…loading (#47037)

before

<img width="1360" height="564" alt="Screenshot 2026-06-17 at 11 30 45"
src="https://github.com/user-attachments/assets/69516e92-0c3d-42ad-b893-2d6ffdf9e9f8"
/>

after

<img width="1351" height="596" alt="Screenshot 2026-06-17 at 11 31 29"
src="https://github.com/user-attachments/assets/8c78e4c5-f46b-4419-87bd-d689982c3f5b"
/>


<!-- This is an auto-generated comment: release notes by coderabbit.ai
-->

## Summary by CodeRabbit

* **Style**
* Refined the visual presentation of the "Open positions" call-to-action
button on the careers page with improved spacing and layout alignment.

<!-- end of auto-generated comment: release notes by coderabbit.ai -->

…-138 (#47026)

## Problem

The unified log dashboards' Postgres detail panel hid the `query` and
`detail` fields. They were present in the raw log message but never
surfaced in the structured view, making them harder to use when
debugging.

## Fix

- Select `pgl_parsed.query` and `pgl_parsed.detail` in the Postgres
service flow query.
- Add `Query` and `Details` field configs to the Postgres primary
fields, both with `wrap: true` so long values display in full instead of
truncating.

The parsed Postgres field is `detail` (singular); it is labeled
"Details" in the UI.

## How to test

- Open Studio and navigate to the unified logs dashboard for a project.
- Filter to Postgres logs and select a log row to open the detail panel.
- Confirm the Postgres section now shows `Query` and `Details` rows
below `User`.
- Expected result: rows render the parsed query and detail text,
wrapping for long values, and show an em dash when empty.

<!-- This is an auto-generated comment: release notes by coderabbit.ai
-->
## Summary by CodeRabbit

* **New Features**
* PostgreSQL service flow logs now show additional always-visible
**Query** and **Details** fields, bringing parsed database query content
and expanded information directly into the log view.
* **Tests**
* Updated log inspection coverage to ensure the new parsed fields are
correctly surfaced in the flattened inspection output.
<!-- end of auto-generated comment: release notes by coderabbit.ai -->

---------

Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

## Problem

It's hard to find a function that references another database entity:
users have to open each of them and look for matches themselves.

## Solution

Add a search input dedicated to function content filtering. Reusing the
existing input to match both names and content may be worse than before
as it would match too many functions if some of them have common sql
keywords in their name.

## Screenshots

<img width="2908" height="672" alt="image"
src="https://github.com/user-attachments/assets/38e35512-d733-434e-8b44-6ff043c01c7e"
/>

<img width="2904" height="560" alt="image"
src="https://github.com/user-attachments/assets/36643865-a1c8-4943-8f13-00272e44eea1"
/>


<!-- This is an auto-generated comment: release notes by coderabbit.ai
-->
## Summary by CodeRabbit

* **New Features**
* Search now performs fuzzy matching over function names and bodies,
ranks exact-name matches higher, and respects
schema/return-type/security filters via centralized filtering logic.
* **Style / UI**
  * Search input placeholder updated to "Search for a function by name".
* **Documentation / Messaging**
* Empty-state messaging clarified to distinguish no functions vs. no
search matches.
* **Tests**
  * Added tests covering the new filtering and ranking behavior.
* **Chores**
  * Added runtime dependency for fuzzy-search library.
<!-- end of auto-generated comment: release notes by coderabbit.ai -->

## I have read the
[CONTRIBUTING.md](https://github.com/supabase/supabase/blob/master/CONTRIBUTING.md)
file.

YES/NO

## What kind of change does this PR introduce?

Bug fix, feature, docs update, ...

## What is the current behavior?

Please link any relevant issues here.

## What is the new behavior?

Feel free to include screenshots if it includes visual changes.

## Additional context

Add any other context or screenshots.


<!-- This is an auto-generated comment: release notes by coderabbit.ai
-->
## Summary by CodeRabbit

## Release Notes

* **New Features**
* Added “Continue with {provider}” sign-in and sign-up flows using
enabled external identity providers.
* Enabled inbound branding to focus a specific provider for customized
sign-in/sign-up experiences.

* **Improvements**
* Refined the sign-in options layout and “last used” tracking for
clearer authentication choices.
* Updated account identity/provider connection experiences (link/unlink
and management UI).

* **Bug Fixes**
  * Fixed hydration mismatches in sign-in and password-related layouts.
<!-- end of auto-generated comment: release notes by coderabbit.ai -->

---------

Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
Co-authored-by: Joshen Lim <joshenlimek@gmail.com>

## Context

Tiny UI nudge to just use the full "Continue with ___" text in the
`SignInWithExternalProvider` UI (can repro by adding a `?method=` param
in the URL on the sign in page)

### Before
<img width="520" height="692" alt="image"
src="https://github.com/user-attachments/assets/f120ade1-8b82-459d-9d82-99027a86bdfd"
/>


### After
<img width="469" height="669" alt="image"
src="https://github.com/user-attachments/assets/7076132a-dee5-40d0-a261-d357808b92b0"
/>


<!-- This is an auto-generated comment: release notes by coderabbit.ai
-->
## Summary by CodeRabbit

* **Improvements**
* Standardized the external authentication provider button text across
sign-in and sign-up flows. The primary provider option now consistently
displays **“Continue with [provider name]”**, including the
focused-provider paths, ensuring uniform labeling instead of relying on
caller-provided overrides.
<!-- end of auto-generated comment: release notes by coderabbit.ai -->

## Details of change

Re-lands DATAENG-1468 (docs page feedback to Postgres) with an
**insert-only** design that avoids the cross-project auth issue that
caused the prior revert.

- New insert-only `feedback_comments` table: anon `insert` policy only
(no select/update/delete). Columns: `page`, `vote`, `title`, `comment`,
`user_id`, `metadata`.
- The docs widget writes the free-text comment to `feedback_comments`
via the **anon key**. The votes `feedback` table is untouched (one row
per vote).
- No user token is sent to the content project anymore (that was the
cause of the previous failure): the feedback client uses the anon key
only.
- The commenter's account user id (gotrue UUID) is read client-side from
the session and stored as a plain `user_id` column for attribution
(comments are logged-in-only). Org/project association is derived
downstream in BigQuery via that id; docs pages aren't project-scoped, so
there's no project_ref/org to capture here.
- Removed the previous update-by-id approach, the per-user RLS policies,
and the obsolete unit test.

## Why the previous version was reverted

It authenticated feedback writes with the supabase.com account session
token, but the requests target the docs content project
(`xguihxuzqibwxjnimxev`), which cannot verify that token. Logged-in
users got `PGRST301 / JWSInvalidSignature`. This version removes the
user token entirely, so writes succeed for everyone.

## Verification

Insert-only RLS means a row can be written but not read/updated/deleted
by `anon`. Comments retrievable with `where comment is not null` is not
needed (separate table); just query `feedback_comments`.


<!-- This is an auto-generated comment: release notes by coderabbit.ai
-->
## Summary by CodeRabbit

* **New Features**
* The feedback form now captures a vote rating along with an optional
title and detailed comments, saving richer context for review.
* **Refactor**
* Feedback submission has been streamlined to write directly to the
database for both vote and comment submissions.
* **Maintenance**
* Updated the feedback data typings to support the new title, comment,
user, page, and vote fields via the new feedback comments storage.
<!-- end of auto-generated comment: release notes by coderabbit.ai -->

## Context

If a user switches account without an explicit log out via the
dashboard, landing back on `/org` will redirect users to the last
visited organization as stored in local storage, in which it can result
in the following state if the last visited organization does not exist
in the current account
<img width="2538" height="1060" alt="image"
src="https://github.com/user-attachments/assets/270e482a-3515-48ef-898b-87e76fce80d6"
/>

## Changes involved
Am opting to scope the last visited organization to the user profile
instead - this would be a bit more cleaner than trying to actively clear
the last visited org slug from local storage with implicit account
changes as there's no deterministic way to track that (afaik) from FE
side of things

## To test
Can reproduce the problem as such
- Ensure that you have 2 accounts to log in with, and one account has an
org that the other is not a part of
- For the organization that has the "extra" org, ensure that you click
into it so that the last visited org slug is saved in local storage
- Mimic changing accounts by visiting
`/auth/v1/authorize?provider=github` (using the domain for the env that
you're testing on - e.g localhost:8000 for local, or green for staging
preview)

<!-- This is an auto-generated comment: release notes by coderabbit.ai
-->
## Summary by CodeRabbit

* **New Features**
* Unified “last visited organization” handling across the Studio UI with
a shared hook, improving consistency for home/dashboard return,
cancel/back navigation, and account routing.

* **Bug Fixes**
* Updated redirects to only route to an organization when a valid
last-visited value is available; otherwise users go to the general
organizations page.
* Kept MFA enrollment and factor delete/leave flows aligned to the
unified last-visited organization value.

* **Tests**
* Updated onboarding and layout tests to match the new last-visited
organization storage key format and hook/query success behavior.
<!-- end of auto-generated comment: release notes by coderabbit.ai -->

Closes DOCS-994

<img width="784" height="171" alt="Screenshot 2026-06-17 at 3 50 27 PM"
src="https://github.com/user-attachments/assets/4ff4afc4-2bc8-4a71-95d8-655171b317b6"
/>



## I have read the
[CONTRIBUTING.md](https://github.com/supabase/supabase/blob/master/CONTRIBUTING.md)
file.

YES

## Problem

The formatting for `!!! note` breaks out multiple nodes when using `code
notation`.

## Fix

This PR formats the `Annotation` MDX component so that it groups the
components into one paragraph rather than multiple.

It also includes a test file, using the example given in the Linear
issue.

## Tophatting

1. Go to
[Hubspot](https://docs-git-docs-format-note-p-supabase.vercel.app/docs/guides/database/extensions/wrappers/hubspot)
in preview.
2. See that the note is formatted correctly. Compare to production.

<!-- This is an auto-generated comment: release notes by coderabbit.ai
-->
## Summary by CodeRabbit

* **Tests**
* Added a Vitest suite to verify admonition transformation and
round-trip markdown parsing, including inline wrapping and preservation
of multiple indented content blocks.
* **Bug Fixes**
* Improved admonition rendering so inline admonition content is grouped
into a single paragraph and whitespace-only text is ignored.
* When admonitions include multiple sibling blocks, those blocks are
kept as separate paragraph children for consistent output.
<!-- end of auto-generated comment: release notes by coderabbit.ai -->