14859 Commits
Author SHA1 Message Date
Harsh Mahajan 8c1b58a6d2 Merge pull request #3155 from appwrite/fix/variable-key-validation
fix(variables): validate variable keys before writing them
8.7.41
2026-08-14 18:27:52 +05:30
harsh mahajan de9b9a529d Merge branch 'main' into fix/variable-key-validation
The three create-site pages conflicted only in their import blocks: main
dropped the FrameworkAdapterWithStartCommand import along with the adapter
start-command reads, while this branch added validateVariables next to it.
Kept both sides' intent — the type is gone from $lib/stores/sites and no
longer referenced, so only the validateVariables import remains.
2026-08-14 18:19:37 +05:30
Harsh Mahajan 36236db1ef Merge pull request #3160 from appwrite/refactor/remove-cloud-card-route
Remove the Cloud public beta card route
2026-08-14 18:18:56 +05:30
harsh mahajan 8c5d6ee60e Drop svelte-confetti, unused once the card route is gone
The card page was its only consumer.
2026-08-14 18:09:42 +05:30
harsh mahajan f5b26992c5 Remove the Cloud public beta card route
/console/card and /console/card/<uid> served the 2023 Cloud public beta
shareable card. The API no longer serves /cards/cloud, so both loaders
already bounced to / — the pages were unreachable.

getCampaignImageUrl was the only live export under the route and has
nothing to do with the card, so it moves to $lib/helpers/campaign.ts
next to its one consumer, the unauthenticated layout. getCardImgUrls
went with the route; the [uid] loader kept its own copy. windowFocus.ts
had no consumer besides Card.svelte, and the console layout's /card
path check guarded a route that never rendered under that group.
2026-08-14 18:09:39 +05:30
Harsh Mahajan 9e17b05fa8 Merge pull request #3157 from appwrite/fix/sites-hide-start-command
Keep the start command field, drop the adapter reads behind it
8.7.40
2026-08-14 11:28:41 +05:30
harsh mahajan 6fa8402cd0 Pin nanoid past the indefinite-loop advisory
bun audit fails on GHSA-2v37-7h3g-55p8: the lockfile pinned nanoid 3.3.17
under postcss and @ai-sdk/provider-utils, and neither `bun update nanoid`
nor updating those parents re-resolved it. Bun ignores npm's nested
override syntax, so the override has to be flat.

That takes the direct dependency to the 3.x line as well. The only use is
nanoid/non-secure, which 3.x provides, and the declared range now matches
what installs rather than claiming a 5.x that the override replaces.
2026-08-14 11:24:14 +05:30
harsh mahajan 58b37773f3 Reset the start command from the framework select, not the effect
The effect is skipped whenever the URL supplied any command, so a deploy
link carrying ?start= kept that entrypoint through a framework switch.
Resetting inside the effect could not fix it either: the effect also runs
on load, where the URL value has to survive.

Move the reset onto the select, so only a user changing framework clears
it and the link keeps working as written.
2026-08-14 11:17:34 +05:30
harsh mahajan a175d022e2 Clear the start command when the framework changes on create too
The settings form already resets it; the create flows kept the previous
framework's entrypoint in state and submitted it, and kept sending it
after a switch to a framework that only builds static.

Reset it alongside the other adapter-derived fields in both places.
2026-08-13 17:19:13 +05:30
harsh mahajan 40588237ed Keep the start command field, drop the adapter reads behind it
The field stays where it is. What goes is the code reading startCommand
off a framework adapter: the SDK never declared it and the API no longer
sends it, so the prefill and the placeholder were pointing at nothing, and
Reset restored a value that does not exist.

Also stop carrying a start command across a framework or adapter change -
it names an entrypoint only the previous framework has - and send it only
for the ssr adapter, matching the fallbackFile line beside it.
2026-08-13 16:16:05 +05:30
Harsh Mahajan ee7514c69b Merge pull request #3156 from appwrite/fix/sites-hide-start-command
fix(sites): keep framework start commands out of the console
8.7.39
2026-08-13 13:50:46 +05:30
Levi van Noort e10ae018bf fix(variables): filter valueless entries before validating an import
An entry with an empty value is discarded rather than written, so an
invalid key on one of them was rejecting the whole file and taking the
valid variables down with it.
2026-08-13 08:50:46 +02:00
harsh mahajan 8fab5505c9 Clear the start command when the framework or adapter changes
A start command names an entrypoint belonging to one framework. Carrying it
across a switch submits it against a runtime that has no such entrypoint,
and a switch to static leaves a command the form no longer shows.

Reset it alongside the other framework-derived fields, and send it only for
the ssr adapter, matching the fallbackFile line beside it.
2026-08-13 11:20:18 +05:30
harsh mahajan 6e64d0ce57 Match the start command copy used by the new console 2026-08-13 00:54:31 +05:30
harsh mahajan 40a286f6e0 fix(sites): keep framework start commands out of the console
The create-site flows read startCommand off a framework adapter and
offered it as a prefilled value and a placeholder. The API never returns
that field - FrameworkAdapter does not declare it - so the reads always
resolved to undefined, and the cast that made them typecheck was the only
thing hiding it. Had the field ever been returned, the console would have
put an internal runtime command in front of every user creating a site.

Drop the field from the create flows entirely and let the server pick its
own default. In site settings the input stays, but it moves under a
collapsed Advanced section, loses the Reset-to-default button that had
nothing to reset to, and takes a generic placeholder.
2026-08-12 23:55:44 +05:30
Levi van Noort 0feeb2a627 fix(variables): drop the key from the draft map before writing it
A rejected update left its key in vars, so the second pass picked it up
as a new variable and resent the stored key — the exact call the sparse
update is meant to avoid for keys that predate the identifier rule.
2026-08-12 12:28:58 +02:00
Levi van NoortandClaude Opus 5 1d4cfad9fa fix(variables): validate variable keys before writing them
The API now rejects variable keys that are not valid environment variable
names, but the console had no key validation anywhere — only a value
length check. Keys with hyphens, dots, spaces or a leading digit reached
the API and came back as a bare server error, and three flows handled
that badly:

- create-function and create-site create the resource (and its proxy
  rule) before the variables, so a rejected key left a function or site
  with no deployment behind
- the update modal resends the stored key on a value-only edit, so a key
  stored before the rule existed could no longer have its value changed
- promote-to-global deletes the existing global variable before creating
  its replacement, so a rejected key destroyed the original

Add a shared validator and check keys before submitting, in every entry
point that writes them. Send the key on update only when it changed, so
the API keeps the stored one and existing keys stay editable. Flag keys
that cannot be used as environment variable names in the variables
tables, since those are ignored at build and runtime. Report per-key
failures from the batch writes instead of a single rejection.

Server-side rule: appwrite/appwrite#13181

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-08-12 06:38:51 +02:00
Matej Bačo 00ba9b1a6a Merge pull request #3154 from appwrite/fix-cimd-client-id-resolution
Fix compatibility with external CIMD URLs
8.7.38
2026-08-11 16:55:04 +02:00
Matej BačoandClaude Fable 5 6b7c35ae15 chore(deps): raise the nanoid floor past GHSA-28wg-ghj8-5hjv and GHSA-2v37-7h3g-55p8
bun audit fails the CI build on three high nanoid advisories. All fixes are
within the existing semver ranges: the direct dependency and @melt-ui/svelte's
nested copy move to 5.1.16, and the 3.x copies under @ai-sdk/provider-utils
and postcss move to 3.3.17.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-11 16:44:21 +02:00
Matej BačoandClaude Fable 5 8ed7be09ae refactor(oauth2): inline the one-off CIMD helpers
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-11 16:36:53 +02:00
Matej BačoandClaude Fable 5 9dde997830 Fix compatibility with external CIMD URLs
Appwrite's apps.get()/apps.list() no longer resolve CIMD URLs — appId is
always a plain ID now. Detect URL-shaped OAuth2 client IDs in the console
and fetch the Client ID Metadata Document directly from the browser,
mapping its RFC 7591 fields onto Models.App for consent/device/applications
rendering. Fetch or validation failures fall back to hostname-only branding
so the flow never blocks.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-11 16:34:30 +02:00
Harsh Mahajan df7f976af5 Merge pull request #3151 from appwrite/fix-array-column-item-type 8.7.37 2026-08-07 12:02:20 +05:30
harsh mahajan a7c75f7be5 chore(deps): raise the js-yaml floor to 4.3.1
GHSA-5p4m-2wfm-xmqj (CVE-2026-59870) covers >=4.0.0 <4.3.1 and the fix was
not backported, so the existing >=4.3.0 override stopped clearing it and the
audit step in tests.yml now fails on every branch. It reaches us transitively
via eslint > @eslint/eslintrc > js-yaml.

Only the pinned floor and the resolved version move; integrity hashes are
untouched.
2026-08-07 11:14:28 +05:30
harsh mahajan 9f0ae45537 fix(databases): render array column items with their element type
The string fallback in the row column renderer exists for the spreadsheet
inline cell editor, where an entire array is edited as one comma-separated
textarea. It keyed off `column.array` (the column definition) rather than
the `array` prop (the render mode).

In the full row create/edit form, columnItem renders one Column per array
element and deliberately omits the `array` prop so each item gets its
element-type input. Since `column.array` was still true, every item
collapsed to a string input -- a Boolean[] column showed "Enter string".

This also caused intermittent save failures: the string component still ran
parseValue() against column.type, so literal true/false/1/0 coerced and
saved, while anything else yielded null, and a null inside a typed array is
rejected by the API.

Keying off the `array` prop restores per-item typed inputs (boolean select,
number input with min/max, date picker) while leaving the spreadsheet cell
editor and its "Advanced edit" link untouched, since only the two
fromSpreadsheet branches pass `array`.
2026-08-07 10:49:32 +05:30
Harsh Mahajan 6aa937295c Merge pull request #3149 from appwrite/fix/free-plan-readonly-blocks-project-creation
fix(billing): stop a zero seat limit turning free orgs read-only
2026-08-07 10:43:03 +05:30
harsh mahajan 750ec8eb82 refactor(billing): trim the comments down to what the code can't say
Two of the three added comments restated the expression next to them. Keep
only the one carrying the upstream fact — that the API serializes a missing
seat `limit` as `0` — and leave the rest of the reasoning in the commit
message and PR body.
2026-08-06 19:02:48 +05:30
harsh mahajan 2f91099c17 fix(billing): stop a zero seat limit turning free orgs read-only
`getServiceLimit('members')` returned `plan.addons.seats.limit ?? 1`. The
Appwrite starter plan's seats addon omits `limit`, and the API serializes
the missing value as `0` — which is not nullish, so the `?? 1` fallback
never applied and the limit came back as `0`.

`checkForUsageLimit` then read that as "no members allowed":

    membersOverflow = Math.max(0, organization.total - 0)  // always >= 1

so every free-plan organization was flipped to read-only, which silently
disables project creation, member invites and the other write actions
gated on `readOnly`. The explanatory banner never showed either, because
`limitReached.svelte` is gated on `!billingPlanDetails.usage` and the
starter plan's `usage` serializes to a truthy object — leaving a
greyed-out "Create project" button with no reason given.

Normalise a zero limit to `Infinity` in `getServiceLimit` so every caller
agrees, and only treat a real, positive cap as exceedable so a bad limit
can't take an organization read-only again. The UI call sites already
worked around this with `getServiceLimit('members', ...) || Infinity`;
`checkForUsageLimit` was the only consumer that did not.

Plans that set an explicit seat limit (the Imagine plans use 1) are
unaffected, as are Pro, Scale and Education, which declare no seats addon
and already returned `Infinity`.
2026-08-06 18:54:59 +05:30
Harsh Mahajan a41838a86d Merge pull request #3145 from appwrite/fix/delete-processing-message
fix(messaging): show delete for a processing message
8.7.36
2026-08-04 15:25:03 +05:30
Harsh Mahajan b041769cc9 Merge pull request #3146 from appwrite/chore/bump-brace-expansion
chore(deps): raise the brace-expansion floor to 5.0.9
2026-08-04 15:19:02 +05:30
harsh mahajan d3ae09c292 Merge remote-tracking branch 'origin/chore/bump-brace-expansion' into fix/delete-processing-message 2026-08-04 15:16:23 +05:30
harsh mahajan 54f52a894a fix(messaging): gate the delete card on write scope
The card had no permission check, so a reader saw a delete control that
only failed once confirmed. The rest of the feature already derives that
from messages.write: the list gates its create button and row selection
on it.

Gate the card the same way, so it tracks permission rather than status.
2026-08-04 15:13:27 +05:30
harsh mahajan 9bce5ac7bd chore(deps): raise the brace-expansion floor to 5.0.9
GHSA-rgw5-rvv9-x895 widened the affected range to <5.0.9, so the
existing >=5.0.8 override stopped clearing it and the audit step in
tests.yml now fails on every branch. Only the pinned floor and the
resolved version move; integrity hashes are untouched.
2026-08-04 15:10:51 +05:30
harsh mahajan 1f43673992 fix(messaging): show delete for a processing message
The delete card was hidden while a message sat in processing, so a
message a worker never finished could not be removed from the console at
all. The server no longer rejects that delete, so the card can always
render.

Editing stays gated where it already was; only removal is unblocked.
2026-08-04 15:04:42 +05:30
Harsh Mahajan b1ac625e1d Merge pull request #3144 from appwrite/codex/fix-sites-fallback-placeholder
Fix Sites fallback file placeholder
8.7.35
2026-07-31 18:58:50 +05:30
harsh mahajan c5b9428478 fix sites fallback placeholder 2026-07-31 18:26:47 +05:30
Harsh Mahajan 826bccc7f4 Merge pull request #3143 from appwrite/fix-grid-relationship-query-limit
fix: don't fail the grid when relationship resolution exceeds the query value limit
8.7.34
2026-07-31 12:58:36 +05:30
harsh mahajan 218e12864e fix: materialize filter queries so the relationship fallback keeps them
loadGridRows builds its queries twice on the fallback path, and loadPage
passed a Map iterator that the first build exhausted - the retry then ran
without the active filters and paged in rows outside them.
2026-07-31 12:39:17 +05:30
harsh mahajan 1ee99447d8 fix: don't fail the grid when relationship resolution exceeds the query value limit
Tables with wide relationships make the API exceed its own 500 value
limit while resolving them for a full page of rows, failing the whole
listRows request and rendering a 400 error page instead of the table.

Retry the page without relationship selects when that happens, then fill
relationships in over chunks of 10 rows. Rows whose chunk still fails
render without relationship data instead of taking down the page.
Tables that load today are unaffected - the fallback only runs after a
'greater than N values' failure.
2026-07-31 12:35:10 +05:30
Harsh Mahajan 5c5c83e261 Merge pull request #3142 from appwrite/fix-usage-tab-breakdown-crash
fix: prevent project Usage tab crash on missing usage breakdowns
8.7.33
2026-07-27 19:30:24 +05:30
harsh mahajan 7e8011976c fix: prevent project Usage tab crash on missing usage breakdowns
The project Settings > Usage tab failed to render entirely, while every
other Settings tab worked. The API request succeeded; the crash was in
the page component's render:

    Uncaught TypeError: Cannot read properties of undefined (reading 'length')
        at +page.svelte:324

`data.usage.executionsBreakdown` was undefined. The console SDK types it
as a required `MetricBreakdown[]`, so the guard read `.length` on it
unconditionally. Because a throw during render aborts the whole render
pass, nothing mounted -- which is why the tab highlight moved to Usage
while the previously active tab's content stayed on screen instead of an
error page being shown.

Optional-chain both breakdown-table guards (`executionsBreakdown` and
`authPhoneCountryBreakdown`) so a missing breakdown hides just its table.

Also guard two latent instances of the same class on this page, since the
response demonstrably omits array fields the SDK types as required:

- `legendData` reduced over `databasesReads`/`databasesWrites` outside any
  `{#if}`, so it would have crashed even earlier had those been missing.
- The db chart's `{#if dbReads || dbWrites}` passes when only one array is
  present, but the body mapped both.

These are null-guards rather than defaulting the arrays to `[]`, because
`[]` is truthy and would flip several `{#if}` empty-state checks, silently
replacing "No data to show" cards with empty charts.

Separately, the attribution ping in `+layout.svelte` was fire-and-forget
with no `.catch()`, so a blocked request surfaced as an unhandled
`Failed to fetch` rejection that obscured real errors in the console.
2026-07-27 19:24:28 +05:30
Chirag Aggarwal c2c43acf78 Merge pull request #3140 from appwrite/fix/oauth-failure-show-error
fix: show real OAuth error on failure relay page
2026-07-26 11:36:00 +05:30
Chirag AggarwalandCursor 9559127a80 fix: pin postcss and brace-expansion to clear high-severity audit
Co-authored-by: Cursor <cursoragent@cursor.com>
2026-07-25 09:18:58 +05:30
Chirag Aggarwal d4e22eb344 fix: show real OAuth error on failure relay page
When project is missing but an error query param is present, display the actual OAuth failure instead of the misleading "Missing redirect URL" message.
2026-07-24 20:14:09 +05:30
Atharva Deosthale 84eb30dd27 Merge pull request #3138 from appwrite/fix-remove-dev-key-creation
fix(dev-keys): remove dev key creation and add deprecation notice
8.7.32
2026-07-22 19:20:23 +05:30
Atharva Deosthale b2991238c2 fix(dev-keys): state deprecation dates in the alert 2026-07-22 19:16:14 +05:30
Atharva Deosthale d9db09f0fb fix(dev-keys): remove dev key creation, add deprecation notice
Dev key creation is no longer allowed. Removes the create button,
create modal, empty-state create action, and command palette entry,
and adds a deprecation alert linking the changelog:
https://appwrite.io/changelog/entry/2026-07-22
2026-07-22 19:14:12 +05:30
Harsh Mahajan c424116653 Merge pull request #3136 from appwrite/fix-remove-site-deployment-metrics-usage
fix(sites): remove deployment metrics card and bump immutable audit override
8.7.31
2026-07-22 16:21:29 +05:30
harsh mahajan ca1e002230 Merge branch 'fix-audit-immutable' into fix-remove-site-deployment-metrics-usage 2026-07-22 16:17:16 +05:30
harsh mahajan 62867648a7 fix: bump immutable override to clear high-severity audit
bun audit --audit-level high was failing on immutable (pulled in via
sass) for two DoS advisories (32-bit trie overflow, hash-collision
complexity). Bumped the override from ^5.1.5 to ^5.1.8.
2026-07-22 16:15:51 +05:30
harsh mahajan 3924835e96 fix(sites): remove deployment metrics card calling sites.getUsage()
sites.getUsage() is being dropped from the SDK (appwrite/sdk-for-console#105),
along with the other usage endpoints removed in #3134. This inline
build-metrics card on the site deployments page was the one live
consumer that cleanup missed.
2026-07-22 16:13:09 +05:30