes
> <gustavonalle at gmail.com> wrote:
>
>> I agree with Sanne regarding the hiding of Solr framework behind an
>> adaptor or whatever, even better if it'd possible to provide Solr
>> integration
>
> FYI, we are talking about these classes:
>
> ArabicLetterTokenizerFactory
> ArabicNormalizationFilterFactory
> ArabicStemFilterFactory
> ASCIIFoldingFilterFactory
> BaseCharFilterFactory
> BaseTokenFilterFactory
> BaseTokenizerFactory
> BaseTokenStreamFactory
> BrazilianStemFilterFactory
> BufferedTokenStream
> CapitalizationFilterFactory
> CharFilterFactory
> ChineseFilterFactory
> ChineseTokenizerFactory
> CJKTokenizerFactory
> CollationKeyFilterFactory
> CommonGramsFilter
> CommonGramsFilterFactory
> CommonGramsQueryFilter
> CommonGramsQueryFilterFactory
> DelimitedPayloadTokenFilterFactory
> DictionaryCompoundWordTokenFilterFactory
> DoubleMetaphoneFilter
> DoubleMetaphoneFilterFactory
> DutchStemFilterFactory
> EdgeNGramFilterFactory
> EdgeNGramTokenizerFactory
> ElisionFilterFactory
> EnglishPorterFilterFactory
> FrenchStemFilterFactory
> GermanStemFilterFactory
> GreekLowerCaseFilterFactory
> HTMLStripCharFilter
> HTMLStripCharFilterFactory
> HTMLStripReader
> HTMLStripWhitespaceTokenizerFactory
> HyphenatedWordsFilter
> HyphenatedWordsFilterFactory
> ISOLatin1AccentFilterFactory
> KeepWordFilter
> KeepWordFilterFactory
> KeywordTokenizerFactory
> LengthFilterFactory
> LetterTokenizerFactory
> LowerCaseFilterFactory
> LowerCaseTokenizerFactory
> MappingCharFilterFactory
> NGramFilterFactory
> NGramTokenizerFactory
> NumericPayloadTokenFilterFactory
> PatternReplaceCharFilter
> PatternReplaceCharFilterFactory
> PatternReplaceFilter
> PatternReplaceFilterFactory
> PatternTokenizer
> PatternTokenizerFactory
> PersianNormalizationFilterFactory
> PhoneticFilter
> PhoneticFilterFactory
> PorterStemFilterFactory
> PositionFilterFactory
> RemoveDuplicatesTokenFilter
> RemoveDuplicatesTokenFilterFactory
> ReversedWildcardFilter
> ReversedWildcardFilterFactory
> ReverseStringFilterFactory
> RussianLetterTokenizerFactory
> RussianLowerCaseFilterFactory
> RussianStemFilterFactory
> ShingleFilterFactory
> SnowballPorterFilterFactory
> SolrAnalyzer
> StandardFilterFactory
> StandardTokenizerFactory
> StopFilterFactory
> SynonymFilter
> SynonymFilterFactory
> SynonymMap
> ThaiWordFilterFactory
> TokenFilterFactory
> TokenizerChain
> TokenizerFactory
> TokenOffsetPayloadTokenFilterFactory
> TrimFilter
> TrimFilterFactory
> TypeAsPayloadTokenFilterFactory
> WhitespaceTokenizerFactory
> WordDelimiterFilter
> WordDelimiterFilterFactory
> WordDelimiterIterator
>
>
> In many cases the factory classes are just wrapping the creation of a
> specific Filter/Tokenizer. Often these Filter/Tokenizer are even provided
> by Lucene. Delegating in these cases seems almost silly.
> More interesting are actual Filter classes provided by Solr which are not
>  from Lucene, for example SynonymFilter or TrimFilter. Here we have actual
> logic in the implementations and one could consider delegation. Here also
> lies the real additional value of the framework and you would like to keep
> in sync with the Solr work instead of forking.
>
>
> --Hardy
> _______________________________________________
> hibernate-dev mailing list
> hibernate-dev at lists.jboss.org
> https://lists.jboss.org/mailman/listinfo/hibernate-dev
>