~ knotless Search/
↑↓ move↵ open⇧↵ all results

chadtmiller.com/knotless

git that lives in your Atmosphere account

knotless.dev
1 collaborator

knotless 0.13.0 Latest 2 packages, 1 image

1
10

Docs say how a code search orders files scored alike

Merged chadtmiller.com asked to merge into main from chadtmiller.com/knotless:docs/search-order
docs/design/serve.md+5−3
@@ -446,9 +446,11 @@ Zoekt ranks the files it finds but has no cursor, and files it scores
alike come back in whatever order its shards answered. So each page asks
it for the best 200 files, sorts those it scores alike by repository and
path, and answers its own slice of them: a page goes on from the one
before rather than showing a file again or passing one over. The pages
stop 200 files deep: a search that finds more is narrowed rather than
read further. Only the first page asks for the
before rather than showing a file again or passing one over. Only files
tied at the 200th place can differ from one ask to the next, as Zoekt
cuts among them as they come, so only the last page's last files may.
The pages stop 200 files deep: a search that finds more is narrowed
rather than read further. Only the first page asks for the
languages.
### Shared steps
docs/reference/api.md+2−2
@@ -164,8 +164,8 @@ alone.
| `limit` | 20 | the most files answered, clamped to 1 to 50 |
| `cursor` | | from the page before: how many files it held |
Answers `files`, the best first, and those found alike by repository and
path, each with its `repo` (as `listRepos`
Answers `files`, the best first, files scored alike ordered by
repository, then path, each with its `repo` (as `listRepos`
answers one), `path`, `language?`, `commit?` (the commit of the default
branch that was indexed, which its lines are of) and `chunks`: runs of lines holding
matches, each its `startLine`, its `content` and its `ranges`, a match