PSeitz

@ppodolsky sorry I missed that review. Yes support for such queries is definitely interesting. Could you check the performance impact on `bench_union_3_low` for removing the static dispatch?

> > Did you observe a performance regression / improvement? Did it shave off the memory peaks you observed before during indexing? I didn't see an impact on indexing performance....

Let me explain how it's handled in veloci, which is neat in some parts I think, but I don't know how applicable it would be in tantivy. Before indexing, you...

I think this is a great way of describing a format: https://github.com/mocobeta/lucene-postings-format, well done @mocobeta

Nice, rucene also has a bkd implementation https://github.com/zhihu/rucene/blob/master/src/core/util/bkd/bkd_writer.rs

Yes, it's easy to reproduce, but I don't get the stacktrace anymore. cargo nextest seems to behave more consistent there. ``` NUM_FUNCTIONAL_TEST_ITERATIONS=2000000 cargo test indexing_sorted -- --ignored running 1 test...

Just to confirm my understanding. We would stack all small files in the .segbundle and large files are always on its own?

> @PSeitz I updated the RFC for IP address field. I think the data suggested we do not need to go for the more complicated code to handle high-freq /...

I'll just reserve same space in the codec to be able to handle that later on. @kstaken What's the cardinality of the ip field, single or multi values?

@kstaken We plan to support both and it should also be seamless. The current fast field API requires to define `Cardinality`, but we consider to drop that and detect it...