Hey Uwe, This problem would not be observed by bwc tests because they always make sure to use the right codec. The problem only arises if things get mixed up for some reason, and the wrong codec gets used to read a segment. Hopefully no user would ever see this problem in practice.
The main reason why I'm suggesting respinning is because digging corruption issues is no fun at all, and I'd rather not introduce a case when Lucene would report a corruption with matching checksums instead of a codec header mismatch. Ideally we would have tests that make sure that different file formats always put different codec names in their index headers, but I don't know if there is a reasonable way we could test this. On Mon, Sep 26, 2022 at 12:23 PM Uwe Schindler <u...@thetaphi.de> wrote: > Hi again, > > About the original bug you reported: This is really serious and we MUST > respin! > > That opens the following question: Why was this bug not discoveredby the > backwards compatibility tests? If you open a 9.0 index from backwards > indexes with 9.4 it should complain? Or does this pass fine, because the > segments file opens the correctly. Nevertheless, we should have a check > that also validates all Codec Identifiers matching the class names on > opening an index? > > Uwe > Am 26.09.2022 um 12:01 schrieb Uwe Schindler: > > Hi, > > cool. Could we please wait with the respon till Adoption released also > OpenJDK 19, so we can include the Java 19 MMapDirectory as discussed > before. Everything is ready to merge! > > Uwe > Am 26.09.2022 um 11:42 schrieb Adrien Grand: > > Ignacio made me notice than the new Lucene94FieldInfosFormat did not > change the codec name compared to Lucene90FieldInfosFormat: > https://github.com/apache/lucene/blob/branch_9_4/lucene/core/src/java/org/apache/lucene/codecs/lucene94/Lucene94FieldInfosFormat.java#L376. > > > The issue with this is that if you mistakenly open a Lucene90FieldInfos > file with the 9.4 codec, then it will tell you that there is a corruption > despite the fact that checksums match, instead of pointing out the codec > mismatch. > > I suggest that we respin to address this problem. > > On Sun, Sep 25, 2022 at 8:24 PM Anshum Gupta <ans...@anshumgupta.net> > wrote: > >> +1 (binding) >> >> The smoke tests passed. >> SUCCESS! [2:49:17.242508] >> >> On Fri, Sep 23, 2022 at 12:22 PM Patrick Zhai <zhai7...@gmail.com> wrote: >> >>> (non-binding) +1 >>> >>> SUCCESS! [1:11:00.934249] >>> >>> On Fri, Sep 23, 2022 at 9:44 AM Vigya Sharma <vigya.w...@gmail.com> >>> wrote: >>> >>>> The smoke tests passed for me too.. >>>> >>>> >>>> (no vote) >>>> >>>> SUCCESS! [1:12:31.588303] >>>> >>>> On Thu, Sep 22, 2022 at 2:27 AM Ignacio Vera <iver...@gmail.com> wrote: >>>> >>>>> +1 >>>>> >>>>> >>>>> SUCCESS! [0:46:00.508949] >>>>> >>>>> On Thu, Sep 22, 2022 at 9:53 AM Adrien Grand <jpou...@gmail.com> >>>>> wrote: >>>>> >>>>>> +1 SUCCESS! [0:45:35.275017] >>>>>> >>>>>> On Wed, Sep 21, 2022 at 9:19 PM Michael McCandless < >>>>>> luc...@mikemccandless.com> wrote: >>>>>> >>>>>>> +1 >>>>>>> >>>>>>> >>>>>>> SUCCESS! [0:27:16.514391] >>>>>>> >>>>>>> >>>>>>> Mike McCandless >>>>>>> >>>>>>> http://blog.mikemccandless.com >>>>>>> >>>>>>> >>>>>>> On Wed, Sep 21, 2022 at 2:32 PM Dawid Weiss <dawid.we...@gmail.com> >>>>>>> wrote: >>>>>>> >>>>>>>> >>>>>>>> +1. >>>>>>>> SUCCESS! [0:53:33.891603] >>>>>>>> >>>>>>>> >>>>>>>>> Ran the smoketester with both java 11 and 17: >>>>>>>>> >>>>>>>>> SUCCESS! [2:41:19.024193] >>>>>>>>> >>>>>>>>> On Tue, Sep 20, 2022 at 10:10 PM Michael Sokolov < >>>>>>>>> msoko...@gmail.com> wrote: >>>>>>>>> > >>>>>>>>> > Please vote for release candidate 1 for Lucene 9.4.0 >>>>>>>>> > >>>>>>>>> > The artifacts can be downloaded from: >>>>>>>>> > >>>>>>>>> https://dist.apache.org/repos/dist/dev/lucene/lucene-9.4.0-RC1-rev-f5d0646daa5651f2192282ac85551bca667e34f9 >>>>>>>>> > >>>>>>>>> > You can run the smoke tester directly with this command: >>>>>>>>> > >>>>>>>>> > python3 -u dev-tools/scripts/smokeTestRelease.py \ >>>>>>>>> > >>>>>>>>> https://dist.apache.org/repos/dist/dev/lucene/lucene-9.4.0-RC1-rev-f5d0646daa5651f2192282ac85551bca667e34f9 >>>>>>>>> > >>>>>>>>> > The vote will be open for at least 72 hours i.e. until >>>>>>>>> 2022-09-24 02:00 UTC. >>>>>>>>> > >>>>>>>>> > [ ] +1 approve >>>>>>>>> > [ ] +0 no opinion >>>>>>>>> > [ ] -1 disapprove (and reason why) >>>>>>>>> > >>>>>>>>> > Here is my +1 >>>>>>>>> > >>>>>>>>> > >>>>>>>>> --------------------------------------------------------------------- >>>>>>>>> > To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org >>>>>>>>> > For additional commands, e-mail: dev-h...@lucene.apache.org >>>>>>>>> > >>>>>>>>> >>>>>>>>> >>>>>>>>> --------------------------------------------------------------------- >>>>>>>>> To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org >>>>>>>>> For additional commands, e-mail: dev-h...@lucene.apache.org >>>>>>>>> >>>>>>>>> >>>>>> >>>>>> -- >>>>>> Adrien >>>>>> >>>>> >>>> >>>> -- >>>> - Vigya >>>> >>> >> >> -- >> Anshum Gupta >> > > > -- > Adrien > > -- > Uwe Schindler > Achterdiek 19, D-28357 Bremenhttps://www.thetaphi.de > eMail: u...@thetaphi.de > > -- > Uwe Schindler > Achterdiek 19, D-28357 Bremenhttps://www.thetaphi.de > eMail: u...@thetaphi.de > > -- Adrien
