zeekay ed50a30052 fix(brand): error links point at docs.hanzo.ai, not the upstream domain ...

Every error this engine returns carried a link to the upstream project's docs
domain, so a visitor to search.hanzo.ai was handed another company's URL —
the brand policy forbids referencing upstream at a user-facing surface. Live
today: GET https://search.hanzo.ai/ returns
{"code":"invalid_api_key","link":"https://www.meilisearch.com/..."}.

ONE source builds that link (Code::url in meilisearch-types/src/error.rs); the
snapshots and the OpenAPI examples in routes/tasks.rs assert it literally, so
they are updated to match — that is exactly what regenerating them would
produce, done deterministically instead of by a full test run.

Both spellings are covered: `/errors#name` and `/errors/#name`.

Pure string-literal edits: the format! arity and types are unchanged, so this
cannot alter behaviour beyond the emitted URL.

NOTE: this fork is at 1.45.2 while production runs 1.37.0, so shipping it is a
deliberate version jump, not a side effect of this commit.

2026-07-25 16:38:41 -07:00

..

datasets_paths.rs

fix(brand): error links point at docs.hanzo.ai, not the upstream domain

2026-07-25 16:38:41 -07:00