Skip to content

Upgrade Lucene #850

Open
Open

Description

@mcserep

We use Lucene 4.9.0 in the search parser, while the current major version is 10.

The goal of this ticket is to analyze that what breaking API changes and dependency changes would be required to upgrade Lucene to a newer version.

Activity

  1. added this to the Upcoming Release milestone on Apr 14, 2026
  2. self-assigned this
    on Apr 14, 2026
  3. LakatosMark1 commented on Apr 14, 2026

    @LakatosMark1
    Collaborator

    Lucene 5.5.X: Csak az importok és a verzió-konstansok cseréjét igényli, nem kell újraírni az adatbázis sémáját, és a régi indexek is megmaradhatnak
    Lucene 6.6.X: Minimum Java 8, a régi számkezelés (NumericField, IntField) végleg megszűnik, az új PointValues és DocValues logikát kell használni
    Lucene 10.X: Java 21: újra kell írni a teljes indexelőt és Query engine-t (DocValues bevezetése, régi lekérdezés-építők tiltása, átszervezett analizátorok)

  4. mcserep commented on Apr 15, 2026

    @mcserep
    CollaboratorAuthor

    @LakatosMark1 Thanks for your results!

    We use the English language on GitHub for the project. Please update your remark accordingly.

  5. LakatosMark1 commented on Apr 15, 2026

    @LakatosMark1
    Collaborator

    Excuse me:
    Lucene 5.5.X: Only requires replacing imports and version constants, no need to rewrite the database schema, and old indexes can be preserved
    Lucene 6.6.X: Minimum Java 8, old number handling (NumericField, IntField) is permanently discontinued, new PointValues ​​and DocValues ​​logic must be used
    Lucene 10.X: Java 21: entire indexer and Query engine must be rewritten (introduction of DocValues, disabling old query builders, reorganized analyzers)

  6. mcserep commented on Apr 28, 2026

    @mcserep
    CollaboratorAuthor

    @LakatosMark1 As discussed today, we shall proceed with smaller steps. Try to upgrade to version 6, as that task shall fit in the timeframe for this semester.

  7. mcserep commented on Apr 28, 2026

    @mcserep
    CollaboratorAuthor

    @LakatosMark1 Additionally, as discussed on today's meeting, please investigate the UTF-8 / UTF-16 parsing issue. As far as I remember, originally it was reported when parsing C# code with such characters. Create a simple, Hello World program in C#, but e.g. use emojis as variable names, as that is permitted in C#. Try to parse it and observer if there are any issues?

    If there is a problem, but not related to the search parser, we will create a separate GitHub ticket for it, but first it should be analyzed in more detail.

  8. LakatosMark1 commented on May 9, 2026

    @LakatosMark1
    Collaborator

    @mcserep I have made a pull request for this issue, it solves the search parser error.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

Type

No type

Projects

  • Status
    To do
  • Status
    To do

Relationships

None yet

Development

No branches or pull requests

Issue actions