NightOwl888 commented on issue #419:
URL: https://github.com/apache/lucenenet/issues/419#issuecomment-782869640


   Thanks. I took a close look at these patches.
   
   My first thought was that since most analyzers exist in a vacuum, to upgrade 
hunspell to the latest version of Lucene. However, FST had some major changes 
in 6.x, so that is not an option.
   
   But since there were quite a few hunspell issues patched before 5.x, I am 
proposing upgrading the whole `Lucene.Net.Hunspell` namespace (and tests) to 
version 
[4.10.4](https://github.com/apache/lucene-solr/tree/releases/lucene-solr%2F4.10.4),
 the latest 4.x version. This reduces the risk of incompatible changes and 
gives us a point in time we can debug the code against without too much mental 
anguish, if necessary.
   
   "LUCENE-5778: support hunspell morphological description fields" isn't a big 
deal, but "LUCENE-5864: Split BytesRef into BytesRef and BytesRefBuilder" is a 
major API change to the `Lucene.Net.Util` namespace. That being said, it is not 
a necessary piece of the puzzle - we just need to "downgrade" the following 
classes appropriately.
   
   
   When Lucene 4.10 Uses  |  Replace With
   -- | --
   `BytesRefBuilder`  |  `BytesRef`
   `CharsRefBuilder`  |  `CharsRef`
   `IntsRefBuilder` | `Int32sRef`
   `LongsRefBuilder` | `Int64sRef`
   
   Thoughts?


----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

For queries about this service, please contact Infrastructure at:
[email protected]


Reply via email to