* fix(docs): update versioning strategy and tag policies for Docker images
* fix(docs): clarify versioning strategy in environment configuration
fix(docker): update docker-push script to remove local latest tag handling
* fix(docs): clarify VERSION variable usage in Docker push script
* fix(dashboard): show course cards in landing page settings preview
Fetch public courses for the org and pass them to the theme preview
on the landing page settings screen, matching the editor preview.
Co-authored-by: Rotimi Best <[email protected]>
* refactor(dashboard): skip redundant public course fetches
Add loadPublicCoursesIfNeeded so settings, editor, and live landing
page reuse cached courses for the same org site instead of clearing
and refetching on every mount.
Co-authored-by: Rotimi Best <[email protected]>
* fix(dashboard): always refetch courses on live org landing page
Keep loadPublicCoursesIfNeeded for settings preview and editor only.
The public-facing landing page should fetch fresh on each visit.
Co-authored-by: Rotimi Best <[email protected]>
* feat(ui): add shared empty state for org landing page catalogs
Introduce OrgLandingPageCoursesEmpty using the dashboard Empty pattern
with landing-theme colors. Show it across all ten org landing templates
when no courses are published, and wire translated labels through
buildOrgLandingPageProps.
Co-authored-by: Rotimi Best <[email protected]>
* fix(dashboard): address Greptile review on public courses cache
- Add coursesLoaded prop so empty catalog state waits for fetch
- Deduplicate concurrent public course requests via in-flight promise
- Refresh courses on settings mount; reuse cache in editor only
- Expose invalidatePublicCourses for explicit cache clearing
Co-authored-by: Rotimi Best <[email protected]>
* fix(dashboard): load landing preview courses when org store hydrates
Replace onMount course fetch with a reactive effect keyed on the route
slug and currentOrg.siteName so public courses load after app init.
Mark the fetch as complete in finally so failed requests show the empty
state instead of a permanently blank catalog section.
Co-authored-by: Rotimi Best <[email protected]>
---------
* fix(website): restore radial hero backgrounds on page headers
PageHeader lost its decorative background when hero-bg was replaced by
DotField. Add a bottom-centered radial glow via a CSS class so it renders
reliably without Tailwind arbitrary-value parsing issues.
Also move the signup CTA gradient from an inline arbitrary Tailwind class
to the same CSS-class pattern for consistency.
Co-authored-by: Rotimi Best <[email protected]>
* fix(website): restore dotted DotPattern hero background on PageHeader
Revert the radial glow added to page headers and bring back the static
DotPattern grid that was replaced by DotField in c31ad6395.
Co-authored-by: Rotimi Best <[email protected]>
---------
Co-authored-by: Cursor Agent <[email protected]>
Co-authored-by: Rotimi Best <[email protected]>
* fix(api): return Zod validation errors from public API routes
Public API handlers now surface ZodError details as 400 VALIDATION_ERROR
responses instead of generic 500 INTERNAL_ERROR payloads. Course import
services also rethrow ZodError so validation failures are not wrapped as
server errors.
Co-authored-by: Rotimi Best <[email protected]>
* refactor(utils): centralize course type values without db dependency
Define COURSE_TYPE_VALUES once in @cio/utils/constants and use it for
ZCourseType validation and the Drizzle COURSE_TYPE pgEnum. Removes the
circular @cio/utils -> @cio/db import while keeping course-import schemas
aligned with all four course types including PUBLIC.
Co-authored-by: Rotimi Best <[email protected]>
* fix(validation): default missing course metadata allowNewStudent to true
Courses stored without allowNewStudent in metadata failed structure export
validation. Align Zod preprocessing with runtime behavior where absent
values mean enrollments are open.
Co-authored-by: Rotimi Best <[email protected]>
* chore(db): drop unused COURSE_TYPE_VALUES re-export from schema
Co-authored-by: Rotimi Best <[email protected]>
---------
Co-authored-by: Cursor Agent <[email protected]>
Co-authored-by: Rotimi Best <[email protected]>
Embed iframes were placed inside flex column layouts without shrink
protection, so authored width/height attributes could be ignored when the
preview panel was narrower than the iframe. Parse dimensions from pasted
embed code and size a dedicated host wrapper so the iframe renders at the
intended dimensions while still scaling down with max-width: 100%.
Co-authored-by: Cursor Agent <[email protected]>
Co-authored-by: Rotimi Best <[email protected]>
* fix(dashboard): API page snippet highlighting and breadcrumb i18n
Stop double-translating nav titles in breadcrumbs (t was called on
already-localized strings like Audience and API).
Use Code.Root default typescript highlighting for all API quick-start
snippets instead of unsupported bash/python/go langs.
Co-authored-by: Rotimi Best <[email protected]>
* revert(dashboard): drop breadcrumb i18n change from API page PR
Translation warnings were unrelated; keep only the Shiki lang fix.
Co-authored-by: Rotimi Best <[email protected]>
---------
Co-authored-by: Cursor Agent <[email protected]>
Co-authored-by: Rotimi Best <[email protected]>
* feat(api): purge Cloudflare CDN before and after OpenAPI spec upload
Purge api.cdn.clsrio.com edge cache for openapi-latest and the dated
spec URL so deploys are not blocked by the 1h Cache-Control TTL on R2
objects. Uses the same CLOUDFLARE_API_TOKEN and CLOUDFLARE_ZONE_ID
pattern as embed uploads; skips when credentials are unset.
Co-authored-by: Rotimi Best <[email protected]>
* fix(api): purge OpenAPI CDN cache only after upload
Match the embeds upload pattern: update R2 first, then invalidate
edge cache so clients refetch the new spec.
Co-authored-by: Rotimi Best <[email protected]>
* fix(api): treat Cloudflare purge 2xx as success when body is unparseable
Avoid false upload failures when the purge API returns 200 with a
non-JSON body. Addresses Greptile review on #704.
Co-authored-by: Rotimi Best <[email protected]>
---------
Co-authored-by: Cursor Agent <[email protected]>
Co-authored-by: Rotimi Best <[email protected]>
* feat(cohorts): rename Programs to Cohorts with data migration
- Add cohort database tables and Drizzle migration alongside legacy program tables
- Introduce migrate-programs-to-cohorts script to copy existing program data safely
- Rename API routes (/cohort), services, queries, validation, and error codes
- Rename dashboard feature module, routes (/cohorts), navigation, and translations
- Update docs, emails, and internal compliance cron endpoints for cohort terminology
- Keep legacy program tables in place for rollback until explicitly removed
Co-authored-by: Rotimi Best <[email protected]>
* fix(db): restore migrate:answer-data script alongside cohort migration
Co-authored-by: Rotimi Best <[email protected]>
* fix(db): cast program enums when migrating data to cohort tables
Co-authored-by: Rotimi Best <[email protected]>
* fix(db): truncate cohort tables before program-to-cohort copy
Co-authored-by: Rotimi Best <[email protected]>
* fix(db): report accurate row counts in programs-to-cohorts migration
Co-authored-by: Rotimi Best <[email protected]>
---------
Co-authored-by: Cursor Agent <[email protected]>
Co-authored-by: Rotimi Best <[email protected]>
* fix(automation): align public API plan gating across cloud and self-hosted
Add shared canUsePublicApi helper so cloud allows Early Adopter and
Enterprise while self-hosted requires Enterprise. Update the API key
dashboard page and backend key creation checks to use the same rules.
Co-authored-by: Rotimi Best <[email protected]>
* docs(api): document bearer auth on public API reference
Add shared OpenAPI enrichment with bearerAuth security scheme and
authentication instructions. The interactive reference now loads an
enriched spec with Authorize support, and the API guide explains how
to create and use org API keys.
Co-authored-by: Rotimi Best <[email protected]>
* feat(dashboard): add API quick-start code snippets on automation page
Mirror the MCP client setup section with cURL, JavaScript, and Python
examples that show how to authenticate public API requests with org keys.
Co-authored-by: Rotimi Best <[email protected]>
* ci(openapi): republish spec when public API OpenAPI helpers change
Co-authored-by: Rotimi Best <[email protected]>
* feat(dashboard): add Go snippet to API quick-start examples
Co-authored-by: Rotimi Best <[email protected]>
* docs(api): use /org/*/api and clarify dashboard key generation path
Co-authored-by: Rotimi Best <[email protected]>
* fix(api): escape backticks in OpenAPI description template literal
Greptile flagged that unescaped backticks around * in
PUBLIC_API_OPENAPI_DESCRIPTION evaluated to NaN and dropped auth docs
from the Scalar spec. Also use plan-specific snackbar message on upgrade.
Co-authored-by: Rotimi Best <[email protected]>
---------
Co-authored-by: Cursor Agent <[email protected]>
Co-authored-by: Rotimi Best <[email protected]>
- Add Types section to the Filter sidebar with all 4 course types
- Add full-text search input centered in the hero (title + description)
- Add Pricing section (Free / Paid) based on course cost
- Compact hero prop across all 10 themes to halve vertical padding
- Children snippet support in all hero themes for the search input
- All filters are URL-driven (?types=, ?search=, ?tags=, ?pricing=)
- Full-stack: validation schema, DB queries, service, API route, server load, UI
The embed section passed code through SafeHtmlContent which ran DOMPurify
with FORBID_TAGS including 'iframe', silently removing the element.
Replaced SafeHtmlContent with a new sanitizeEmbedHtml() that uses a
whitelist of ALLOWED_TAGS: ['iframe'] and restricts src to https: only,
so Google Forms, YouTube, and other https embeds render while javascript:
sources and non-iframe tags are still blocked.
Items with isUnlocked = null were evaluated as teacher-locked due to
`?? false` defaulting to false before negation. This caused content to
appear locked immediately after a student marked a lesson complete, even
in free-mode courses.
Changed to `?? true` in both client-side (progression-utils.ts) and
server-side (core/progression.ts) so null/undefined isUnlocked defaults
to unlocked — consistent with the sidebar's own `?? true` guard.
* feat: make upload file size limits configurable via .env
Add UPLOAD_MAX_*_MB environment variables for documents, images, videos,
exercise assignment files, AI assistant documents, landing page images,
and media thumbnails. API presign routes validate file size server-side;
dashboard UI reads the same limits from layout data.
* chore: wire upload limit env vars through Docker compose and docs
Pass UPLOAD_MAX_*_MB from the root .env into api and dashboard containers
in both compose files, document them in .env.example and self-hosting guides.
Co-authored-by: Rotimi Best <[email protected]>
* docs: cross-link upload limits in README, Railway, Coolify, and Cloudflare guides
Co-authored-by: Rotimi Best <[email protected]>
* fix: address PR review feedback on upload limits
- Restore 5 MB default for thumbnail uploads (edit-asset-dialog regression)
- Add getResolvedUploadLimits() fallback for Storybook/tests
- Cache core getUploadLimits() like the dashboard helper
- Document presign fileSize as advisory; allow zero-byte fileSize in schema
- Only send fileSize to presign when greater than zero
* docs: remove upload limits from Cloudflare setup guide
---------
- Add .odt (application/vnd.oasis.opendocument.text) as a supported upload
format; extract text via JSZip + content.xml parsing
- Add attach_document_to_lesson agent tool so the AI can attach an uploaded
document to a lesson as a downloadable material
- Instruct the agent to create a "Course Materials" lesson (first lesson of
the first section) and attach the uploaded document to it on implementation
- Suppress hallucinated references: uploaded documents do not count as fetched
URLs — omit the References section when no fetch_documentation_url calls
were made
- Surface specific upload error snackbars (file_too_large, unsupported_file_type)
instead of silently logging; fix generic fallback message
- Allow typing in the chat textarea while a document is uploading; only block
Enter/send, not the whole input
- Fix retry endpoint to return the current run instead of erroring when the run
is already active (queued/running/waiting_for_input/paused)
- Connect @cio/core Redis singleton in the jobs bootstrap so course-generation
workers no longer fail with "The client is closed"
- Expose documentAssets (assetId + fileName) alongside documentText in
AgentContext and pass it through both the chat route and course runner
* feat(lms): add student certificates page and sidebar nav
Add a Certificates item to the student LMS sidebar that lists courses
where the student has earned a certificate. Each course card includes a
View Certificate action linking to the existing per-course certificate
page.
* refactor: reuse isOrgOnFreePlan helper from org store
* chore(i18n): add certificates translations for all locales
---------
exercise.svelte was auto-adding a QUESTION_TEMPLATE to the store whenever
questions were empty, which raced with hydrateExercisePageData.
applyMySubmission fired with the blank template in place, set
isFinished=true, and the guard prevented re-application once real
questions loaded.
Two fixes: guard the auto-add effect with isOrgStudent (students never
need to add questions), and gate applyMySubmission on isExerciseLoaded
so it always waits for hydrateExercisePageData to populate real questions
first.
Also fix off-by-1 in applyMySubmission (questions.length → questions.length + 1)
so currentQuestion is undefined at end-of-exercise, satisfying the
isFinished derived. Clear questionnaire.questions in onDestroy to prevent
stale questions flashing on the next exercise.
Fix lesson delete not updating sidebar/content list by filtering the
removed lesson from sections on success.
When a lesson has two or more uploaded videos each calls mintHlsCookie,
but all wrote to the same `cio_hls` cookie name. The last write won,
leaving every other video with a mismatched token and a 403 from the
CF Worker.
Cookie name is now `cio_hls_${assetId}` so each video owns its own
cookie. The tenant-router Worker, the Hono HLS fallback route, and the
transcripts route all read the per-asset name.
Replace the per-handler boolean guard with a shared 300ms cooldown
timestamp so all three Plyr seek-enforcement paths (progress-bar
listener, seeking, seeked) produce at most one snackbar per skip
attempt.
The registry hardcoded numeric ids that drifted from the database
`question_type` table: MATCHING and HOTSPOT were inserted mid-list
(ids 9 and 11) but never seeded, shifting ORDERING/LINK/WORD_BANK/
STAR/VIDEO_RECORDING one or two ids past their real rows. Every id
emitter inherited the wrong values — the exercise editor, the AI
question generator, and the grading/scoring gate — so e.g. an ORDERING
question was stored as id 10 (LINK in the DB), came back as a manual
type, and never auto-graded.
- Move MATCHING/HOTSPOT to the end (ids 14/15) and flag them
`disabled`, restoring the append-only id convention and realigning
ORDERING=9, LINK=10, WORD_BANK=11, STAR=12, VIDEO_RECORDING=13 with
the DB.
- Add a registry test asserting every enabled type's id matches the
canonical `question_type` rows and disabled types stay past the live
range, so ids can't silently drift again.
- Constrain the agent tool schema to enabled, DB-backed ids so the
model can't emit a disabled id with no row.
- Drop disabled types from the teacher prompt's type list and guidance.
- Centralize the dashboard question-type picker on the `disabled` flag.
This also fixes video-recording detection, which compared against the
stale id 15 instead of the real 13.
Generated migration for the narrowed
asset_usages_asset_target_slot_unique constraint (asset_id, target_type,
target_id, slot_type, slot_key). Run the duplicate cleanup before
applying — existing rows that differ only by position will otherwise
violate the new constraint.
Drop `position` from the `asset_usages` unique constraint so the same
asset can't be attached to one target/slot more than once, and update
the insert's onConflict target + lookup to match the narrower key.
Guard it at the service layer too: attachAssetService now rejects an
attach when the asset already has a usage for that target with a new
ASSET_ALREADY_ATTACHED (409) error, backed by an
assetUsageExistsForTarget query.
Add an admin flow to permanently delete a media asset and clean up
everything it leaves behind.
- Delete dialog with a destructive, irreversible confirmation; deletion
is blocked while the video is still attached and shows the in-use
lessons so the admin can detach first.
- Usage list resolves lesson/course context for each attachment, links
to the lesson editor via an anchor (right-click "open in new tab"),
and exposes a refresh button to re-check usage after detaching
elsewhere.
- Storage cleanup runs as a background maintenance job: the producer
captures the asset's object keys/prefixes up front and an idempotent
worker sweeps HLS segments, thumbnails, audio, transcript, and the
raw upload from object storage.
Also harden usage accounting so stale attachments cannot trap an asset:
- asset_usages has no FK to its polymorphic target, so prune a lesson's
usage rows whenever the lesson is hard-deleted (single + bulk paths).
- Usage resolution drops orphaned lesson usages (lesson gone or course
soft-deleted) so they neither appear in the list nor block deletion.
Document the icon-only button convention (variant="secondary") in
AGENTS.md.
The env module was moved to @cio/core/config/env but the upload-openapi-spec
script still imported from the removed apps/api/src/config/env path, causing
CI upload:openapi to fail with ERR_MODULE_NOT_FOUND.
Co-authored-by: Rotimi Best <[email protected]>
* feat(course): add reset student course progress action
Give course admins and tutors a destructive reset-progress flow on the
student people detail page. The backend clears learner progress for one
(courseId, memberId) pair in a single transaction while keeping the
groupmember enrolled, and logs an audit summary. The dashboard shows a
confirmation dialog with impact counts from analytics before deleting.
Co-authored-by: Rotimi Best <[email protected]>
* fix(i18n): add localized reset progress strings for all locales
Co-authored-by: Rotimi Best <[email protected]>
* fix(course): scope reset progress to course data only
Stop deleting org-level community questions and answers during course
progress reset. Fix course newsfeed deletion order so comments on a
student's posts are removed before the posts themselves. Restrict reset
to student members, move analytics loading to the person detail layout,
and rename impact copy to course newsfeed only.
Co-authored-by: Rotimi Best <[email protected]>
* fix(course): address Greptile review on reset progress
Make progressImpact opt-in via includeProgressImpact query param so
analytics requests avoid extra DB work unless the reset dialog needs it.
Count only student-authored course newsfeed posts and comments in the
impact summary to avoid double-counting replies on the student's posts.
Co-authored-by: Rotimi Best <[email protected]>
---------
Co-authored-by: Cursor Agent <[email protected]>
Co-authored-by: Rotimi Best <[email protected]>
Disabled items were overriding props.class with a custom class string,
stripping the Sidebar.MenuButton's height, padding, and color styles.
Drop the inline class and rely on the existing aria-disabled CSS rules.
* fix(dashboard): course menu icons, delete scroll, and settings tags auth
Remove inconsistent dropdown icons, scroll to settings sections for delete
and related actions when already on the settings page, and skip org tag
group fetches for non-admin users to avoid ORG_TEAM_NOT_AUTHORIZED errors.
* fix(api): allow org members to fetch organization tags
Use orgMemberMiddleware for GET /organization/tags instead of restricting
reads to admins. Revert the dashboard settings workaround that skipped tag
group fetches for non-admin users.
---------
Disabled items were overriding props.class with a custom class string,
stripping the Sidebar.MenuButton's height, padding, and color styles.
Drop the inline class and rely on the existing aria-disabled CSS rules.
Add View course site and Publish course options to org course context menus.
Replace the course header View as student button with View Course Site and
a companion overflow menu that includes View as student plus list actions.
Show a publish prompt modal for unpublished courses and support #publish
deep links in course settings.
* feat: enhance self-hosting support and CI/CD workflows
- Updated .env.example to clarify auth secret generation and self-hosting mode.
- Added CI workflow for build and lint checks on PRs and main branch pushes.
- Improved docker-publish workflow to trigger on relevant file changes and added multi-arch support.
- Enhanced upload-openapi-spec workflow with caching for dependencies.
- Removed obsolete Dockerfile for API service and added jobs worker to Docker setup.
- Updated self-host documentation to reflect new jobs worker requirement.
- Modified Dockerfiles for API and Dashboard to optimize dependency fetching and installation.
- Enhanced docker-compose.yaml with health checks and jobs worker configuration.
- Updated docker-push.sh to support multi-architecture builds and improved logging.
- Revised documentation for publishing images and self-hosting to include new features and requirements.
* feat(self-host): safe versioned releases + pull-based image pipeline
Builds on the pull-based images compose (docker-compose.images.yaml) to make
pulling pre-built images safe for self-hosters:
- docker-publish: stable-releases + edge tag policy. Git tag vX.Y.Z publishes
X.Y.Z/X.Y/X/latest; a push to main publishes only :edge (latest never moves
on a routine merge).
- Pre-publish smoke-test gate: each image is booted against ephemeral
postgres+redis and healthchecked before any push (docker/smoke-test.sh).
- CIO_VERSION pinning in .env.example; run-docker-full-stack.sh --images flag
to pull instead of build.
- Docs: SELF_HOST pull-from-images + versioning/upgrade/rollback sections,
PUBLISHING_IMAGES tag policy + pnpm release flow, self-hosting.mdx quickstart.
* feat: enhance self-hosting support and CI/CD workflows
- Updated .env.example to clarify auth secret generation and self-hosting mode.
- Added CI workflow for build and lint checks on PRs and main branch pushes.
- Improved docker-publish workflow to trigger on relevant file changes and added multi-arch support.
- Enhanced upload-openapi-spec workflow with caching for dependencies.
- Removed obsolete Dockerfile for API service and added jobs worker to Docker setup.
- Updated self-host documentation to reflect new jobs worker requirement.
- Modified Dockerfiles for API and Dashboard to optimize dependency fetching and installation.
- Enhanced docker-compose.yaml with health checks and jobs worker configuration.
- Updated docker-push.sh to support multi-architecture builds and improved logging.
- Revised documentation for publishing images and self-hosting to include new features and requirements.
* ci: run build-only on PRs, defer lint until repo-wide format
* docs: point Railway template link at the new template URL
* fix: address Greptile review + dashboard CI build env
- docker-push.sh: only also-tag :latest for real semver releases, so
VERSION=edge/main/dev can't clobber the stable latest pointer (P1).
- run-docker-full-stack.sh: stop deriving OBJECT_STORAGE_PUBLIC_ENDPOINT from
the dashboard origin (broke presigned URLs for videos/documents buckets);
derive only the public media base URL and document the S3-API route.
- ci.yml: provide PUBLIC_IS_SELFHOSTED at build time (dashboard reads it from
SvelteKit's strict $env/static/public); bump cache actions v3 -> v4.
- upload-openapi-spec.yml: bump cache actions v3 -> v4.
- smoke-test.sh: dashboard gate now fails on 5xx instead of accepting any response.
* ci: scope PR build gate to the api + dashboard closures
Building every package surfaced pre-existing rot in apps that aren't part of
this PR (storybook's missing @lucide/svelte import, course-app prettier debt).
The gate's purpose is to verify the two Docker-image apps; the rest are built on
PRs by their own Cloudflare Workers builds. Filter to @cio/api... and
@cio/dashboard... so the gate covers exactly what it should.
* fix(docker-push): make latest_arg always return 0
Use an `if` instead of `is_release_version && printf` so the helper returns 0
even when it emits no tag. A non-zero return inside the $(latest_arg ...) build
argument is fragile under set -e (and would abort with inherit_errexit), which
would break the edge/main builds the policy explicitly supports.
* fix(docker): copy patches/ into jobs image so pnpm install resolves patchedDependencies
The jobs Dockerfile ran `pnpm install` without copying patches/, so pnpm
failed hashing package.json#pnpm.patchedDependencies (layerchart patch),
breaking the self-host full-stack build. api/dashboard already copy patches/.
Co-Authored-By: Claude Opus 4.8 <[email protected]>
* fix(docker): build jobs worker deps in topological order
The jobs Dockerfile built workspace packages with a hand-written list that
omitted transitive deps (e.g. @cio/question-types, consumed by @cio/utils),
so `tsc` failed with TS2307 during the self-host build. Use the recursive
`pnpm -r --filter @cio/jobs-worker... build` like the api/dashboard images,
which builds the full dependency closure in the correct order.
* refactor(docker): move jobs Dockerfile to docker/Dockerfile.jobs and align dep install
Move apps/jobs/Dockerfile -> docker/Dockerfile.jobs so all three image recipes
live together in docker/, and update every reference (docker-compose.yaml,
docker-push.sh, docker-publish.yml, railway.jobs.json, self-hosting.mdx).
Also align its dependency install with the api/dashboard images: replace the
hand-maintained list of workspace package.json copies (which silently missed
transitive deps like @cio/core / @cio/ai-assistant, leaving their node_modules
uninstalled and breaking the build with TS2307) with the proven pnpm-fetch
pattern: fetch the full lockfile into a cached store, COPY . ., then an offline
filtered install of @cio/jobs-worker... This installs the complete dependency
closure regardless of how the workspace graph grows.
* refactor(docker): move docker-compose.yaml to repo root for Coolify compatibility
Coolify's Docker Compose build pack forces Base Directory to the repo root and
resolves build contexts from there, so the compose's `context: ..` (needed when
the file lived in docker/) climbed above the repo and failed with
`lstat /artifacts/docker: no such file or directory`.
Move docker/docker-compose.yaml -> ./docker-compose.yaml and change the build
contexts from `..` to `.` so the context is the repo root with no upward path.
This works identically for run-docker-full-stack.sh, manual `docker compose up`,
and Coolify. Dockerfiles stay referenced as docker/Dockerfile.*. Update the
references in run-docker-full-stack.sh and the docs accordingly. The images
compose (docker-compose.images.yaml) is unaffected (no build context) and stays
in docker/.
* Add self-hosting documentation for Coolify and Railway
- Created new documentation for deploying ClassroomIO on Coolify using a pull-based Docker Compose stack.
- Added comprehensive guide for self-hosting ClassroomIO with Docker, covering both pre-built images and building from source.
- Introduced a new meta.json file to organize self-hosting documentation pages.
- Developed a Railway deployment guide with environment variable configurations for API, dashboard, and jobs worker.
- Updated self-hosting blog post to link directly to the new Railway guide.
- Removed obsolete Railway configuration files (railway.api.json, railway.dashboard.json, railway.jobs.json).
- Enhanced run-docker-full-stack.sh script to improve MinIO credential generation and service readiness checks.
- Updated docker-compose files to enforce strong secret requirements for authentication.
- Adjusted self-hosting references in the Docker documentation to include new guides.
* fix(docker): point references at root-level docker-compose.images.yaml
The pull-based compose file moved to the repo root (alongside docker-compose.yaml).
Update IMAGES_COMPOSE_FILE in run-docker-full-stack.sh, the file's own header
example, and the docs (docker.mdx curl URL, coolify.mdx path + blob link,
SELF_HOST.md relative link + curl URL).
* fix(self-host): address review round 2 — drop unused AUTH_BEARER_TOKEN, pull-by-default, doc fixes
- Remove AUTH_BEARER_TOKEN everywhere (unused in app code; PRIVATE_SERVER_KEY is the real
dashboard<->api key). Simplifies run-docker-full-stack.sh secret generation.
- run-docker-full-stack.sh now defaults to PULLING pre-built images; add --build for source.
- docs: recommend MinIO, 'a way' (not 'recommended way') to self-host, remove Direct API Access
section, trim Common Commands, SMTP vars per-line, fix PRIVATE_SERVER_KEY wording.
* feat(db): versioned migrations for self-host with auto-baseline
- Regenerate a single clean baseline migration; db:setup now runs drizzle-kit migrate (not push).
- Add db:generate / db:migrate / db:baseline scripts.
- baselineMigrationsIfNeeded(): adopt an existing (push/supabase) schema into drizzle's journal so
migrate won't try to recreate existing objects ("type/table already exists"). No-ops on fresh
and already-tracked databases; runs automatically in db:setup before migrate.
* ci(openapi): drop stale ci.yml reference in turbo-cache comment
ci.yml was removed earlier in this review; the openapi workflow's turbo cache still
warms from its own prior runs via restore-keys, so this is a comment-only fix.
* fix(db): make baseline squash-safe for migrate-tracked databases
baselineMigrationsIfNeeded returned early whenever drizzle.__drizzle_migrations
had any rows. After a migration squash, an instance tracked under the old
journal (pre-squash hashes) would then let `drizzle-kit migrate` re-run the
consolidated baseline — whose journal `when` is newer than every old created_at
— and crash on `CREATE TYPE/TABLE "..." already exists` at startup.
Now, when history exists but the consolidated baseline hash is absent, insert
that hash (marking it already applied) so migrate skips it. Only the baseline
entry (idx 0) is reconciled, so genuinely pending future migrations still run.
Verified end-to-end against Postgres: reproduces the original crash, confirms
the fix no-ops migrate on a pre-squash DB, and confirms a later 0001 migration
still applies. Fresh-DB and already-adopted paths unchanged.
* fix(docker): add explicit id to pnpm-store cache mounts
Railway's Metal builder rejects BuildKit cache mounts that omit an
explicit id (it does not default id to target like standard Docker/
BuildKit do), failing the build at parse time with:
dockerfile invalid: flag '--mount=type=cache,target=/pnpm/store'
is missing an id argument
Add id=pnpm-store to all six cache mounts across the api, dashboard,
and jobs images. Harmless on Docker Desktop / CI / GitHub Actions, and
required for the Railway self-host build path.
* fix(docker): drop BuildKit cache mounts for Railway compatibility
Railway's builder rejects cache-mount ids not prefixed with its
per-service cacheKey, which can't be hardcoded in a Dockerfile shared
across CI, local Docker, and three Railway services. Remove the
--mount=type=cache flag (keep --store-dir /pnpm/store); the store is
baked into the fetch layer and reused by the offline install, so
behaviour is unchanged. Supersedes 9c4633f2c.
* fix(docker): force LF for shell scripts and Dockerfiles via .gitattributes
cio-api crash-looped on Railway with 'exec /app/entrypoint.sh: No such
file or directory' because the entrypoint shipped with CRLF endings
(Windows checkout, core.autocrlf=true) — the '#!/bin/sh\r' shebang is
unresolvable in the container. Add .gitattributes pinning *.sh and
Dockerfile* to LF so container-executed files are never CRLF regardless
of contributor OS.
* fix(db): serialize db-setup with an advisory lock to prevent baseline double-insert
Concurrent db:setup runs (e.g. multiple API replicas booting together) could both
pass the baseline 'already present?' check and insert duplicate rows into
drizzle.__drizzle_migrations, which has no unique constraint on hash. Wrap baseline
+ migrate in a pg_advisory_lock held on a dedicated connection so concurrent
startups serialize: the loser blocks, then sees the baseline applied and no-ops.
* docs(railway): update template deploy URL to classroomio-1
---------
Co-authored-by: Rotimi Best <[email protected]>
Clamping currentTime in the seeking/seeked handlers is ignored on progressive
MP4 (the forward seek takes effect immediately), so the must-watch lock leaked
even though the snackbar showed.
Block at the source: a Plyr `seek` listener returns false for targets past the
watched point, which skips Plyr's default seek so the media never moves.
Rewinding and re-seeking up to the watched point stay allowed; the snackbar
fires when a forward skip is blocked. furthestSeconds is now shared between the
listener and the playback tracker. The seeking/seeked clamp is kept as a
backstop for keyboard/programmatic seeks that bypass the progress bar.
* fix(analytics): show per-student last seen from login and session activity
Return ISO last-seen timestamps from course analytics instead of
pre-formatted strings so the dashboard can format them per student.
Batch-fetch last seen for all students and include session activity
alongside login events for more accurate, distinct values.
Co-authored-by: Rotimi Best <[email protected]>
* fix(analytics): destructure first row from getLastLogin query results
Promise.all returns an array per Drizzle select; nested destructuring
is required so loggedInAt and updatedAt are read from row objects.
Co-authored-by: Rotimi Best <[email protected]>
---------
Co-authored-by: Cursor Agent <[email protected]>
Co-authored-by: Rotimi Best <[email protected]>
The lock clamped currentTime only during the seeking event. HLS defers the
forward seek until the fragment loads, so that clamp caught it; but for
progressive MP4 the forward seek takes effect immediately and the mid-seek
write is ignored, so playback leaked past the watched point (snackbar showed,
no snap-back).
Keep the immediate seeking-time clamp (HLS relies on it to avoid buffering
ahead) and add an authoritative re-clamp in seeked: if the playhead still
landed past the limit, snap back via a backward seek into buffered content,
which holds. Snackbar now fires once per blocked gesture and only when a
clamp actually happens.
Resetting currentTime inside the seeking handler doesn't reliably stick
across sources (HLS re-buffering, emulated YouTube events, coalesced native
seeks), so the snackbar showed but playback continued ahead. The timeupdate
handler then ignored the large forward jump via its delta < 2 guard, making
the skip permanent.
Re-clamp in seeked once the position settles, and add a timeupdate backstop
that snaps back forward jumps (delta >= 2) landing past furthestSeconds.
Rotimi Best