Fix bug in prefix DB indexing

[loiclec]

Where the batch's information was not properly updated in cases where only the proximity changed between two consecutive word pair proximities.

Closes partially https://github.com/meilisearch/meilisearch/issues/3043

[ryanrussell]

@loiclec

Thanks for submitting this fix crazy-fast!

Just for semantics, the test function should be test_batch_bug_3034() -> test_batch_bug_3043() to make it easier to trace back...

[loiclec]

@ryanrussell You're right! Thanks for spotting it :) And thanks again for reporting those two bugs in the release candidate. It helped a lot!

[curquiza]

@loiclec since this PR is in milli, it does not close the Meilisearch issue, can you remove it from the "Development" section on the right?

[Kerollmops]

Hum looks like there is an issue with the heed version we use here. It should not be the lmdb-master-3 branch of heed here, @irevoire can you help us? We should come back to the v0.12.3 IIRC.

[irevoire]

Actually, we should use the v0.12.4 of heed.

Rebasing on main should be enough to fix the issue since we've merged #699

[loiclec]

Thanks everyone :)

bors merge

[bors[bot]]

Build succeeded:

• Run Clippy
• Run Rustfmt
• Tests on macos-latest
• Tests on ubuntu-20.04
• Tests on windows-latest