Repository navigation
chore: production deploy - #7549
Merged
Merged
Conversation
supabase-cli-releaser
Bot
commented
Nov 10, 2026
Contributor
- chore(stack): bump realtime to v2.154.1-r0 (chore(stack): bump realtime to v2.154.1-r0 #7543)
- chore(stack): bump storage to v1.90.2-r0 (chore(stack): bump storage to v1.90.2-r0 #7544)
- chore(api): sync Management API OpenAPI spec (chore(api): sync Management API OpenAPI spec #7541)
- chore(stack): bump realtime to v2.155.0-r0 (chore(stack): bump realtime to v2.155.0-r0 #7547)
- chore(stack): bump realtime to v2.156.0-r0 (chore(stack): bump realtime to v2.156.0-r0 #7548)
This bumps realtime from v2.154.0-r0 to v2.154.1-r0. https://github.com/supabase/slim-services/releases/tag/realtime-v2.154.1-r0 Planned after supabase/slim-services published v2.154.1-r0. This branch is rewritten from develop whenever a newer relevant release arrives. Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
This bumps storage from v1.90.1-r0 to v1.90.2-r0. https://github.com/supabase/slim-services/releases/tag/storage-v1.90.2-r0 Planned after supabase/slim-services published v1.90.2-r0. This branch is rewritten from develop whenever a newer relevant release arrives. Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
This PR was automatically created to sync the generated `@supabase/api` package with the latest Management API OpenAPI document. Changes were detected in the upstream OpenAPI documents exposed by `https://api.supabase.com/api/v1-json` and `https://api.supabase.com/api/v2-json`. --------- Co-authored-by: jgoux <1443499+jgoux@users.noreply.github.com> Co-authored-by: Julien Goux <hi@jgoux.dev>
This bumps realtime from v2.154.1-r0 to v2.155.0-r0. https://github.com/supabase/slim-services/releases/tag/realtime-v2.155.0-r0 Planned after supabase/slim-services published v2.155.0-r0. This branch is rewritten from develop whenever a newer relevant release arrives. Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
This bumps realtime from v2.155.0-r0 to v2.156.0-r0. https://github.com/supabase/slim-services/releases/tag/realtime-v2.156.0-r0 Planned after supabase/slim-services published v2.156.0-r0. This branch is rewritten from develop whenever a newer relevant release arrives. Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
supabase-cli-releaser
Bot
requested review from
a team and
supabase-oss
as code owners
November 10, 2026 02:15
This PR was automatically created to sync the generated `@supabase/api` package with the latest Management API OpenAPI document. Changes were detected in the upstream OpenAPI documents exposed by `https://api.supabase.com/api/v1-json` and `https://api.supabase.com/api/v2-json`. Co-authored-by: jgoux <1443499+jgoux@users.noreply.github.com>
This bumps storage from v1.90.2-r0 to v1.91.0-r0. https://github.com/supabase/slim-services/releases/tag/storage-v1.91.0-r0 Planned after supabase/slim-services published v1.91.0-r0. This branch is rewritten from develop whenever a newer relevant release arrives. Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
…7545) ## TL;DR stops the stack refusing to start when `[storage.analytics]` is enabled or vector buckets are declared ## whats hurting? `supabase start` on the stack exits w `storage.analytics.enabled is unsupported by the experimental stack` or `storage.vector settings are unsupported by the experimental stack`while the legacy backend starts fine with the same config ## now fixed by? dropping both checks, so the stack accepts analytics and seeds vector buckets on first start like the legacy backend the settings the stack still rejects now have their error messages documented > [!NOTE] > `/storage/v1/iceberg` still returns 404 locally, left for a separate decision ## ref: - closes: CLI-2582
This bumps storage from v1.91.0-r0 to v1.92.0-r0. https://github.com/supabase/slim-services/releases/tag/storage-v1.92.0-r0 Planned after supabase/slim-services published v1.92.0-r0. This branch is rewritten from develop whenever a newer relevant release arrives. Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
This bumps storage from v1.92.0-r0 to v1.92.2-r0. https://github.com/supabase/slim-services/releases/tag/storage-v1.92.2-r0 Planned after supabase/slim-services published v1.92.2-r0. This branch is rewritten from develop whenever a newer relevant release arrives. Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
This bumps postgres from 17.11.0.018-r0 to 17.11.0.019-r0. https://github.com/supabase/slim-services/releases/tag/postgres-17.11.0.019-r0 Planned after supabase/slim-services published 17.11.0.019-orioledb-r0. This branch is rewritten from develop whenever a newer relevant release arrives. Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
This adds postgres 17.11.0.019-orioledb-r0 as a new release line. https://github.com/supabase/slim-services/releases/tag/postgres-17.11.0.019-orioledb-r0 Planned after supabase/slim-services published 17.11.0.019-orioledb-r0. This branch is rewritten from develop whenever a newer relevant release arrives. Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
This bumps postgres from 15.19.0.018-r0 to 15.19.0.019-r0. https://github.com/supabase/slim-services/releases/tag/postgres-15.19.0.019-r0 Planned after supabase/slim-services published 17.11.0.019-orioledb-r0. This branch is rewritten from develop whenever a newer relevant release arrives. Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
This bumps storage from v1.92.2-r0 to v1.92.3-r0. https://github.com/supabase/slim-services/releases/tag/storage-v1.92.3-r0 Planned after supabase/slim-services published v1.92.3-r0. This branch is rewritten from develop whenever a newer relevant release arrives. Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
This bumps realtime from v2.156.0-r0 to v2.156.1-r0. https://github.com/supabase/slim-services/releases/tag/realtime-v2.156.1-r0 Planned after supabase/slim-services published v2.156.1-r0. This branch is rewritten from develop whenever a newer relevant release arrives. Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
This bumps storage from v1.92.3-r0 to v1.94.0-r0. https://github.com/supabase/slim-services/releases/tag/storage-v1.94.0-r0 Planned after supabase/slim-services published v1.94.0-r0. This branch is rewritten from develop whenever a newer relevant release arrives. Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
This bumps realtime from v2.156.1-r0 to v2.156.2-r0. https://github.com/supabase/slim-services/releases/tag/realtime-v2.156.2-r0 Planned after supabase/slim-services published v2.156.2-r0. This branch is rewritten from develop whenever a newer relevant release arrives. Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
I moved plan-gate handling out of each command into shared code. Since #7271, `mapHttpError` attaches `entitlement` from the Management API's `entitlement_required` body, so the upgrade hint, the `cli_upgrade_suggested` event and the `plan_limit` classification now all come from that one field, with no per-command wiring and no extra requests. **Changed:** - **One hint and one event per gated denial, wired or not**: the text error renderer prints the upgrade hint before the error message, and `withCommandTelemetry` captures `cli_upgrade_suggested` (same name, same `{feature_key, org_slug}` payload) after the command, just before `cli_command_executed`. Commands that never had wiring, such as `branches unpause`, now get both. `vanity-subdomains check-availability` now fires the event too; its no-event exception only existed for parity with the removed Go CLI. - **No per-command gate code or follow-up lookups**: I deleted the `gateMapError`/`suggestUpgrade` wrapping at `branches create`/`update`, the four `domains` commands and the three `vanity-subdomains` commands. `branches update` no longer fetches the project and org entitlements after a 4xx, since the persistent-branch 402 carries the envelope. Only the SSO commands keep the entitlements lookup. - **`plan_limit` from the envelope**: an error carrying `entitlement` classifies as `plan_limit` whatever its class declares, so gated denials from `branches unpause`, `domains get`/`create`/`activate`/`reverify` and `vanity-subdomains get` move from `api_status` to `plan_limit` (fingerprint suffix `:api_status` becomes `:plan_limit`). **Note:** the event moved from mid-handler to post-run. Payload and identity stay the same; it uses the same capture context as `cli_command_executed`. ## Linear - part of GROWTH-1050
## TL;DR
`stack start` now supports `[api.tls]`, serving https on the api port
with legacy's cert while containers keep plain http over loopback.
## whats introduced?
- one port, both protocols. a tls hello gets https from anyone, plain
http is served only to loopback, and any other peer is closed with
nothing sent back
- host urls move to https (`API_URL`, auth links, studio's public url,
the functions banner), while containers and `pg_net` keep their plain
runtime urls
- legacy's embedded cert by default, or your own `cert_path` and
`key_path`. the cert must cover `127.0.0.1` and is checked before
anything starts, then saved with the stack for `storage` commands
- every start and restart that serves tls on docker or podman ends with
the db container reading a reply only this stack and this start can
give. if containers don't reach the api from loopback, start fails with
`unsupported networking` and stops the stack
- works on docker desktop, colima and podman machine on macos. linux
docker and podman deliver containers from a bridge or host address, so a
tls start fails there
- tls on or off applies on the next `stack start`, and without it
nothing changes
## visual rep:
```mermaid
flowchart LR
host["cli, sdk, browser<br/>on this machine"] --> port{{"api port<br/>reads the first byte"}}
ctr["auth, studio,<br/>functions, db"] -- "via loopback" --> port
lan["lan peer"] --> port
port -- "tls hello" --> https["https<br/>the api, for anyone"]
port -- "anything else" --> who{"from loopback?"}
who -- "no" --> closed["closed<br/>nothing sent back"]
who -- "yes" --> named{"stack Host?"}
named -- "no" --> misdirected["421"]
named -- "yes" --> plain["plain http<br/>the api, /email templates<br/>and the start check"]
https -. "/email" .-> hidden["404"]
classDef t fill:transparent
class host,ctr,lan,port,https,who,closed,named,misdirected,plain,hidden t
```
## ref
- closes: CLI-2584
- supersedes: #6803
## Summary Adds `supabase config convert`, which rewrites `supabase/config.toml` as `supabase/config.jsonc` and removes the TOML file. - **Lossless or nothing.** Values come from the TOML parser. The JSONC output is re-parsed and compared with the source before anything is written; if they differ, the command refuses and leaves `config.toml` untouched. - **Same document.** Keys and `env()` strings are unchanged. `$schema` comes from the `#:schema` directive, or the current schema URL when there is none. - **Comments survive.** Comments are carried to the matching member. Commented-out TOML examples (as in the `init` template) become commented-out JSONC inside the right table, and each one can be uncommented on its own. Any comment that has to move is reported. - **Warnings before writing.** It prints the minimum CLI version that reads `config.jsonc`, with an undo command. `--dry-run` shows the result without writing. Running it on a project that already uses JSON exits 0 and does nothing. - **VS Code.** When `.vscode/settings.json` exists, it trusts the Supabase schema domain without overriding a value already set. `init`'s settings merge now keeps comments in that file. - **Schema.** The CLI config JSON schema sets `allowTrailingCommas: true`, so VS Code doesn't flag the trailing commas the converter writes. The converter lives in `@supabase/config/internal`. The TOML scanner moved out of `config-edit.ts` into `toml-scan.ts` so both can share it. ## Linked issue CLI-2867
This PR was automatically created to sync the generated `@supabase/api` package with the latest Management API OpenAPI document. Changes were detected in the upstream OpenAPI documents exposed by `https://api.supabase.com/api/v1-json` and `https://api.supabase.com/api/v2-json`. Co-authored-by: jgoux <1443499+jgoux@users.noreply.github.com>
…7571) ## TL;DR fixes edge functions on the stack connecting to the database as the `supabase_admin` superuser instead of `postgres` ## whats hurting? on the stack, `SUPABASE_DB_URL` inside a function pointed at `supabase_admin`, while the legacy backend and deployed functions get `postgres` so a function could run superuser only sql locally that fails once deployed tables a function created were also owned by `supabase_admin`, so migrations and `db query` as `postgres` could not alter them ## now fixed by? the stack hands functions a `postgres` role url, on `supabase start` and on `functions serve` a stack created before this picks it up on its next `stack start`, while `stack restart` keeps the saved value ## ref: - missed in: #6684
This bumps realtime from v2.156.2-r0 to v2.157.0-r0. https://github.com/supabase/slim-services/releases/tag/realtime-v2.157.0-r0 Planned after supabase/slim-services published v2.157.0-r0. This branch is rewritten from develop whenever a newer relevant release arrives. Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
Towards FUNC-954
## Current Behavior There's no way to check a config file on its own. Problems only show up when some other command loads the config, and that load stops at the first one it hits, so fixing a broken config means running a command, fixing one thing and running it again. ## Expected Behavior `supabase config validate` checks the selected config file (`config.toml`, `config.json` or `config.jsonc`) as committed and reports every problem it can decide in one go: - syntax errors, the schema version check, the two-files rule, removed keys, unknown keys (with a "did you mean" suggestion), type/enum/value errors, the schema's table rules and the cross-key checks `start` applies - every `[remotes.*]` block, each checked as base merged with that remote and labelled with the remote it applies to - the file, line and column of each problem, located in the raw TOML/JSON text By default every `env()` value is unknown, so the result is the same on every machine and in CI. A check that reads an unknown value isn't run, and one info line says how many values come from `env()` and how many checks were skipped. `--resolve-env` looks each reference up the way the loader does (shell, then the `.env*` files picked by `SUPABASE_ENV`). Set values are checked like literals. An unset or empty variable is an error when a check needed it or it feeds a number or boolean key, and a warning otherwise. The message names the variable and every place it's used. Text output follows the `config convert` report style from #7517. It never prints source text or configured values ([ADR 0034](docs/adr/0034-config-diagnostics-print-no-source-text.md)). `--output-format json` emits a versioned payload with codes, key paths, positions, the deferred checks and a `complete` flag. The command exits 1 on any error and never contacts the Supabase API. The design is in the Linear doc on CLI-2866 (revision 4). A few things worth calling out for review: - Validate has its own engine (`commands/config/validate/validate.engine.ts`) instead of a validation mode in the shared loader. Unknown values decode as their schema default or a typed stand-in on a copy of the document, and a problem is only reported when everything it read is known. - The shared local config checks are now small scoped units (`command-internal/config-check-*.ts`). `start`, `db` and `validate` all run them, and `start`/`db` keep their order and first error message; the parity test pins that. `start`'s inline duration checks moved into units that only `start` and `validate` run. - Shared steps no longer drop a key or branch on a value that could be unknown, i.e. SMTP, external/third-party providers, passkey and SMS now enumerate by presence and leave the value read to the check. Nothing changes for configs `start` accepts. - `@supabase/config/internal` gains a decode with the table-level rules stripped, the table rules themselves, `SMS_PROVIDER_REQUIRED_FIELDS`, `EXTERNAL_PROVIDER_IDS` and `scanCliConfigDocument`. Nothing new is public, but it touches `packages/config` so it will release. - Having both `config.toml` and `config.json` is an error in `validate`, while other commands still warn and pick one. It felt wrong for a validator to pass a project that's ambiguous about which file it reads, but happy to change that if it's too strict. - A run that finds problems is recorded in telemetry as a failed command with a user-actionable `invalid_config` error (`ConfigValidateProblemsError`), the same way `db lint` classifies its failures. ## Related Issue(s) Fixes CLI-2866
This bumps edge-runtime from v1.78.1-r0 to v1.79.0-r0. https://github.com/supabase/slim-services/releases/tag/edge-runtime-v1.79.0-r0 Planned after supabase/slim-services published v1.79.0-r0. This branch is rewritten from develop whenever a newer relevant release arrives. Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
This bumps postgres from 15.19.0.019-r0 to 15.20.0.000-r0. https://github.com/supabase/slim-services/releases/tag/postgres-15.20.0.000-r0 Planned after supabase/slim-services published 15.20.0.000-r0. This branch is rewritten from develop whenever a newer relevant release arrives. Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
This bumps postgres from 17.11.0.019-r0 to 17.12.0.000-r0. https://github.com/supabase/slim-services/releases/tag/postgres-17.12.0.000-r0 Planned after supabase/slim-services published 15.20.0.000-r0. This branch is rewritten from develop whenever a newer relevant release arrives. Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
Adds `--local <path>` to the CLI reference publisher to write the generated spec into an existing `supabase/supabase` checkout without cloning, committing, pushing, or opening a pull request. Local preview and release publishing share the same write and formatting path. Local mode validates input before writing, supports `--dry-run`, and rejects missing target directories, absolute spec paths, and parent-directory traversal outside the checkout. Documents preview setup and links to the website's local development instructions. This change covers only the CLI reference spec, not sidebar entries or JSON Schemas. Closes CLI-2679
This bumps edge-runtime from v1.79.0-r0 to v1.79.1-r0. https://github.com/supabase/slim-services/releases/tag/edge-runtime-v1.79.1-r0 Planned after supabase/slim-services published v1.79.1-r0. This branch is rewritten from develop whenever a newer relevant release arrives. Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
This bumps realtime from v2.161.0-r0 to v2.162.0-r0. https://github.com/supabase/slim-services/releases/tag/realtime-v2.162.0-r0 Planned after supabase/slim-services published v2.162.0-r0. This branch is rewritten from develop whenever a newer relevant release arrives. Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
This bumps realtime from v2.162.0-r0 to v2.163.0-r0. https://github.com/supabase/slim-services/releases/tag/realtime-v2.163.0-r0 Planned after supabase/slim-services published v2.163.0-r0. This branch is rewritten from develop whenever a newer relevant release arrives. Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
This bumps realtime from v2.163.0-r0 to v2.163.1-r0. https://github.com/supabase/slim-services/releases/tag/realtime-v2.163.1-r0 Planned after supabase/slim-services published v2.163.1-r0. This branch is rewritten from develop whenever a newer relevant release arrives. Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
This bumps realtime from v2.163.1-r0 to v2.164.0-r0. https://github.com/supabase/slim-services/releases/tag/realtime-v2.164.0-r0 Planned after supabase/slim-services published v2.164.0-r0. This branch is rewritten from develop whenever a newer relevant release arrives. Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
This bumps realtime from v2.164.0-r0 to v2.165.0-r0. https://github.com/supabase/slim-services/releases/tag/realtime-v2.165.0-r0 Planned after supabase/slim-services published v2.165.0-r0. This branch is rewritten from develop whenever a newer relevant release arrives. Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
## TL;DR warns on stderr when a `gen types --schema` name is missing from the generated types ## whats missing? `gen types` silently drops any `--schema` name the database does not return and still exits 0 a typo, a stray space after a comma, a migration that has not run yet, or a schema the connecting role cannot use all leave a valid looking types file with that schema gone ## now fixed by: each such name now prints one warning on stderr ``` Warning: schema "<name>" is missing from the generated types: it does not exist, the connecting role cannot use it, or its name starts with pg_. ``` stdout and the exit code stay the same, and only names passed to `--schema` warn, never the defaults > [!NOTE] > types from the Management API carry no schema list to check, so that path is unchanged ## ref: - closes: #7625
This bumps pooler from v2.10.0-r0 to v2.10.1-r0. https://github.com/supabase/slim-services/releases/tag/pooler-v2.10.1-r0 Planned after supabase/slim-services published v2.10.1-r0. This branch is rewritten from develop whenever a newer relevant release arrives. Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
…7607) ## TL;DR fixes `supabase start` and `functions serve` failing on every restart on Docker Desktop when `supabase/functions/.env` holds a multiline value ## what's missing? a multiline value such as a quoted PEM key was staged under `supabase/.temp/start-secrets` and bind mounted into the edge runtime. `stop` deletes that folder and the next `start` recreates it, and Docker Desktop then fails the `docker cp` of the main service with `must be a directory`, so only the first start in a project ever worked ## now fixed by: the multiline env script and its values now ride in the same `docker cp` archive as the main service, so the edge runtime no longer needs that host folder. a `multiline-env` folder left by an older CLI is removed on the next start <details> <summary>visual rep</summary> ```mermaid flowchart TD A["supabase/functions/.env has a multiline value<br/>supabase start, stop, start again on Docker Desktop"] A -->|before| B["value staged in .temp/start-secrets/…/multiline-env<br/>and bind mounted into the edge runtime"] B --> C["stop deletes that folder, start recreates it"] C -.-> D["docker cp fails with must be a directory<br/>first start works, every restart fails"] A -->|now| E["script and values ride in the docker cp archive<br/>next to the main service, no host bind"] E --> F["every restart and functions serve start<br/>value arrives byte for byte"] classDef t fill:transparent class A,B,C,D,E,F t ``` </details> ## ref: - closes: #7603 - missed in: #6273
This PR was automatically created to sync the generated `@supabase/api` package with the latest Management API OpenAPI document. Changes were detected in the upstream OpenAPI documents exposed by `https://api.supabase.com/api/v1-json` and `https://api.supabase.com/api/v2-json`. Co-authored-by: jgoux <1443499+jgoux@users.noreply.github.com>
## TL;DR `supabase bootstrap` on macOS asked for the keychain password up to 4 times, now it asks once ## what's hurting the user(s)? every part of the cli that needs your access token read it from the keychain again instead of reusing the first read one `bootstrap` run does that 4-7 times, and `secrets list`, `projects api-keys`, `projects list` and `orgs list` 2 to 3 times macOS counts each read as a new access, so anyone who clicks Allow instead of Always Allow gets a password prompt for every one of them ## now fixed by: reading the keychain once and sharing that token with every caller in the same run login, logout and the credentials sweep clear it so a new or removed token is picked up right away, and `--debug` output stays the same ## ref: - missed in: #5331 - closes CLI-3135
## TL;DR prefills the `bootstrap` directory prompt with the current dir so you see and edit where the project lands ## whats improved? the prompt starts with the current dir and a trailing slash, so Enter keeps it in plain sight and typing `my-app` lands in a fresh subfolder piped stdin and dumb terminals get the path as a placeholder, so scripts and `--workdir` behave as before ## why? blank meant the current dir without ever showing it, so a quick Enter inside an existing project went straight to an overwrite prompt that defaults to yes <!-- linear:table-colwidths:364,363 --> <table> <thead> <tr> <th>before</th> <th>after</th> </tr> </thead> <tbody> <tr> <td> https://github.com/user-attachments/assets/e1ca203b-6d73-48dd-a5ec-0f349fd52d15 </td> <td> https://github.com/user-attachments/assets/88028b6d-e22d-4a57-8a3e-ba948c9788f4 </td> </tr> </tbody> </table> ## ref: * closes: [CLI-3137](https://linear.app/supabase/issue/CLI-3137/make-the-bootstrap-destination-directory-default-clearer)
This bumps studio from 2026.11.09-sha-1567384-r0 to 2026.11.16-sha-51bd16d-r0. https://github.com/supabase/slim-services/releases/tag/studio-2026.11.16-sha-51bd16d-r0 Planned after supabase/slim-services published 2026.11.13-sha-64fad1a-r0. This branch is rewritten from develop whenever a newer relevant release arrives. --------- Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com> Co-authored-by: Julien Goux <hi@jgoux.dev>
This bumps storage from v1.95.0-r0 to v1.96.0-r0. https://github.com/supabase/slim-services/releases/tag/storage-v1.96.0-r0 Planned after supabase/slim-services published v1.96.0-r0. This branch is rewritten from develop whenever a newer relevant release arrives. Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
## TL;DR adds `supabase stack prune`, which destroys managed local stacks whose project directory no longer exists ## whats introduced? `supabase stack prune` finds the stacks whose saved project directory is gone, like a deleted worktree or a moved project, and destroys them with their data after one confirmation. `--dry-run` only lists them, `--yes` skips the question for CI and agents, and `--output-format json` lists what was removed. a stack whose directory still exists is never selected, and each directory is checked again right before its stack is destroyed. `stack destroy` and `stack list` behave as before. ## why? stacks of deleted worktrees and moved directories keep their data and can keep running. finding them meant reading `stack list` and checking each project path by hand before passing the ids to `stack destroy`. agents create and remove worktrees all the time, so these piled up with no way to find them. ## ref: - related: #7568 - related: #7269 - closes: CLI-2548
A newer slim release rebuilt its `slim-bump/<service>` branch from `develop` and force-pushed it, which dropped any commit a maintainer had pushed to make the bump pass CI (as on #7634). The apply step still generates every bump from the `develop` checkout, so no code from the branch runs while the app token is available. When the branch's open PR carries commits authored by someone other than the workflow, those commits are cherry-picked onto the fresh bump; merge commits and the workflow's own superseded commits are dropped. A replay conflict fails the run with manual steps. A leftover branch without an open PR is rebuilt as before. Every push now uses `--force-with-lease` on the tip read just before, so a push that lands mid-run is never overwritten. The PR body the workflow writes and ADR 0026 "Hotfix and upgrade pickup" describe the new behavior.
## Summary Most `db` commands have pinned to the Supabase CA. `db dump` and `db push` did not. Ensure the CA is pinned for all commands ## Linked issue Closes #cli-3140 ## Checklist - [x] The PR title follows [Conventional Commits](https://www.conventionalcommits.org/) (e.g. `fix(cli): …`). - [x] Tests added or updated for the change. - [x] From the repository root, `pnpm check:all` passes; relevant package tests pass for every touched workspace, and `pnpm types:check` passes for each touched TypeScript workspace (or workspace declaring it). --------- Co-authored-by: 7ttp <117663341+7ttp@users.noreply.github.com>
This bumps realtime from v2.165.0-r0 to v2.167.0-r0. https://github.com/supabase/slim-services/releases/tag/realtime-v2.167.0-r0 Planned after supabase/slim-services published v2.167.0-r0. A newer relevant release rewrites this branch from develop and replays any commits others pushed to it. Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
## TL;DR Teardown now waits up to 60s (exponential backoff capped at 2s) for another stack's Docker prune, instead of a fixed ~10s attempt budget. - `stop.e2e.test.ts › stop --no-backup --debug reports real pruned containers, volumes, and network`: `stop` failed with `failed to prune containers: Error response from daemon: a prune operation is already running`. Docker serializes prunes per daemon, and another stack's container/volume prune in the same shard held the lock longer than the 40 × 250ms retry window in `docker-remove-all.ts`. The wait is now bounded by elapsed time, and an exhausted wait names how long it waited and how many attempts it made. This affects `stop` and the `start` rollback teardown for users running several stacks too. ## ref spotted in: https://github.com/supabase/cli/actions/runs/42601796698/job/129107821033 --- _Generated by [Claude Code](https://claude.ai/code/session_01MBEk1MjMxzzyP1gVeMiWys)_ --------- Co-authored-by: Claude <noreply@anthropic.com>
## TL;DR Fixes one whole-stack flake and adds named diagnostics to two flakes whose cause is still unknown. - `whole-stack.podman.lifecycle.e2e.test.ts › writes stay confined…`: Studio's metadata call returned 502. On a cold engine, Studio was still in its `preparing` stage when the traffic wake budget ran out. The flow now calls `prepare` on Studio before the request, so the request only exercises the lazy wake. - `Pooler.integration.test.ts › serves Pooler SQL against the owned database`: this test only ever reported its 120s test timeout. Each phase (database start/readiness/relay, and the pooler's start/readiness/connect/query/stop for each runtime and mode) now has a named guard deadline, so a stall names the step that hung. The `phaseDeadline` helper moves to `packages/stack/tests/` and the storage suite now shares it. - `advisors.live.test.ts › reads advisor findings over the database connection` (and its sibling, which failed in the same run): when `db advisors` returns no findings, the test now reports the table's RLS flag, the `anon` and `authenticated` SELECT privileges, and the exposed schemas as the lint session's role sees them. ## ref spotted in: - https://github.com/supabase/cli/actions/runs/43413521407/job/131845624205 - https://github.com/supabase/cli/actions/runs/43413521407/job/131845624199 - https://github.com/supabase/cli/actions/runs/43401298717/job/131804505472 --- _Generated by [Claude Code](https://claude.ai/code/session_01MBEk1LMouEhCXP5bABWjgJ)_ --------- Co-authored-by: Claude <noreply@anthropic.com>
## TL;DR Two flakes with unknown causes now fail with the details needed to find the real cause. - `contract.native-tls.integration.test.ts › consumer contract (stack/native over api.tls, project pair)`: the run failed with `Owner response unavailable; the request outcome is uncertain: HttpError: `. The RPC client error message for a serialized HTTP failure drops the transport kind and the nested causes. The stack error now names the transport kind, each nested message and the socket error code, for example `HttpError TransportError: Transport error (POST …): fetch failed: other side closed (UND_ERR_SOCKET)`. - `Mail.integration.test.ts › keeps a captured email after a native mail stop and start`: the mailbox came back empty (`expected [] to include 'persistence-test'`) without showing which process replied. Readiness accepts any 2xx on the port. Mailpit prints no line after its HTTP bind, so a bind-line readiness check can't confirm the bind. After each start the test now checks that the Mailpit answering uses this instance's database (`/api/v1/info`), and adds Mailpit's output and port to each failure. ## ref spotted in: - https://github.com/supabase/cli/actions/runs/43201489194/job/131125766890 - https://github.com/supabase/cli/actions/runs/43176200373/job/131040928003 🤖 Generated with [Claude Code](https://claude.com/claude-code) https://claude.ai/code/session_01MBEk1MLMsKMitQvScEUGE1 --------- Co-authored-by: Claude <noreply@anthropic.com>
This bumps auth from v2.197.0-r0 to v2.198.0-r0. https://github.com/supabase/slim-services/releases/tag/auth-v2.198.0-r0 Planned after supabase/slim-services published v2.198.0-r0. A newer relevant release rewrites this branch from develop and replays any commits others pushed to it. Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
Coly010
approved these changes
Nov 16, 2026
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.