- Instagram carousel and Facebook multi-image crop tests now run as datasets
over 1:1 and 4:5, asserting each image's real cropped ratio (previously 1:1
only) — proving the chosen ratio flows through the multi-image paths.
- Add a Facebook test that a multi-image post aborts entirely (throws
FacebookPublishException) when one image can't be downloaded for cropping.
- CropsImageForAspectRatio: drop the per-publisher path-prefix argument; crops
from both platforms go to a single `social-crops/` directory (the folder has
no functional meaning — the platform fetches by URL and nothing reads it back;
UUID filenames prevent collisions).
- Close the IG/FB test asymmetry surfaced in review: Instagram single-image crop
now runs across 1:1/4:5/16:9 (dataset) and has a crop-download-failure test;
Facebook gains an explicit 'original' bypass test. Both platforms now cover the
same matrix (crop per ratio + uploaded-URL is the crop, original/no-meta
bypass, multi-image, download-failure exception).
Follow-up tracked in #83 (orphaned crops are never pruned from storage).
Facebook feed posts now have the same aspect ratio selector as Instagram
(1:1, 4:5, 16:9, Original), letting users control how their image is cropped
before publishing instead of leaving it to Facebook's auto-crop.
The Facebook Graph API has no crop/aspect-ratio parameter for organic Page
photo posts (image crops are ads-only) — Facebook displays the bytes it
receives at their native ratio. So, exactly like Instagram, the crop is done
server-side before upload.
- Extract the Instagram crop logic into a shared CropsImageForAspectRatio
concern (download -> crop -> store -> URL), with a per-publisher
cropFailureException hook. InstagramPublisher refactored to use it
(behaviour identical; existing tests unchanged).
- FacebookPublisher reads meta.aspect_ratio and crops single- and multi-image
feed posts (facebook-crops/ prefix). Reels/Stories are unaffected.
- FacebookSettings: aspect ratio picker on the Feed variant (default Original
-> no crop). ScheduleTab wires meta; FacebookPreview reflects the chosen
ratio. Adds posts.form.facebook.aspect.* in en/es/pt-BR.
Closes#69
Instagram carousels were stored as content_type=instagram_carousel, which the
publisher's match() did not handle — publishing failed with "Unsupported
Instagram content type: instagram_carousel" for any post created via API, MCP,
or template (the AI flow worked only thanks to an inline band-aid that rewrote
carousel to feed before saving).
A carousel is just an Instagram feed post with multiple images: the editor,
preview, and publisher already treat a multi-image feed as a carousel. So
instagram_carousel is a generation format, not a stored content type. Remove it
from the ContentType enum entirely; it now lives only as an AI generation-format
string (wizard card + slide structure + carousel templates are untouched), and
posts always persist as instagram_feed.
- ContentType: drop the InstagramCarousel case; InstagramFeed maxMediaCount 1 -> 10
- StreamPostCreation: resolvedContentType() maps carousel -> feed; band-aid removed
- StartPostCreationRequest: accept instagram_carousel as a generation format
- Frontend: carousel becomes a wizard-local AiFormat; UI/UX unchanged
- API/MCP now reject instagram_carousel as a content_type (Rule::in no longer lists it)
The calendar read each post's caption from post_platforms[0].content,
but post_platforms has no content column — the caption lives on
posts.content. As a result every cell fell back to the 'no content'
placeholder. Read post.content instead (matching the list view) and
drop the phantom content field from the PostPlatform interface.
Also surface the caption in the month view (it previously showed only
the time and platform icons), add horizontal padding to the month post
list so the card border is not clipped inside the overflow-y scroll
container, and standardise the hover effect across day/week/month
views to a shadow-only highlight (no lift).
Closes#70
The Workspace and Brand settings tabs both submit to the same
updateSettings endpoint, but the Workspace tab sends only the name.
UpdateWorkspaceRequest marked brand_font and image_style as required,
so the name-only request failed validation silently (the form only
renders the name error). Mark both brand fields as sometimes|required
so a name-only update succeeds while the Brand tab still validates them.
Closes#74
The Vue computed property returns false for allow_comments when meta lacks
the key (so the UI checkbox renders unchecked), but the backend defaulted
allow_comments to true if the key was absent — sending disable_comment=false
to TikTok. A post created without the user toggling the comment checkbox
went out with comments enabled even though the UI showed them off.
Align the backend default with what the UI displays: when meta omits
allow_comments, treat it as off (matching duet and stitch, which already
default to off). Adjust the existing default-settings test that locked in
the old behaviour.
Match the literal copy the TikTok app reviewer looks for during the
Content Posting API audit, per the Mixpost playbook (the only public
account of how to pass).
- Privacy dropdown placeholder: "Select who can view this post"
- Tooltip on interaction toggles disabled by creator: "Turned off in
your TikTok account settings"
- Render the "may take a few minutes" notice that was translated but
never displayed (processing_hint)
Mirror both i18n changes in pt-BR and es. Add a feature test that locks
the literal strings so they cannot drift before the next review cycle.
Facebook Stories do not work reliably with images via the Graph API.
Align validation and publisher with video-only (like Reel), while
Instagram Story continues to accept image or video.
Co-authored-by: Cursor <cursoragent@cursor.com>
Multi-image posts were published as text-only because Laravel sent
attached_media as JSON; Facebook accepts message but drops attachments.
Use asForm() for all FacebookPublisher Graph API calls.
Co-authored-by: Cursor <cursoragent@cursor.com>
- Replaced the `regenerate_image` boolean with a `change_mode` string parameter in the PostImageRegenerator schema, allowing for more granular control over regeneration options.
- Updated methods in RegeneratePostMediaImage to handle the new `change_mode` values: `image_only`, `text_only`, and `both`.
- Enhanced documentation and comments to clarify the new regeneration logic.
- Added tests to verify correct behavior for different regeneration modes, ensuring that text and image changes are handled appropriately.
- Replaced direct status checks in multiple controllers and actions with the PostStatusGuard utility, improving code readability and maintainability.
- Updated error messages to utilize a centralized method for consistency across the application.
- Removed the BrandImagePalette class, consolidating color resolution logic into the AiImageClient for better organization and type safety.
Let users adjust AI-generated slides in place via async job and Echo, while applying workspace brand, background, and text colors to image prompts. Autofill swaps site text/background colors for the image palette, and regeneration is blocked on finalized posts with safer job cleanup.
Co-authored-by: Cursor <cursoragent@cursor.com>
Add a trypost config toggle to switch between card-required checkout trials and no-card signup trials, and wire signup, checkout, access gating, UI copy, and tests to both modes.
Co-authored-by: Cursor <cursoragent@cursor.com>
Show the checkout-based trial duration again on the subscribe page and rename the signup billing test file so it matches the current no-trial-at-signup behavior.
Co-authored-by: Cursor <cursoragent@cursor.com>
Revert the no-card signup trial flow so access depends on a Stripe subscription trial started at checkout, preventing app access before a payment method is collected.
Co-authored-by: Cursor <cursoragent@cursor.com>
Use Account::postsCountCacheKey everywhere, avoid findOrFail when syncing
post deletes, dispatch SyncAccountUsage after DeleteWorkspace when enabled,
and add coverage for the new paths.
Co-authored-by: Cursor <cursoragent@cursor.com>
- Add separate canDelete predicate in Index.vue that includes Failed —
the previous EDITABLE_STATUSES gating hid the delete button for
failed posts even though the backend allows deleting them.
- Edit.vue Echo handler navigates to /show when an in-page real-time
status update transitions the post into a read-only state, instead
of leaving the user on a stuck readonly editor.
- FacebookPublisher: stop using empty() for content checks (treats
literal "0" as empty) — compare explicitly against null and "".
- Update terminal-state error messages in API + MCP update tool to
reflect the broadened guard (no longer Published-only). Adjust the
matching test assertion.
- PostController@edit was redirecting Failed→show while show was
redirecting Failed→edit, producing ERR_TOO_MANY_REDIRECTS. Failed
posts now render in show.
- New universal `hasContentOrMedia` rule in Edit.vue blocks publishing
when both text and media are empty (closes the hole where empty posts
could reach the publish button).
- Unified `PLATFORM_VARIANTS` to include Facebook, Instagram and
LinkedIn variants. togglePlatform snaps to a compatible variant when
reselecting a platform whose current content_type is incompatible
with the attached media (fixes the case where Reel+image left the
tile permanently blocked).
- platformIssues suppresses the issue on deselected tiles when a
compatible variant exists, so the tile remains clickable and the
snap can recover state.
- Use ContentType enum in place of string literals.
UpdatePost::execute returns Finalized for Published, PartiallyPublished,
Failed, and Publishing — the four "terminal" states the PR introduced.
The existing tests for UpdatePostTool (MCP) and Api/PostController only
exercised the Published path. Convert both to a Pest dataset over all
four statuses so a future regression that lets one state slip through
the check fails loudly.
Same shape already used for PublishPostTool and App/PostController.
The earlier rename of UpdatePost's terminal short-circuit from
AlreadyPublished → Finalized was applied to UpdatePostTool and
Api/PostController but missed PublishPostTool. Result: calling
publish-post-tool against a post in Published / PartiallyPublished /
Failed / Publishing returned the unchanged post wrapped as a success
response — the MCP client thought it had republished.
The check now accepts both action enums so the existing string
behavior (clear error) is preserved, and a Pest dataset covers all
four terminal statuses.
Production incident: a customer's Facebook Page post failed with 'The post
is empty. Please enter a message to share.' (error code 197) and ended up
with a contradictory DB state (status=published + error_message=set).
Three independent bugs were uncovered:
A. FacebookPublisher sends 'message'/'description' as null when the user
posts media without text. Graph API requires the key be omitted, not
null. Fixed in publishSingleImagePost, publishMultiImagePost,
publishVideoPost, publishReel.
B. markAsPublished/markAsFailed leak stale fields across transitions
(a published row could retain error_message from a prior failure,
vice-versa). Both transitions now explicitly clear the opposite
side's fields.
C. status='failed' was editable in the UI and the backend, so users
were re-clicking Publish, generating duplicate failure emails and
the contradictory state from bug B. The frontend isReadOnly check
and the UpdatePost backend guard now treat Published/PartiallyPublished/
Failed/Publishing as terminal. To retry, the user duplicates the post.
11 new tests guarantee these can't regress silently: FB payload shape
per content type, PostPlatform field-clearing on transitions, and the
terminal-status block at the controller level.
Members joining via workspace invite share the owner's account_id.
ProfileController::destroy was unconditionally calling $account->delete()
in every profile-deletion path, so any member could wipe the whole
organization (cascade: workspaces, posts, social accounts, signatures,
labels) just by clicking Delete on their own profile.
Gate the account/subscription teardown behind isAccountOwner(). For
members the path now only detaches them from workspaces and deletes
the user row — owner's data is untouched.
Tested in both SELF_HOSTED=true and false.
- POST /register with ?invite query and no prior session passes the gate.
- UserSeeder seeds the admin + workspace on an empty DB and skips when a
user already exists.
- Login page exposes selfHosted prop in both modes (proves the flag the
frontend uses to hide the "Sign up" link reaches the client).
Self-hosted installs (SELF_HOSTED=true, the default) now close /register
to the public. Workspace invites still work — the AcceptInvite page
links into /register with ?invite={id}, the middleware persists that
into the session, and POST /register passes through.
- EnsureRegistrationEnabled middleware gates GET/POST /register.
Accepts ?invite=… (URL) or pending_invite_id (session) as the pass.
- RegisteredUserController::store clears the marker after signup.
- AcceptInvite.vue passes invite.id in the register link's query string.
- Login.vue hides the "Sign up" link when self_hosted.
- UserSeeder bootstraps a single admin (admin@trypost.it / password).
Idempotent; not wired into DatabaseSeeder — operator runs
`php artisan db:seed --class=UserSeeder` per the install docs.
- Tests cover both flag values for every changed surface.
Docs PR: see trypost-docs self-hosting/installation.mdx step 3.
Audit found three behaviors with no explicit assertion:
- Job is dispatched with a 10-minute delay (would silently regress if
the duration changed). Uses Carbon::setTestNow + Bus::assertDispatched
inspecting \$job->delay.
- error_context.last_attempt_at is recorded at the moment of failure.
- updatePostStatus does NOT finalize the parent Post while any of its
platforms is in Retrying — covers the central invariant of the
feature (the post must stay Publishing until every platform lands
in Published or Failed).
- A platform currently in Retrying transitions to Published when the
next attempt succeeds — proves the loop terminates.
Before: a scheduled post hitting a platform outage was marked Failed —
user had to manually retry. Now the job reschedules itself for 10
minutes later and the PostPlatform shows status "Retrying". Loops
indefinitely until the platform accepts the post.
- Adds PostPlatformStatus::Retrying (existing string column, no migration)
- PublishToSocialPlatform: PlatformUnavailable catch now calls
rescheduleForRetry() which (a) updates the row to Retrying with
retry_count + next_attempt_at in error_context, and (b) dispatches
itself with a 10-minute delay. updatePostStatus() naturally leaves
the parent Post in Publishing because Retrying is neither Published
nor Failed.
- Same treatment for the retry-refresh edge case (publisher throws
TokenExpired, refresh subsequently fails with PlatformUnavailable).
- i18n + frontend status config updated (en, pt-BR, es) for both
posts.status.retrying and posts.edit.status.retrying.
- Tests: 3 new tests covering the dispatched job, the edge case path,
and retry_count increment across attempts.
Edge case from the prior commits: if the publisher throws TokenExpired
(401 path), PublishToSocialPlatform attempts refreshAccountToken() to
recover. That internally goes through ConnectionVerifier::verify, which
can now raise PlatformUnavailable (5xx). The old catch (\Throwable)
swallowed it but the loop still fell through to markAsTokenExpired —
meaning a transient platform outage during a retry could still flip the
account to expired.
Adds an explicit PlatformUnavailable catch in the retry block: marks
the post failed with category platform_unavailable and breaks before
touching the account status.
Same rule applied to production code in earlier commits now applies to
tests: Http::fake patterns and assertions read from
config('trypost.platforms.*.oauth_api' / '.api' / '.default_service')
instead of hardcoded strings. Hardcoded URLs in tests drift silently
when the config changes.
Also documents the rule in CLAUDE.md under "External Service URLs" so
new code (and tests) start in the right place — only the host comes
from config, path/RPC segments stay inline next to the call.
- TokenRedactorTest (7 unit tests): all four regex patterns, multiple
secrets in one body, null input, and the no-op pass-through case.
Guards against silent regression of the redaction regexes (which
previously drifted across three duplicated copies).
- ConnectionVerifierTest: HTTP 429 during refresh raises
PlatformUnavailableException, not TokenExpiredException. Locks in
the rate-limit-as-transient behavior added when consolidating the
refresh logic.
The previous commits in this PR closed the loophole on the hourly /
daily token-refresh jobs. The same loophole remained on the publish
path: every per-platform publisher (LinkedIn, X, YouTube, TikTok,
Threads, Instagram, Pinterest, Bluesky and their Analytics siblings)
has its own refreshToken() called before publishing a scheduled post,
and all of those treated any non-2xx as TokenExpired — including 5xx.
Result before this commit: a Bluesky outage that coincided with a
scheduled publish would mark the account as expired and fail the post.
Changes:
- Route every refreshToken() in the 16 publisher / analytics classes
through TokenRefreshClient::for(Platform::X)->send(...).
- TokenRefreshClient now also fills platformErrorCode from the HTTP
status and pulls error_description / error.message from the JSON
body, preserving the richer info LinkedIn / X / TikTok / Pinterest /
Threads used to put on their TokenExpiredException.
- PublishToSocialPlatform catches PlatformUnavailableException
explicitly: the post is marked failed (category: platform_unavailable,
with http_status in error_context) but the account stays Connected.
No retry inside this job — the scheduler reattempts the next run.
Test added: publish flow does NOT mark account expired when the
publisher throws PlatformUnavailable. Full suite: 1569 passing.
LinkedInPageAnalytics and MastodonAnalytics were the only two of the 11
files refactored to read OAuth host / default instance from config that
had zero test coverage. Adds smoke tests that assert the HTTP request
hits the configured URL, so a typo in the config key (e.g. linkedin.api
vs linkedin.oauth_api) would now fail loudly.
Review follow-ups:
- verifyMastodon was the last hardcoded host left after the PR moved
LinkedIn/YouTube/Bluesky to config. Adds trypost.platforms.mastodon
.default_instance (env MASTODON_DEFAULT_INSTANCE) and reads from it.
- refreshToken() docblock now declares @throws PlatformUnavailableException
(the whole point of the PR was missing from its contract).
- Strip the new explanatory comments inside catch blocks and tests —
rationale lives in the commit / PR, not inline. The two comments
inside empty `catch (TokenExpiredException) {}` blocks stay because
there the comment is the only thing telling the reader why the
exception is swallowed.
When a provider's API was down (5xx, timeout, DNS), the hourly
RefreshSocialToken job and daily VerifyWorkspaceConnections job were
treating it as "token revoked" and emailing the user to reconnect.
Bluesky going offline triggered false-positive disconnect notifications
because Bluesky access tokens are short-lived (2h) so every hourly
refresh failed during the outage.
- New PlatformUnavailableException: API unreachable / 5xx, transient.
TokenExpiredException stays for 4xx (token is provably bad).
- New TokenRefreshClient: normalizes failure semantics for OAuth
refresh HTTP calls across all providers. Takes a Platform enum so
typos fail at compile time and the user-facing label comes from
one source.
- ConnectionVerifier: all 8 refresh*Token methods route through the
new client. Hardcoded OAuth URLs (LinkedIn, YouTube) and Bluesky's
default PDS host moved into config/trypost.php alongside the
existing per-platform entries.
- RefreshSocialToken job: PlatformUnavailableException → log warning
and stop. Do NOT markAsTokenExpired, do NOT notify the user. Next
scheduled tick retries.
- VerifyWorkspaceConnections job: PlatformUnavailableException from
the inner refresh propagates and is treated as a transient skip.
Onboarding/lifecycle workflows in PostHog (and downstream tools like SendKit)
need to segment users by how many social accounts they've connected and how
many posts they've created. The existing SyncUser job only re-emitted these
counts on signup and billing changes, so the values went stale the moment a
user did anything meaningful.
This wires up two new paths that refresh the account group automatically:
- SocialAccountObserver (#[ObservedBy] on the model) fires SyncAccountUsage
on created/deleted, covering all 14 OAuth callback paths in one hook.
- SyncUsageOnPostCreated / SyncUsageOnPostDeleted listeners (auto-discovered)
fire SyncAccountUsage on the corresponding events dispatched by CreatePost
and DeletePost.
SyncAccountUsage is the new dedicated job for group properties only
(groupIdentify account + workspace). SyncUser was slimmed to just identify
the user and delegate the group sync, removing the duplicated property
mapping between the two jobs.
All entry points (observer + both listeners) short-circuit when PostHog is
disabled, so self-hosted instances without PostHog configured see zero
queued jobs and zero overhead.
posts_count cache is invalidated before each sync so the job reads fresh
counts from the database instead of stale cached values.
The post editor lost the Pinterest board picker during a UI rewrite,
causing scheduled posts to fail in production with 'Pinterest board_id
is required'. This restores the picker and locks the contract with
validation + tests so the regression cannot happen silently again.
Backend:
- PostController: pinterestBoards is now Record<account_id, Board[]>
(mirrors the TikTok creator-info pattern); supports multi-account.
- UpdatePostRequest: 'platforms.*.meta.board_id' rule + after-validator
rejects Publishing/Scheduling Pinterest posts without board_id.
Frontend:
- PinterestSettings.vue: Combobox board picker with empty-state warning;
emits update:meta with board_id.
- ScheduleTab / PostEditorSidebar / Edit pass pinterestBoards down by
social_account_id.
Tests (6 new):
- UpdatePostRequestTest: rejects publishing/scheduling without board_id
across pin/carousel/video pin; allows draft without board_id;
pinterest error doesn't block sibling platforms in multi-platform.
- PinterestPublisherTest: publisher throws for carousel + video pin
when no board_id (existing image-pin case kept).
1542 tests passing.
- StripeEventListener::handleSubscriptionCreated nulls account.trial_ends_at
when a Stripe subscription is created. Prevents 'Trial' badge from
lingering for users who convert mid-generic-trial to paid.
- Drop unused trialDays global Inertia prop (no frontend consumers after
/subscribe redesign).
Tests added (10 new, 0 regressions, 1533 total):
- AccountTest: isOnTrial + activeTrialEndsAt across 4 scenarios
(no trial, generic only, subscription only, both — subscription wins)
- StripeEventListenerTest: subscription created clears generic trial
- TrialMiddlewareAccessTest: trialing-with-card subscription passes
- BillingControllerTest: index exposes onTrial/trialEndsAt for the 3
trial states (generic-only, subscription-only, paying); subscribe
page no longer exposes trialDays prop
