Lucene: Status=Open

CH-Abhinav 2.8 hours ago:	I tried removing the config overrides and relying strictly on forceMerge(1) for those 5 tests, but ... I guess it is because some methods directly access via dir.listAll() files ignoring setUseCompoundFi ... At least that is what I think.
mccullocht 23.9 hours ago:	I think you can remove all of these. I don't think the buffering options are needed, and setMergePolicy(NoMergePolicy) means that ...

3.4 hours ago 9 comments 0 votes 0 watches Pranshu-S, kaivalnp, mccullocht, mikemccand, romseygeek, shubhamvishu, xande

mccullocht 3.4 hours ago:	I merged the data blind changes (thanks @kaivalnp) with some expanded scope: you can choose whether ... We disallow dropping the float vectors for asymmetric quantization schemes (1 and 2 bit docs) ... The next step here is to optionally quantize residuals.
mccullocht 1.1 months ago:	We had a similar discussion about rotation and centering on a PR for TurboQuant. Quantizers often perform better with a Gaussian component distribution -- RaBitQ and TurboQuant ... If you don't _need_ to be data blind, centering is still tremendously helpful.

4.1 hours ago 2 comments 0 votes 0 watches HoustonPutman, github-actions[bot], romseygeek

HoustonPutman 4.1 hours ago:	I agree, when doing this I was really confused why this wasn't built in to the existing query ... But thought I was just missing something. I'll try to get this integrated into the actual sub queries and deprecate this class.
romseygeek 12.1 hours ago:	Having thought about this a bit more..

4.1 hours ago 12 comments 0 votes 0 watches ChrisHegarty, github-actions[bot], jimczi

jimczi 4.1 hours ago:	Done, on both states.
jimczi 4.1 hours ago:	I added this and then took it out again. Every call site passed a null access hint, so it returned the context it was given. A format that can be wrapped has no fallback to apply: whoever wraps it puts the hint on the state, ...

5.7 hours ago 4 comments 0 votes 0 watches Abdellox, kkewwei, mkierus

kkewwei 5.7 hours ago:	@mkierus Thanks for the detailed investigation and reproducer. I’ll pick this up and work on a fix soon.
mkierus 5.9 hours ago:	@kkewwei Thanks for confirming, and for the offer! I don't have enough Lucene internals experience to take this on myself with confidence, so I'd ... If it helps, I'm happy to run any candidate fix against the real-world reproducer from the issue ( ...

6.1 hours ago 25 comments 0 votes 0 watches github-actions[bot], iverase, john-mlika, thecoop

thecoop 6.1 hours ago:	Isn't this type already known?
thecoop 6.2 hours ago:	Is there some way to use/share what the reader does, so there's not two methods producing the same ...

7 hours ago 2 comments 0 votes 0 watches costin, github-actions[bot], mayya-sharipova

costin 7.1 days ago:	Thanks for the review. This is going to be a long answer as my understanding of the algorithm might be incomplete hence ... Now threads "steal" leftover work as (source graph, source ordinal) and look up the merged ordinal ...
mayya-sharipova 11.3 days ago:	@costin Thanks for tackling concurrent merges. Before you invest more here: I'm not convinced by this approach. The sequential algorithm was designed around the join set, which gives two guarantees — coverage ...

2,661 issues

[22.7 msec search, 24.5 msec total]