-
Notifications
You must be signed in to change notification settings - Fork 81
Fix typo initial candidates computation #737
Conversation
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Nice, from what I understand it works.
I looked at all the other criterion and it look like some (proximity, geo, arc_desc, attribute, exactness) are doing this;
match initial_candidates {
Some(initial_candidates) => {
self.initial_candidates |= initial_candidates
}
None => self.initial_candidates.map_inplace(|c| c | &candidates),
}
While words is doing what you wrote here.
Is there a reason to chose one method over the other?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Looks amazing to me!
bors merge
@irevoire, |
Poke @curquiza about this fix. |
Is it normal we haven't added any tests here? At least both of the cases reported by the users |
739: Cherry pick bug fixes for v0.37.3 r=curquiza a=curquiza Integrate #734 and #737 into the `release-v0.37.3` branch to avoid to release on `main` <s>Wait for meilisearch/meilisearch#3235 investigation without merging this PR. Indeed, if a bug fix on milli's side, we might wait for it before merging this PR/</s> -> no need to wait, not a bug Co-authored-by: ManyTheFish <many@meilisearch.com> Co-authored-by: Loïc Lecrenier <loic.lecrenier@me.com>
When
Typo
criterion was after a different criterion thanWords
and the previous criterion wasn't returning any candidates at the first iteration of the bucket sort, then theinitial_candidates
were lost.Now,
Typo
ensure to keep theinitial_candidates
between iterations.related to meilisearch/meilisearch#3200 (comment)
related to meilisearch/meilisearch#3228