Re: [JENKINS] Lucene-main-Linux (64bit/hotspot/jdk-17.0.9) - Build # 45858 - Still Unstable!

2023-12-19 Thread Michael McCandless
Oh this is the new (awesome) check Adrien recent added to CheckIndex, so maybe this check is catching some pre-existing bugs in one of our (hopefully experimental, not default!) codecs? Mike McCandless http://blog.mikemccandless.com On Tue, Dec 19, 2023 at 3:43 PM Michael McCandless <

Re: [JENKINS] Lucene-main-Linux (64bit/hotspot/jdk-17.0.9) - Build # 45858 - Still Unstable!

2023-12-19 Thread Michael McCandless
Hmm anyone know why this test suddenly started failing...? Mike McCandless http://blog.mikemccandless.com On Tue, Dec 19, 2023 at 3:35 PM Policeman Jenkins Server < jenk...@thetaphi.de> wrote: > Build: https://jenkins.thetaphi.de/job/Lucene-main-Linux/45858/ > Java: 64bit/hotspot/jdk-17.0.9

Re: [JENKINS] Lucene-main-Linux (64bit/hotspot/jdk-19) - Build # 45856 - Unstable!

2023-12-19 Thread Adrien Grand
This looks like a real bug with the default codec when the prefix compares greater than every indexed term. I'll look into it tomorrow if nobody beats me to it. On Tue, Dec 19, 2023 at 12:35 PM Policeman Jenkins Server < jenk...@thetaphi.de> wrote: > Build:

Re: UTF-8 well-formedness for SimpleTextCodec

2023-12-19 Thread Adrien Grand
Hey Michael, Writing well-formed UTF-8 with SimpleTextformat sounds desirable indeed, e.g. your PR makes sense. I don't think we would want to be heroic about it, but if we can serialize the same information easily, then it sounds like something we should do. Thanks for improving SimpleTextCodec!