dweiss commented on pull request #2094: URL: https://github.com/apache/lucene-solr/pull/2094#issuecomment-734138279
The fst test verifies byte order of internal data serializer (reverse vs. normal byte order). Every time I look at this diff I have a gut feeling that it should be possible to do this switch in a more transparent way... Even with all the temporary buffers we write to, etc. IndexInput and IndexOutput extend from DataInput/DataOutput. What if those classes were forced to carry the byte order with them (much like byte buffers do)? Initially this would mean all of DataInput/DataOutput we create would have an explicit big endian. We could then gradually switch these to LE for new codecs (in theory all the paired read/write of more complex types should still work) and eventually prohibit BE in DataInput/DataOutput constructor once all the backward codecs are gone... I am tempted to try out this approach this weekend, just out of curiosity. What do you think, @iverase ? ---------------------------------------------------------------- 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: us...@infra.apache.org --------------------------------------------------------------------- To unsubscribe, e-mail: issues-unsubscr...@lucene.apache.org For additional commands, e-mail: issues-h...@lucene.apache.org
