Repository navigation
chore: version packages (rc) - #674
Open
github-actions[bot] wants to merge 1 commit into
Open
github-actions[bot] wants to merge 1 commit into
github-actions[bot] wants to merge 1 commit into
Conversation
github-actions
Bot
force-pushed
the
changeset-release/main
branch
7 times, most recently
from
November 1, 2026 23:51
7b1b13d to
688f364
Compare
github-actions
Bot
force-pushed
the
changeset-release/main
branch
from
November 2, 2026 01:36
688f364 to
0eeed5e
Compare
This branch has not been deployed
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
This PR was opened by the Changesets release GitHub action. When you're ready to do a release, you can merge this and the packages will be published to npm automatically. If you're not ready to do a release yet, that's fine, whenever you add more changesets to main, this PR will be updated.
mainis currently in pre mode so this branch has prereleases rather than normal releases. If you want to exit prereleases, runchangeset pre exitonmain.Releases
everything-dev@2.0.0-rc.28
Minor Changes
bos upgrade), effect 4.0.0-rc.118 with the vendored clone retagged and the unstable/http subpaths promoted to effect/http/*, alchemy 2.0.0-beta.83 (beta.84 removed the Docker/Util surface), module-federation 2.9.2, tanstack query 5.104.5 / table 9.3.1 / form 1.33.5 / intent 0.5.6, zod 4.6.7, yaml 2.9.1, drizzle patches.pglite:.bos/db, per workspace process), so concurrent local projects can never collide on DB ports and data survives restarts without any.env.bos devmigrates leftovers once: compose-erapostgres://everythingdev:…@localhost:543x/…URLs in.envare rewritten to the pglite defaults and a scaffold-shapeddocker-compose.ymlis relocated to.bos/compose-retired/.db:*commands work against pglite (drizzle-kit pglite dialect selected by URL scheme); the CLI/scaffold drop the compose template,dev:postgres*scripts, post-init docker prompt, and preflight auto-start; regression tests and the deploy smoke run on pglite (TEST_DATABASE=postgres+ an external URL remains the parity escape hatch). Productionpostgres://URLs are unchanged.Patch Changes
282ede6: Auth plugin surfaces rate-limit and client-IP configuration (auth: Better Auth rate limiter falls back to a single shared bucket behind proxies — expose advanced.ipAddress (and rateLimit) config #688). New authored variables under
auth.variables:rateLimit: { enabled, window (seconds), max, storage: "memory" | "database" }andipAddress: { ipAddressHeaders, trustedProxies }. A newrate_limittable backsstorage: "database"— the multi-replica option, since in-memory buckets are per instance (migration applied on boot).trustedProxiesis strictly opt-in: Better Auth walks the forwarded chain right-to-left and takes the first untrusted address, so list your proxies' own addresses, never a range that also covers clients (bug: Rate limiter keys off spoofable x-forwarded-for; 15-min unswept growth window #338). Deployments behind appending proxies (Cloudflare + Railway) should name a single trusted header (cf-connecting-ip,x-real-ip) — the defaultx-forwarded-forcannot resolve a client IP through an untrusted comma-chain, which is why the boot WARN fell back to a shared per-path bucket. Absent configuration leaves Better Auth's defaults unchanged;BETTER_AUTH_RATE_LIMIT_DISABLED=1still wins. The root runtime's canonical posture inbos.app.tsisipAddress: { ipAddressHeaders: ["cf-connecting-ip"] }with norateLimit— library defaults stand; the root is Cloudflare-proxied, children inherit it throughextends, and each deployment can override.8c916ff: Bump the platform catalog: better-auth 1.7.7 → 1.7.14 (with @better-auth/*), @orpc suite 2.0.0-beta.35 → 2.0.0-beta.45 (
PublisherResumeOptionstakesttlin milliseconds instead ofseconds), @tanstack/react-router 1.170.32 → 1.170.46 (router-core 1.171.39, router-plugin 1.168.48, router-generator 1.167.46), and the Effect toolchain trio in lockstep (@effect/tsgo 0.63.0, oxlint 1.90.0, oxlint-tsgolint 7.0.2005). The pnpm pin moves to 12.16.1.Overrides now reference the catalog through pnpm's
catalog:protocol so shared-dep pins cannot drift from the catalog (bos upgrade doesn't reconcile stale pnpmoverridesthat shadow the platform catalog — children fail the #546 skew gate pointing at the wrong fix #659/bos upgrade should reconcile v2-migration pnpm overrides with the platform catalog, and audit/retire child ui/rsbuild.config.ts overrides #682);bos upgradereconciles child overrides, propagates the pnpm pin forward-only to existing v2 children, and regenerates the frameworkminimumReleaseAgeExcluderows from the catalog withminimumReleaseAgeExcludePruneenabled — the list replaces instead of accumulating. The v1→v2 migration'sV2_CATALOGbaseline is drift-guarded against the parent catalog (row-for-row, both directions), closing upgrade --from v1: drops app.api.variables, overwrites custom authed-mount guards, writes outdated catalog + router-client shim #650's outdated-catalog vector.680ea78: Concurrent
bos devsessions in one config dir no longer crash artifact generation:.bos/bos.resolved-config.jsonis now written atomically (tmp + rename) so a sibling session never reads a torn file, and the contract-bridge stale sweep removes gen files withrmSync(…, { force: true })so a file vanishing between the existence check and the removal no longer fails the boot withENOENT … lstat.024e9c7: Release-scoped slot stamping no longer skips a slot the child only configures. A child that sets
auth: { variables, secrets }(or a remote-onlyapi: { variables, secrets }) had that slot counted as authored, so the deploy train left it unpinned and the auth server (or api) kept following the parent's live config pointer — under a host and auth UI pinned to the child's release. A slot now counts as authored only when it names a code source (production,development,pin,extends,proxy, a bare plugin ref, orui: false); configuration-only entries get the release'sproduction+pinstamped with theirvariables/secretskept. Affected children pick this up on their next deploy after upgrading.2f72431: Auth plugin onboard page: a rescan by someone who has left the community now says they're no longer a member instead of "You're in". The page checks a fresh organization list against the redeemed organization ID, clears a stale active organization, and shows a retryable error if the check fails. The redeem's refusal for a member who left now carries
data.action: "rejoin-organization". Child apps can configure the joined screen through theonboardauth variable:links.next/links.home(with an{organizationId}placeholder) andlinks.explorefor the success and error panels (defaults stay/build,/dashboard,/explore;next: falsehides the build link), pluscommunity.url, an app endpoint that returns the joined community's page and channels for a "Visit {community}" button and channel links.86eb1df: SSR session resolution is tri-state and never round-trips the public edge on the hot path (SSR session resolution fetches the public edge URL — resolve in-process (or internal base URL) and degrade gracefully #689). The host's per-request in-process
getContextresolution now distinguishes unresolved (no auth plugin, or resolution failed →undefined) from authoritative sign-out (getContext succeeded with no session →null). The SSR query-cache seeding follows: authoritative null seeds the cache (the same explicit signed-out entry sign-out writes), undefined leaves it empty so the guard's fetch still runs — closing the dev:ssr SSR session resolution returns null — authed routes 307 to /login despite a valid session #527 failure mode where a wedged auth resolution redirected signed-in users to /login without ever fetching. A server-side session-fetch failure degrades to a signed-out render instead of failing the whole document: an edge rate-limit or block page can no longer 500 every SSR page for every visitor.Updated dependencies [8c916ff]
Updated dependencies [150b092]
every-plugin@3.0.0-rc.15
Patch Changes
MemoryPublisher'sresumeoption takesttl(milliseconds) instead ofseconds.createRpcClientkeeps the request's own headers, includingcontent-type: application/json, when you pass custom headers such as an API key. Before, they were dropped and the server rejected the input.