rmuir commented on pull request #1882:
URL: https://github.com/apache/lucene-solr/pull/1882#issuecomment-694339576
+1
We can always try to improve this thing later in subsequent issues. This
still keeps a hard safety net on the number of dirty chunks, so I can't see it
introducing a
rmuir commented on pull request #1882:
URL: https://github.com/apache/lucene-solr/pull/1882#issuecomment-694177068
we may want to apply new logic to the CompressingTermVectorsWriter as well
for consistency?
This is an automa