4a190b1bd4
Merge branch 'main' into flags-version-header
2026-09-17 09:38:07 +02:00
2e03f56280
op
2026-09-16 13:19:47 +02:00
b167598b77
fix(flags-core): use plural flags config version headers
2026-09-16 12:04:24 +02:00
4848877ca6
[@vercel/flags-core] Add custom waitUntil option ( #497 )
2026-09-16 09:58:56 +00:00
3c12c2c04d
revert(flags-core): remove source debug logging
2026-09-16 11:44:50 +02:00
25154ede50
fix(flags-core): use singular flags config version headers
2026-09-16 11:18:23 +02:00
48afe11165
feat(flags-core): add bundled source debug diagnostics
2026-09-15 12:21:50 +02:00
0fd2f412e9
debug
2026-09-15 11:38:43 +02:00
004baf818f
fix
2026-09-15 10:25:46 +02:00
da0b044c2f
fix(flags-core): handle cancelled header refreshes and add changeset
2026-09-15 10:13:50 +02:00
923ea09467
stuff
2026-09-15 10:03:39 +02:00
7332270a2a
[vercel-flags-core] add vercel client mode
2026-09-14 18:30:25 +02:00
a3fc275ed3
docs(flags-core): clarify request-time initialization with OIDC ( #496 )
2026-09-10 16:36:37 +02:00
e0eebe6fbc
fix(flags-core): request identity on the stream so Bun receives the first datafile ( #494 )
...
* fix(flags-core): request gzip on the stream so Bun receives the first datafile
* test(flags-core): expect Accept-Encoding on stream requests in black-box tests
* fix(flags-core): send Accept-Encoding: identity on the stream only when running on Bun
2026-09-09 11:16:11 +02:00
c9d28116eb
Experiments ( #486 )
...
* step 1
* single experiment
* reuse variants
* single experiment
* exposures
* undo exposureLogging boolean
* Attach experiment metadata to all evaluated flag outcomes
* version
* hybrid and override exposure reporting
* Initialize adapters before reporting flag overrides
* Update package versions for experiment replacement
* versions
* fix test
* Rename exposure reporting APIs as experimental
* Add APIs for reporting flag exposures and overrides
* Update changeset for exposure reporting APIs
* Remove default exposure reporting
* Defer exposure reporting with waitUntil
* Fix override reporting without a reporter
* Handle optional override reporting in Vercel adapter
* Update override reporting to accept parameter objects
* patch
2026-09-04 12:21:17 +00:00
cc8c26648c
feat(flags-core): add metrics environment option ( #453 )
...
* feat(flags-core): support explicit environment
Co-Authored-By: Claude <noreply@anthropic.com >
Co-Authored-By: Luis Meyer <luis.meyer@vercel.com >
* fix(flags-core): scope environment to metrics ingest
Co-Authored-By: Claude <noreply@anthropic.com >
Co-Authored-By: Luis Meyer <luis.meyer@vercel.com >
* fix(flags-core): clarify metrics environment
Co-Authored-By: Claude <noreply@anthropic.com >
* fix(flags-core): restrict metric environment values
* docs(flags-core): clarify metrics environment fallback
---------
Co-authored-by: vercel[bot] <35613825+vercel[bot]@users.noreply.github.com>
Co-authored-by: Claude <noreply@anthropic.com >
2026-08-26 12:28:13 +02:00
58e1f5bcdf
Rename Edge Config packages to Global Config ( #452 )
...
* Rename edge config packages
* Rename Global Config APIs
* Fix Global Config environment setup
* Rename Global Config endpoints
* Update Global Config documentation links
* Undo lockfile changes
* Update lockfile
* Undo lockfile changes
* Update lockfile
* Fall back to EDGE_CONFIG
---------
Co-authored-by: Andy Bitz <artzbitz@gmail.com >
2026-08-04 12:36:41 +02:00
ceb15198f7
[@vercel/flags-core] strip g and y regex flags ( #450 )
2026-07-27 13:50:24 +00:00
1f000784e5
upgrade vite & vitest ( #445 )
2026-07-26 08:25:37 -04:00
50b1aa4f96
Align rollouts assignments with split assignments ( #427 )
...
* [@vercel/flags-core] align rollotus and splits
* avoid extra array
* 100k
* full range
* fix comments
* rename
* shorten comments
* add changeset
* add stderr test
2026-07-09 22:48:48 +03:00
9dff590bd5
Add FLAGS_EVALUATION metric ( #401 )
...
* Add FLAGS_EVALUATION metric
* Track variant IDs in evaluation results
* report null instead of fallback
* bucket to the previous minute
* replace variant null with undefined
* invert
* use option instead of env var
* Add track option to black-box flag tests
* Add evaluation metrics controls and flush reasons
* increase max_count
* defaul track:true
* increase batch
* Bound ingest POSTs to 2000 events per request
Flushes can overshoot the scheduler's MAX_COUNT because the map read
happens in a microtask, so a single POST could exceed the server-side
maxItems cap and be rejected with a 400. Chunk sendIngestEvents so each
POST carries at most MAX_EVENTS_PER_REQUEST events.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com >
* Remove count-based flushing from the scheduler
Flushing is now purely time-based (idle window, max window, shutdown).
Batches are only cut down to size when the ingest events are sent, via
the 2000-event chunking in sendIngestEvents, which keeps every POST
below the server-side maxItems cap.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com >
* cleanup
---------
Co-authored-by: Claude Fable 5 <noreply@anthropic.com >
2026-07-08 14:09:45 +02:00
f60c99d707
[vercel-flags-core] fix datafile serialization ( #416 )
2026-06-26 17:40:11 +03:00
c35f38dfb9
Support createClient options-first overload ( #406 )
2026-06-23 14:32:06 +02:00
b0150af9c8
[@vercel/flags-core] quiet noisy stream reconnect logs ( #395 )
...
* [@vercel/flags-core] quiet noisy stream reconnect logs
* [@vercel/flags-core] test exhaustion log + guard undefined reason
2026-06-16 14:43:25 +00:00
201f9d5988
Add bulk evaluation ( #385 )
...
* [flags] avoid re-imports
Avoid re-importing next/headers since it adds unnecessary microtask queue overhead
* [flags] avoid iife microtask queue overhead
* [flags] skip awaits where possible
* [flags] allow bulk eval
* wip
* add bulk mode to playground
* first try of bulk eval
* add bulk evaluation to adapters
* reuse
* simplify
* versions
* rm outdated changeset
* rm outdated changeset
* revert playground
* revert playground flags
* rm logs
* reuse cache of resolved flags for bulk
* simplify
* changesets
* support bulk([]) and bulk({})
* flagKey → key
* allow any type in expectPermutations
* avoid unnecessary mapping
* fix changeset
* validate package.json fields
* update package.json fields
* keep covariant
* adjust adapter types
* Merge bulk into evaluate (#392 )
* merge bulk into evaluate
* add pages router compatibility
* better types
* add array test
* add test
* changesets
* use type import
* fix import
* extract readOverrides
* fix import of reportValue
* tracing
* skip bulkDecide for overwritten flags
* update changesets
* [changesets] onlyUpdatePeerDependentsWhenOutOfRange
* add type guard
* adjust changeset
2026-06-05 08:14:29 +00:00
7b5ea9a808
Reapply OIDC ( #390 )
...
* Reapply "Add oidc support (#374 )" (#389 )
This reverts commit 357ca676c6
2026-06-04 16:37:03 +02:00
357ca676c6
Revert "Add oidc support ( #374 )" ( #389 )
...
This reverts commit 72d36511d0
2026-05-28 10:16:43 +03:00
72d36511d0
Add oidc support ( #374 )
...
* Add oidc support
* changeset
* update tests
* update def script
* fix provider data
* update changeset
* fix createVercelAdapter
* comments
* error
2026-05-28 08:59:00 +02:00
4d90e912a4
[flags-core] perf improvements ( #382 )
...
* [flags-core] perf improvements
* add changeset
2026-05-18 13:20:13 +03:00
bd4d01a9b2
vercel-flags-core: jitter ingest retries and batch-wait window ( #371 )
...
Replace the deterministic 100/200ms retry schedule with full-jitter exponential
backoff, jitter MAX_BATCH_WAIT_MS by ±20% to desynchronize concurrent flushes,
and log when a flush exhausts all retries.
Helpers extracted to utils/backoff.ts so they can be reused.
Co-authored-by: Cursor <cursoragent@cursor.com >
2026-05-06 15:57:15 +02:00
80dcdada34
[vercel flags] add progressive rollouts to sdk ( #360 )
...
* [vercel flags] add progressive rollouts
* fix slots
* add changeset
2026-04-14 16:55:57 +03:00
9142790975
only build publishable packages during release process ( #348 )
...
* clean up
* fix release process
2026-03-21 21:17:04 +02:00
b755ffecdf
fix: SDK key detection to avoid false positives with third-party identifiers ( #345 )
...
The SDK key validation now uses a regex (/^vf_(?:server|client)_/) to require
the format vf_server_* or vf_client_* instead of accepting any string starting
with vf_. This prevents false positives with third-party service identifiers
that happen to start with vf_ (e.g., Stripe identity flow IDs like
vf_1PyHgVLpWuMxVFx...).
Adds isValidSdkKey() helper function and updates parseSdkKeyFromFlagsConnectionString()
to use the stricter validation. Updates all tests to use valid SDK key formats.
2026-03-21 17:39:00 +00:00
44460570d4
[vercel/flags-core] allow json flags ( #335 )
...
* rm unused import
* [vercel-flags] allow json flags
* add flags-playground
* wip
* clean up type
* reword changeset
* rm examples/flags-playground
2026-03-20 14:03:53 +02:00
b81963dc14
fix ( #339 )
2026-03-19 16:18:59 +02:00
64619d7c66
@vercel/flags-core: allow specifying entities type when creating clients ( #334 )
...
* changeset
* adapt
2026-03-13 11:09:21 +02:00
4a5f56a8c7
skip sending config reads ( #330 )
...
* skip sending config read for dev and custom envs
* add changeset
* clean up
* wip
2026-03-13 08:16:47 +01:00
dd1396e461
Guard missing Vercel Flags runtime helpers ( #326 )
...
* Guard request-context flag reporting
* Preserve request-context hook binding
* Guard core request-context reporting
* Guard missing bundled flag definitions
---------
Co-authored-by: Dominik Ferber <dominik.ferber@gmail.com >
2026-03-12 16:38:18 +02:00
e8539cd52a
make version dynamic in tests ( #304 )
2026-03-06 12:02:03 +00:00
b70c2ea466
Control ( #278 )
...
* refactor
* remove retries
* simplify
* simplify options
* before
* rename DataSource to Controller
* refactor and better tests
* avoid double polling
* rename
* fix
* simplify test setup
* wip
* unify
* wip
* only track 1 read per build
* make BundledSource lazy
* clean up fallback behavior
* add mode
* mutually exclusive streaming & polling
* rely more on black box testing
* simplify
* Update CLAUDE.md
* various fixes
* add tests
* more fixes
* Update CLAUDE.md
* more fixes
* resolve more issues
* unite black box tests
* tests
* progress
* adjust
* don't report on 401s; use Response.json
* enforce min polling interval
* progress
* types
* capture all logs
* throw with prefix
* added a minimum gap of `BASE_DELAY_MS` (1 second) between connection
attempts
* step
* progress
* use request context in tests
* add separate tests depending on context
* tests
* rm unused option
* add state machine chart
* rm sources
* use fake timers for more tests
* swap remaining tests from msw to mocked fetch
* update comments
* Update CLAUDE.md
* update comment
* fix comment
* fix types
* rm peek
* adds ping checks and sending x-revision header (#282 )
* abort on missed pings
* use revision from bundled definitions
* tests
* simplify
* send x-revsion on reconnects too
* simplify StreamSource
* rename
* rename in test
* wait for stream
* replace startBackgroundUpdates with explicit call
* Update CLAUDE.md
* restore stronger tests
* Update CLAUDE.md
* convert
* simplify tests
* fix test
* fixes
* rm state machine
* await first poll before resoliving init
* rm outdated comment
* avoid leaking _origin
* changeset
* avoid log on reconnect due to missed pings
* don't warn on missed pings
* Update event reporting for control rewrite (#289 )
* Update event reporting for control rewrite
* Log ingest response
* Use string for revision
* Add retries when ingesting events
* adapt changeset
* adjust test
---------
Co-authored-by: Luis Meyer <luis.meyer@vercel.com >
Co-authored-by: Andy <AndyBitz@users.noreply.github.com >
Co-authored-by: Andy Bitz <artzbitz@gmail.com >
2026-03-06 10:05:01 +02:00
722b0d054e
Added new comparators: (not_)contains and case-insensitive versions ( #288 )
...
* Added new comparators: (not_)contains and case-insensitive versions
* Switched to "ci" option instead of creating new condition for every ci-case
* Cleanup
* More cleanup
* Some more cleanup & finetuning
Made-with: Cursor
* Changeset
* Switch options key to just "i" and also allow for passing "i" as string instead of options object
* Apply suggestions from code review
* Update packages/vercel-flags-core/src/evaluate.ts
Co-authored-by: Dominik Ferber <dominik.ferber@gmail.com >
* naming
* Renaming
---------
Co-authored-by: Dominik Ferber <dominik.ferber@gmail.com >
2026-03-06 07:52:32 +02:00
a924044282
Fixed bug with inverted NOT_ONE_OF segment comparator ( #296 )
...
* Fixed bug with inverted NOT_ONE_OF segment comparator
* Feedback
* Changeset
* Updated unit-test
2026-03-05 10:52:50 +01:00
7d7719a255
[@vercel/flags-core] deduplicate concurrent initialize() calls ( #274 )
...
* add shop
* wip
* fix parallel init initialize
* avoid double polling
* add tests
* rm shirt-shop-vercel
* ignore
* refactor
* rm unnecessary fix
* minimal
* changeset
* fix tests
2026-02-13 14:34:46 +02:00
cf9ccd0076
Open Source Docs ( #272 )
...
* Scaffold Geistdocs instance
* Update biome.json
* Replace content
* Re-path docs
* Update docs package name
* Misc fixes
* Update geistdocs.tsx
* Move content to correct folder
* Start migrating custom components
* Finish migrating custom components
* Migrate existing redirects
* Remove #next suffixes from code blocks
* Migrate homepage and flags
* Remove duplicate lockfiles
* Start fixing homepage
* Replace Geist components
* Fix colors
* Improve layout
* Fix code blocks
* Fix hero
* Fix illustrations
* Fix redirect
* Fix nav links
* Patch in FrameworkSwitcher
* Bump Next.js in docs
* Delete turbo.json
* Misc fixes
* More logo fixes
2026-02-13 09:06:03 +02:00
2bfaada64c
Revert "add primed stream message support ( #263 )" ( #270 )
...
This reverts commit 8ae9101ba9
2026-02-11 15:40:45 +02:00
f14aea8e73
respect hashed keys ( #267 )
...
* respect hashed keys
* use cache for pending promises
* refactor
* add try/catch
* typo
2026-02-11 15:22:07 +02:00
8ae9101ba9
add primed stream message support ( #263 )
...
* add primed stream message support
Implements client-side support for version-based caching. When the client
already has the current datafile (indicated by revision), the server can
send a lightweight primed message instead of the full datafile, reducing
bandwidth. Tracks revision from datafile and sends X-Revision header on
reconnections.
Co-Authored-By: Claude Haiku 4.5 <noreply@anthropic.com >
* use getRevision getter for single source of truth and fix type errors
Refactors StreamConfig.revision to a getRevision getter function so
connectStream always reads the latest revision from client state instead
of tracking a separate copy. Fixes CI type-check errors in tests.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com >
---------
Co-authored-by: Claude Haiku 4.5 <noreply@anthropic.com >
2026-02-11 12:51:54 +01:00
84714c3b91
[vercel-flags-core] Forward configUpdatedAt timestamp correctly ( #266 )
2026-02-11 12:50:07 +01:00
5e2d095863
Suppress stream abort error on intentional shutdown ( #264 )
...
When shutdown() is called on a client using FlagNetworkDataSource, the stream abort error is expected and should not be logged. The break statement still exits the loop properly.
Co-authored-by: Claude Haiku 4.5 <noreply@anthropic.com >
2026-02-11 12:45:42 +01:00
d44a9229e3
pass default openfeature client ( #261 )
2026-02-11 12:52:29 +02:00
Luis Meyer