* fix: Facebook Page fetch missing New Pages Experience pages
/me/accounts silently omits Pages that live under Meta's newer
"New Pages Experience" / Business Portfolio model, even when the
token's granular scopes show the Page was explicitly granted -
confirmed via Meta's own Access Token Debugger against a live
account whose Page returned zero results from /me/accounts but
resolved fine when queried directly by ID.
Falls back through Business Manager's owned_pages/client_pages
(via the existing business_management scope) when /me/accounts
comes back empty, so Pages under that model are still found.
* fix: Instagram-via-Facebook has the same New Pages Experience gap
Same root cause and fix as the FacebookController fetchPages()
fallback - the Page/Instagram-linked-Page lookup goes through the
same /me/accounts call and is subject to the same Meta-side gap.
* refactor: one place finds every Page a Meta login can publish to
Both controllers walked /me/accounts and, when it came back empty, the
Business Portfolio edges behind it. ManagedPages now owns that walk for
the Facebook and Instagram-via-Facebook flows alike.
Three behaviour changes come with it:
The portfolio edges are read on every connect, not only when /me/accounts
is empty, and merged by Page id. Someone holding one Page by a classic
role and the rest through a portfolio was auto-connected to that single
Page and never offered the others.
/me/businesses is read only once /me/permissions confirms the login
granted business_management, and a failure anywhere along the portfolio
walk leaves the /me/accounts list standing. It used to escape into the
callback's catch, turning "no pages" into "could not connect" for every
login without the scope.
Pages the login cannot get an access_token for are dropped. Connecting
one produces an account that cannot publish.
* test: pin the Page a login only reaches through a portfolio
Covers the merge with /me/accounts, the access_token filter, the
business_management gate, and a failing portfolio edge leaving the
/me/accounts list intact — plus the connect flow end to end on both
Meta platforms.
The Instagram-via-Facebook request count moves from five to six for
the /me/permissions check.
* refactor: read the portfolio edges without asking permission first
The /me/permissions check saved a rejected /me/businesses call on logins
without business_management, at the cost of running a path nobody had
verified against a live account. The edges already fail soft, so the
check bought log tidiness and nothing else.
* test: cover the portfolio walk's remaining shapes
The multi-Page selection flow behind a portfolio — the case #292 asked a
maintainer to check — plus pages spread across two portfolios, a
paginated edge, a portfolio entry with no id, and a portfolio page
merging with one /me/accounts already returned.
* test: stop the Meta connect tests from calling Graph for real
Http::fake only stubs the URLs it is given; anything else goes out over
the network. These files stubbed /me/accounts and left /me — and now
/me/businesses — unstubbed, so the suite was issuing live requests to
graph.facebook.com on every run. They came back 400 and the code under
test swallowed them, so nothing ever went red while the assertions were
measuring Meta's answer instead of the fixture's.
Every Graph call the connect flow makes is stubbed now, and the files
prevent stray requests so a missing one fails loudly. Inertia's SSR
endpoint is allowed through; it is not what these tests are about.
* fix: tell a denied portfolio edge apart from a throttled one
GraphPaginator throws so no caller reads a failed fetch as an empty list
and auto-connects whatever arrived first. Swallowing that exception on
the portfolio edges gave the invariant away: a 429 on owned_pages left
the merged list holding only the /me/accounts page, and the callback
connected it with no picker.
The exception now carries whether the failure was transient, classified
by GraphError, which already owns Meta's rate-limit and transient code
table. A denied permission reads as "this login reaches no portfolio
pages"; a throttle, a 5xx or a truncated walk is raised.
The walk also stops at MAX_PORTFOLIOS and logs what it skipped. Each
portfolio costs two more paginated edges inside a synchronous OAuth
callback, and nothing bounded that loop.
* fix: three ways the portfolio walk misread what Meta returned
Concurrent Instagram lookups. The Instagram description ran one request
per Page, in sequence, at a 15s timeout each. That list used to be the
Pages someone holds a role on — a handful. It is now the union with
every portfolio's owned_pages and client_pages, so a portfolio holding
hundreds of Pages serialised the OAuth callback past any gateway
timeout, for exactly the accounts the portfolio walk exists to reach.
Meta's ids= batching is no help: each Page carries its own access token
and one call takes one token. The lookups run in concurrent rounds.
A Page without a token is not a Page you don't have. Meta lets someone
decline pages_read_engagement on its per-permission toggles and still
lists the Page, without an access_token. Dropping it inside the walk
left the caller saying "no Pages found, you need to be an admin of at
least one" to an admin. ManagedPages returns everything Meta listed and
publishable() separates what can be posted to, so the callbacks can
tell the two apart and say which one happened.
Stored scopes are what Meta granted. The scope list was written to the
account's scopes column straight from the request, claiming access the
login may have refused — business_management above all, which needs
Advanced Access and is declined by default without it. It now comes
from /me/permissions, falling back to the request when Meta cannot be
asked.
* fix: only drop a scope Meta says was refused
PublishToSocialPlatform::failForMissingScopes() blocks a post when a
platform's required publish scope is absent from the account's scopes
column, so writing that column from /me/permissions can dead-end an
account. Meta does not document that the endpoint echoes scope strings
verbatim, and the edge is paginated, so a scope it never mentions is
unknown rather than refused and stays. Only declined and expired drop.
* fix: keep the portfolio walk honest and cheap
Raise instead of truncating. The ceiling logged a warning and returned
whatever fit, which is the one thing this module refuses to do
everywhere else: if the walk cannot finish, the real list is unknown,
and a truncated list holding exactly one Page would have been
auto-connected without ever showing the picker. It now raises, and the
ceiling rises to GraphPaginator::MAX_PAGES' 100 since the walk no
longer pays for it serially.
Read the edges concurrently. Up to two paginated edges per portfolio ran
back to back inside the OAuth callback. They run in rounds now; a URL
that does not come back cleanly still goes through GraphPaginator, which
owns the single place that logs a Graph failure and decides whether it
is a rejection or an unknown.
Prefer the record that carries a token. Merging kept whichever copy of a
Page id arrived first, and /me/accounts always arrives first — so a Page
listed there without a token buried the portfolio copy that had one, and
the login was told its permission was missing for a Page it could reach.
Describe only the Instagram accounts that survive. The per-Page lookup
ran before filterConnectableIdentities discarded them, spending a
BUC-rate-limited call on every Page only to throw the answer away. The
filter reads instagram_business_account.id straight off the raw Page, so
it needs no lookup to run first.
Two Instagram tests mocked Socialite without usingGraphVersion, so the
callback threw, the generic catch answered, and asserting only
success=false passed on the error path instead of the one under test.
* test: pin that pages survive past the first pooled round
The concurrency test drove 30 portfolios with every edge empty, so the
merge across rounds was never exercised with data in it.
* fix: keep the paging-host guard on the pooled edge walk
GraphPaginator refuses to follow a paging.next that points off the host
the walk started from, so a tampered response cannot carry the access
token somewhere else. Reading the first page of each edge through the
pool and handing its paging.next straight back to GraphPaginator made
that URL the *start* of a new walk, which is the one URL the guard
trusts implicitly — so the first hop went unchecked.
The host is compared before the hand-off now, and a mismatch re-walks
the edge from the beginning so GraphPaginator's own guard is what
refuses it, with its logging.
* fix: stop a cut-short walk from passing for a complete one
optional() was written for "this edge may be forbidden" and answers a
rejection with an empty list. Following paging.next through it gave a
rejected cursor the same answer: page one of a 250-Page portfolio came
back and the rest was dropped, and a single connectable Page in that
fragment would have been auto-connected with no picker. The same hole
sat on /me/businesses, where GraphPaginator is all-or-nothing — a
failure on page two threw away the portfolios page one had already
listed, degrading the connect back to /me/accounts alone in silence.
The exception now carries how many pages arrived. Only a rejection on
the very first request reads as "this edge is not readable"; anything
after that is a fragment and raises. Cursors skip optional() entirely.
A login Meta reports as refusing business_management also stops walking
the edges at all. The controllers already read /me/permissions for the
scopes column, so the answer costs nothing, and the walk was otherwise
spending a request on a certain 403 — and logging it at error level —
on every successful connect by such a login.
An Instagram account with an empty Name connected as display_name null:
data_get's default only fires on an absent key, and describeRound always
writes the key.
* test: pin reconnecting a card only the portfolio still reaches
A card whose Page moved behind a portfolio is the reconnect shape of the
bug this branch fixes, and nothing covered it: the walk has to find the
Page, and filterConnectableIdentities has to keep the original card
rather than offering the portfolio's other Pages.
* fix: an unreadable portfolio must not deny the pages that were readable
The portfolio edges are additive, but every failure in them was raised
and the callback's generic catch turned it into "error connecting" — so
one throttled edge among sixty denied a login the Pages /me/accounts had
already returned, and each retry burned more of the quota that caused it.
Only /me/accounts failing is fatal now; everything else marks the walk
incomplete and keeps what arrived.
What the raise was protecting is kept where it belongs: a lone Page is
only taken without asking when the walk saw everything, or when a
reconnect has already pinned which Page is wanted. Otherwise the picker
opens, and the login can see for itself that its Page is not there.
The ceiling stops pretending. It compared the count after walking every
page of /me/businesses — up to ten thousand ids — so the runaway it
existed to bound had already happened. One request, one page, and more
portfolios than that is an incomplete walk rather than a failed one.
A pooled edge that fails is classified where it lands instead of being
re-fetched, halving the cost of the common client_pages rejection, and
GraphPaginator logs a confirmed rejection at warning: it is Meta
answering the question, not something going wrong.
A login that declined the permission its platform needs to publish is
refused at connect. Meta issues a Page token off pages_show_list, so
declining pages_manage_posts still produced a green account whose every
scheduled post was then hard-failed by failForMissingScopes.
Test fakes address Graph through the config rather than a literal host,
which is what CLAUDE.md asks for and what the newer tests already did.
* fix: an incomplete walk must not answer as if it were sure
Marking the walk incomplete stopped the auto-connect, but every dead end
after it still gave a definitive answer. A login whose only Pages sit
behind a throttled portfolio was told "no Facebook Pages found, you need
to be an admin of at least one" — the exact sentence this branch exists
to stop showing to admins, now arriving for a different reason. The
already-connected and missing-permission answers were equally sure of
themselves.
When the walk could not see everything and there is nothing to offer,
it says so and asks for a retry.
* fix: stop every Inertia test from calling an SSR server
inertia.ssr.enabled defaulted to true and nothing in phpunit.xml turned
it off, so every test rendering an Inertia page issued a real request to
the SSR endpoint. The project does not use SSR, so those calls only ever
failed and fell back to client rendering — quietly, on every run.
Defaulting it off is what the project already assumed, and it retires
the allowStrayRequests hole the Meta connect tests were carrying to work
around it. INERTIA_SSR_ENABLED still turns it back on.
* fix: a taken slot is a fact, not a guess about the listing
Routing every short listing to "try again in a moment" swallowed
network_taken: a workspace that already holds its one Facebook account
was told to retry, forever, whenever a portfolio edge was throttled. That
answer comes from our own rows and does not depend on how far the walk
got. all_connected and page_not_found do, and still yield.
Also: an off-host cursor now stops the edge instead of re-reading page
one, which cost a request and could follow an on-host cursor on the
retry, quietly undoing the guard. Cursor follow-ups are budgeted, since
they cannot be pooled and were the one unbounded serial path left. The
exception's fetched count lost its last reader two commits ago and is
gone. Comments trimmed throughout.
* fix: bound the cursor walk by requests, and keep what it read
MAX_CONTINUATIONS counted edges, not requests: each one then handed off
to GraphPaginator, which follows up to a hundred more pages by itself.
The budget the docblock promised was fifty times larger than it claimed.
Cursors are now followed one budgeted request at a time, so the count
means what it says, and pages already read survive a cut-off instead of
being thrown away with the exception.
A refused /me/businesses is no longer read as "this login has no
portfolios". For a single edge a rejection answers the question; for the
index of edges it means we could not look — and answering complete there
auto-connected the one /me/accounts page while hiding every portfolio
Page, which is this branch's own bug wearing a different hat.
SSR goes back to its shipped default. Turning it off in config to quiet
the test suite would have disabled it wherever it is actually started —
docker/Dockerfile builds the bundle. phpunit.xml carries the switch now,
next to PULSE, TELESCOPE and NIGHTWATCH, and CLAUDE.md records why.
* refactor: one Meta connect flow instead of two kept in step by hand
The Facebook and Instagram-via-Facebook callbacks ran the same twenty-five
lines: the profile touch Meta's review wants, the granted-scope read and
the publish-scope refusal, the page walk, and the answer for a walk with
nothing to offer. They only matched because both were edited side by side,
every round, which is a guarantee nobody should be making by hand.
graphApi() moves to SocialController and reads the host by platform value,
so it serves every network rather than the two that had copied it, and
graphVersion() derives from it instead of reading config a second time.
select() stays as it is. The two differ in the middle — different identity
keys, different connect shapes — and folding them would be abstraction for
its own sake.
* fix: default Inertia SSR off, where this project already stands
Nothing in the repo starts an SSR process, so the shipped default was
describing a setup that does not exist. With it off the test env needs no
override of its own, and CLAUDE.md records that turning it on means
starting the process, not just flipping the env.
* fix: one rule for a refused portfolio, and a clock on the walk
Last round I made a refused /me/businesses mark the walk incomplete, on
the argument that refusing the index means "we could not look". That was
wrong in the case that matters most: an app without Advanced Access for
business_management gets that refusal on every single connect, so every
login on such an install lost auto-connect and every login without Pages
was told to retry forever. Self-hosted in Live mode is exactly that.
The rule that holds everywhere: a Page this login cannot enumerate is a
Page it cannot get a token for, so it was never connectable, and the list
of connectable Pages is complete. Only an unknown — a throttle, a hiccup,
a budget or a ceiling — leaves the walk unable to vouch for itself. Index
and edge now answer the same way, which is also what makes the two
readable together.
The per-request budgets were each bounded while their sum was not: ten
pooled rounds plus twenty-five cursor requests can outlive nginx's
fastcgi_read_timeout of 120s. The walk now carries a deadline and returns
what it has.
touchProfile exists only because Meta's review wants the call. It had no
timeout and no guard, so a hung /me could stall the callback to the
gateway timeout or fail a connect outright, over a response nobody reads.
* docs: the walk's contract changed under its own docblock
It still said any failure marks the walk incomplete, which stopped being
true when a refusal became an answer. A docblock describing an invariant
the code no longer holds is how this branch got two of its bugs.
* fix: put the whole callback inside the budget it advertises
meta_page_walk_seconds bounded the portfolio half of the walk and nothing
else. /me/accounts could paginate a hundred pages at fifteen seconds
each, and the Instagram lookups pooled in rounds that were themselves
serial — a portfolio with three hundred linked Pages is fifteen rounds,
after the walk had already spent its own budget. Both honour the deadline
now. The lookups skip rather than drop: the Page still connects, only its
handle and avatar arrive empty. The first request is always made; the
budget bounds what comes after it.
A Graph body that is valid JSON but not an object — a proxy answering
"throttled" — reached GraphError::isTransient, whose parameter is ?array,
and under strict_types raised a TypeError. That is an Error, so it walked
past both callbacks' catch(\Exception) and 500'd the popup instead of
showing a message.
Refusing a login before any listing has happened no longer borrows the
wording for "we found your Pages but not the permission to post to them".
composer run dev no longer starts an SSR process for SSR that is off, and
CLAUDE.md no longer claims nothing in the repo starts one, which
composer.json contradicted.
* fix: say what actually gets a Page token, per Meta's own reference
The Page node reference is explicit: access_token is "only returned if
the User making the request has a role (other than Live Contributor) on
the Page". Being an admin of the portfolio that owns a Page lists it but
does not grant that role, so the walk can surface Pages this login will
never get a token for.
The popup told those users to reconnect and accept every permission,
which cannot produce a Page role and so could never work. It now names
the role as well.
I rejected this in review on the grounds that a portfolio Page had been
published to successfully in the wild. That proved a token comes back
when the login holds a role, not that one always does.
* fix: one budget for the callback, not one per phase of it
The walk and the Instagram lookups each opened a full
meta_page_walk_seconds, on top of the profile touch and the permission
read, so the callback's worst case was several times the single bounded
budget config/trypost.php advertises. They share one deadline now, taken
once and passed down. META_PAGE_WALK_SECONDS joins .env.example.
An Instagram account described past that deadline arrived with no handle
and no name, and a lone one was then persisted with display_name null —
a blank, unidentifiable card. It falls back to the Page's own name.
Two docblocks were describing behaviour the code does not have.
/me/accounts is the base every other Page is added to, so running out of
budget there aborts rather than degrades, and the class now says so
instead of promising a partial list. GrantedPermissions justified
treating an absent permission as unknown but said nothing about a failed
request, which lands in the same place for a different reason.
* fix: a Pages throttle on a user token was reading as a refusal
Meta's BUC rate-limit table lists code 32 for the Pages API when called
with a User token. GraphError did not carry it, because until this branch
nothing in the app called a Pages surface that way — the publishers use
Page tokens, where the same throttle arrives as 80001.
The portfolio walk does: /me/accounts, /me/businesses and both edges are
read with the user token straight out of OAuth. So an ordinary throttle
came back as code 32, was classified as a confirmed rejection, and the
walk concluded this login simply reaches no portfolio Pages — vouching
for a list missing all of them and auto-connecting whatever /me/accounts
happened to hold. A rate limit was producing the exact silence the
complete flag exists to prevent.
* fix: a reconnect no longer loses its handle to a slow Graph
persistIdentity updates a reconnected card with whatever it is handed,
so a described-with-nulls card overwrote a working account's username
and avatar. Skipping the Instagram lookup — which the shared budget now
does whenever the walk spent it — produced exactly that card. A lookup
that never ran says nothing about a handle the account already has, so
those two keys are left out when it did not.
Refusing the portfolio index goes back to marking the walk incomplete. I
had it that way, reversed it, and this settles it: Meta's Page reference
returns access_token for a Page the login holds a role on, and a Page can
carry that token on a portfolio edge while /me/accounts omits it — which
is this branch's entire premise. So refusing one edge does say those
Pages are unreachable, but refusing the index says no edge was read at
all, and the Pages behind it may well have been connectable. Silently
vouching for a list without them is the original bug.
The budget also starts before the walk and now shapes each request's own
timeout, so no single call can outlive it by fifteen seconds.
composer dev:ssr was a slower alias of composer dev once the SSR process
came out of it.
---------
Co-authored-by: StoriaJames <james@storia.tech>
* feat: expose self-hosted mode to the accounts UI
SocialAccountObserver already bypasses the one-account-per-network
guard when trypost.self_hosted is true, but the frontend had no way
to know that and always collapsed a network to a single card once
any account existed - so self-hosted deployments could not surface
a second LinkedIn (or Instagram) connection even though the backend
would allow creating it.
* feat: allow connecting multiple accounts per network when self-hosted
NetworkConnectGrid always collapsed a network (LinkedIn profile/page,
Instagram standalone/Facebook) to a single card once any account
existed, with no way to trigger another OAuth flow - even though
SocialAccountObserver already allows unlimited accounts per network
in self-hosted mode. A self-hoster connecting their personal LinkedIn
profile had no path back to the connect flow to also add a company
page (or a second company page/showcase page).
Render one card per connected account instead of collapsing to the
first, and keep a standing "Connect another" card available for a
network's existing connections when self-hosted. Hosted mode is
unchanged: still one card per network, matching the backend's
still-enforced one-account-per-network limit there.
* test: cover the selfHosted prop on accounts and onboarding pages
Backend behavior for connecting a second identity per network in
self-hosted mode was already covered (LinkedInControllerTest,
NetworkUniquenessTest) - these just confirm the new prop the frontend
now depends on is actually present and reflects config correctly.
* style: apply prettier formatting
Pre-existing drift in this file unrelated to the selfHosted change.
* refactor: read selfHosted from shared Inertia props
The flag is already shared by HandleInertiaRequests, so the accounts
and onboarding controllers do not need to pass it again.
Co-authored-by: Cursor <cursoragent@cursor.com>
* feat: gate multiple social accounts with a dedicated env
Cloud cannot flip SELF_HOSTED, so one-per-network is now ALLOW_MULTIPLE_SOCIAL_ACCOUNTS (default false).
Co-authored-by: Cursor <cursoragent@cursor.com>
* fix: tighten multiple-account gates after review
Keep every connected identity visible, share occupiesNetwork, and return network_taken instead of a generic connect error.
Co-authored-by: Cursor <cursoragent@cursor.com>
* fix: bind reconnect to the card and unique social identity
Reconnect now updates the selected account, and a unique index plus connectIdentity keep the same platform identity from being inserted twice.
Co-authored-by: Cursor <cursoragent@cursor.com>
* refactor: build the OAuth URL before opening the popup
Keep the popup opener URL-only so reconnect query params are assembled at the call site.
Co-authored-by: Cursor <cursoragent@cursor.com>
* refactor: drop dead social-account guards and slim the connect grid
Skip migration cleanup that production never needs, trust the platform enum in the observer, and move card theming out of the grid.
Co-authored-by: Cursor <cursoragent@cursor.com>
* refactor: scope social reconnect to the current network
Drop dead instanceof/isset guards and filter reconnect targets in the query so a stale session cannot update another network's card.
Co-authored-by: Cursor <cursoragent@cursor.com>
* refactor: slim connectable-identity filtering
Index OAuth identities by id so reconnect and occupancy use only/except instead of hand-rolled filters.
Co-authored-by: Cursor <cursoragent@cursor.com>
* refactor: slim social identity persist helpers
Drop the unused occupiesNetwork exception and persist reconnects with update() instead of fill/save/fresh.
Co-authored-by: Cursor <cursoragent@cursor.com>
* fix: keep reconnect updates on the original social card
Co-authored-by: Cursor <cursoragent@cursor.com>
* test: run the suite with multiple social accounts enabled
phpunit.xml forced ALLOW_MULTIPLE_SOCIAL_ACCOUNTS=false, overriding the
true value in .env.ci. That broke eight tests across Automation, MCP,
PostApi, RefreshExpiringTokens and VerifyUpcomingConnections which only
needed two accounts of one network as a fixture, not as a rule under test.
Match .env.ci instead. Every test that exercises the one-per-network rule
already sets the config itself; the accounts index test was the only one
leaning on the implicit default, so it now pins it.
* fix: align the multi-account fallback with the self-hosted default
allow_multiple_social_accounts fell back to env('SELF_HOSTED', false)
while self_hosted itself defaults to env('SELF_HOSTED', true). A
self-hosted install that never wrote SELF_HOSTED to its .env resolved to
false and silently lost multiple accounts per network on upgrade, which
is the opposite of what the documented fallback promises.
* fix: collapse duplicate identities before adding the unique index
Installs predating the index can hold the same identity twice: the
network guard was bypassed for multi-account installs and Pinterest
always created a fresh row. Creating the index on that data aborts
migrate mid-deploy.
Keep the newest row per identity and move its post_platforms over before
dropping the duplicates - the FK is nullOnDelete, so deleting outright
would orphan drafts and scheduled posts.
* fix: refuse a reconnect that authorized a different identity
connectIdentity overwrote platform_user_id with whatever the provider
returned, so reconnecting a card while signed into another account
repointed the row - and every draft and scheduled post bound to it - at
a stranger. LinkedIn guarded this at the controller and Facebook via its
filtered page list; nothing covered X, TikTok, Threads, Discord,
Bluesky, Mastodon, Pinterest, Instagram or Telegram.
Enforce the identity match at the single choke point every connect flow
goes through. Every call site already maps NetworkAlreadyConnectedException
to network_taken, so the refusal surfaces without new plumbing.
Also restore the null-platform guard in the observer: occupiesNetwork
type-hints a non-nullable Platform, so a row without one died with a
TypeError instead of the database's NOT NULL error.
* fix: filter connectable identities on every picker step
YouTubeController::select re-fetched the channels and matched the posted
id straight off the raw list, unlike callback and selectChannel. With a
live youtube_oauth session it let a POST name any channel the Google
account owns and bind it to the reconnect target. It also read the
reconnect from the session while the connect below it read
youtube_oauth.reconnect_id, so the two could disagree - pass the
resolved account through instead.
filterConnectableIdentities also short-circuited in multi-account mode,
and the unique index is scoped to platform rather than network. That let
one Instagram account connect twice, once directly and once via
Facebook, publishing every Instagram post to it twice. The existing
except() already spans networkPlatformValues(), so dropping the
short-circuit closes it.
* refactor: type the connect cards and drop the dead accounts grid
The cards computed inferred account as a required ConnectedAccount and
then pushed undefined onto it (TS2345). CI only runs eslint so it stayed
green, but vue-tsc and editors flag it.
SocialAccountsGrid is referenced nowhere; its reconnect button was
updated in this branch without passing the card id, which would have
been a bug had anything rendered it.
* test: keep the suite on the cloud one-account-per-network default
CI runs the Cloud build, so the suite baseline should be the Cloud
default rather than the self-hosted one. Put phpunit.xml and .env.ci
back to false and make the eight tests that merely need two accounts of
one network as a fixture opt in for themselves.
This also un-deads the config()->set(true) calls the branch had already
added to AuthenticationTest, SyncAccountUsageTest, HasUsageTraitTest and
SocialAccountObserverTest, which the forced true had turned into no-ops.
* fix: connect standalone instagram instead of reopening the picker
The picker emits an already-resolved connect method, but this branch
rewired @select from openOAuthPopup to startConnect. startConnect sends
a bare 'instagram' straight back into its own picker branch, so choosing
"Instagram" closed the dialog and immediately reopened it - the OAuth
window never opened and the standalone flow was unreachable. Only the
via-Facebook button still worked.
Split the URL-opening tail out of startConnect and let the dialog call
that directly.
* fix: reject a telegram reconnect before burning the connect code
The nonce was consumed before connectIdentity ran, so posting /connect in
the wrong chat spent the one-off code and forced the user to generate a
new one. Check the identity first and report wrong_chat instead of
network_taken, which told them to disconnect an account when the real fix
was posting in the channel they were reconnecting.
* fix: leave one target per post when merging duplicate accounts
post_platforms has no unique on (post_id, social_account_id), so a post
holding a row per duplicate account ended up with two enabled rows aimed
at the surviving account and would publish to it twice. Keep one row per
post, preferring a published one so history survives.
* refactor: collapse the repeated connect-flow boilerplate
Four shapes were copy-pasted across the connect controllers:
- the session + permission guard opening 16 actions, now connectWorkspace()
throwing a ConnectPopupException that renders the popup itself
- the reconnected/connected ternary in 13 places, now connectedCallback()
- the "nothing left to connect" branch in 4 places, now
noConnectableIdentities()
- validatedReconnectId() re-querying what reconnectAccount() already does
Facebook, Instagram-via-Facebook and YouTube also re-queried the reconnect
account three or four times per callback; it is resolved once and passed
down. The three GET pickers skipped the manageAccounts check their POST
siblings had, and pick it up from the shared guard.
Drops the color key from connectableOptions and the matching frontend
field - nothing read it. Platform::color() stays; the disconnection
emails use it.
* fix: keep an expired connect popup out of the error log
ConnectPopupException escapes to the framework handler so it can render
itself, which also meant report() ran first: every session_expired and
workspace_not_found popup filed an ERROR and a Nightwatch issue for what
used to be a silent return. A stale popup is a normal outcome, so it now
implements ShouldntReport.
The Mastodon and Threads guards also cleared their provider session
after connectWorkspace(), so a workspace that vanished mid-flow left the
client secret and the OAuth state behind. Clear first, then resolve.
clearMastodonSession() no longer touches social_connect_workspace -
whatever closes the popup already does.
* fix: stop telling users to disconnect an account that is not the problem
Two flows reused popup_callback.network_taken - "This workspace already
has an account for this network. Disconnect it first." - for situations
where that is neither true nor actionable:
- reconnecting a card while signed into a different account on the
provider, now wrong_account
- an empty picker in multi-account mode, where every page or channel on
that login is simply already connected, now all_connected
NetworkAlreadyConnectedException carries the message key so the catch
sites stay one line. handleCallback() also drops its $platform argument;
it read $this->platform for the reconnect lookup and the identity filter
either way, so a caller passing a different platform would have scoped
the lookup to the wrong network.
* refactor: filter linkedin identities with the shared helper
The picker hand-rolled its own reconnect narrowing because the profile
and the pages arrive in two different shapes. Flatten them into one pool
of LinkedIn identities, run the shared filter, and split them again for
the view - the same path Facebook, YouTube and Instagram already take.
Side effect worth having: the picker previously only narrowed on a
reconnect, so it would offer an identity that is already connected and
only fail once the user picked it. It now hides taken identities up
front and says so when nothing is left.
* fix: keep the linkedin picker's own empty state
Routing the picker through the shared filter made every empty pool look
like "nothing left to take", including the pool LinkedIn never filled.
A self-hoster running pages-only who administers no page was told the
network was already connected, or that every account on the login was
taken - both false - and the picker's own "you are not an admin of any
LinkedIn page" state became unreachable.
Only treat it as taken when filtering is what emptied it. Splitting the
pool back also compared the person id loosely on one side and strictly
on the other; one predicate now drives both.
Threads had two forget() calls for a key the top of the action already
clears, and YouTube's picker resolved the reconnect account twice on the
failure path.
* fix: keep the enabled row when collapsing duplicate post targets
SyncPostPlatforms seeds a disabled post_platforms row for every account
in the workspace, so the usual duplicate is one row the user actually
checked next to one they never saw - both pending, both created in the
same second. Ordering only by published-then-newest made that a coin
flip, and PublishPost iterates enabled() only, so half the time a
scheduled post would silently stop reaching that account and take its
caption and per-platform meta with it. This runs once against production
data and the dropped row is gone, so enabled now beats disabled.
Also: the empty-pool exit from the LinkedIn picker was the only one
leaving linkedin_pending - and its tokens - in the session. The
rationale comments move to the docblocks they belong in, and usePage()
comes out of the cards computed.
* fix: stop the migration destroying publish history and automations
Two ways the one-shot merge lost data that cannot be rebuilt:
Surplus published post_platforms rows were deleted. Two duplicate
accounts really could each have published, and each row carries the
platform_post_id for a live post on the network - dropping one leaves
that post unmanageable and invisible to metrics. The docblock claimed
published beat everything; now the code does, and only unpublished
repeats collapse.
Automation nodes persist social_account_id inside a JSON column with no
foreign key, so deleting the loser left RunGenerateNode skipping that
target, or generating nothing at all when it was the node's only
account. The ids are rewritten - current and legacy shapes both - and
entries the merge just turned into duplicates are collapsed.
Ordering is now total (null created_at sorts oldest on every engine,
then id) so a rehearsal on a replica keeps the same rows as the real
run. The LinkedIn picker also passes onboardingProgress inline: it
clears linkedin_pending on the empty path, and a deferred reload would
re-GET the route and swap the empty state for a session-expired popup.
* fix: make the identity merge auditable and stop a second delivery
Self-hosted installs run this unattended and it cannot be undone, so
each collapsed group now logs the workspace, the identity, which row was
kept, which were dropped, and how many post_platforms and automations it
touched. down() says plainly that it drops the index only.
Two narrower fixes:
A post holding a published row plus an enabled unpublished row for the
same account kept both, and PostPlatform::scopeEnabled() filters on
`enabled` alone with no status check - so a republish would deliver the
same content to that identity twice. Once a published row exists, every
unpublished repeat goes.
The automation dedupe ran on every automation in the workspace, not just
the ones the merge rewrote. A node legitimately holding two entries for
one account under different content types would be collapsed to
whichever came first in the array. It now runs only where an id was
actually substituted.
* test: rehearse the identity merge against a messy database
Every test on this migration so far covered a case someone thought to
write, which is why three separate review rounds each found a defect the
earlier ones missed. This builds a deliberately messy database instead -
three workspaces, four networks, one to three copies of each identity,
posts mixing published, pending and failed rows across the duplicates
with enabled flags varying, and automations referencing them in both the
current and legacy JSON shapes - then runs the real migration and
asserts what must be true afterwards rather than what happens to a
particular fixture.
Invariants: no duplicate identity survives, no published row is ever
destroyed, no post ends up enabled twice against one account, nothing in
post_platforms or automations points at a deleted account, and the
newest row of each identity is the one kept.
The generator is seeded, so a failure reproduces, and it asserts its own
output is adversarial - roughly nine duplicate groups and fourteen
published rows - so it cannot quietly degrade into passing on an empty
problem. Verified by mutation: dropping the automation repoint, the
published guard, or the repeated-target collapse each fails exactly the
invariant that covers it.
* fix: stop the youtube picker refetching itself into a cleared session
HandleInertiaRequests defers onboardingProgress for anyone mid-onboarding
- exactly the people connecting their first accounts - so Inertia
re-GETs the picker route right after it mounts. For Facebook and
Instagram that re-entry is harmless and deliberately left deferred, but
YouTube calls the Google API again, and fetchChannels() turns any
failure into an empty list that clears the connect session and swaps the
mounted picker for an error the user cannot retry from. Same guard the
LinkedIn picker already got.
LinkedIn also answered a reconnect that authorized a different identity
with "Page not found", including in the person branch where no page is
involved. Every other platform says wrong_account, which this PR added.
* refactor: drop the unreachable youtube channel picker
Google's own delegation screen already lists every channel on the
account and makes the user pick one before it issues the token, so
channels?mine=true always answers with that single channel and
count($channels) === 1 always won. The picker behind it was never
reached - its Vue page was deleted back in 7c00c338 (January) and
nothing broke, which is the clearest evidence it was dead.
Removes selectChannel(), select(), both routes, the youtube_oauth
session payload and the tests that drove them. If Google ever does
return more than one, the callback connects the first and logs a warning
rather than routing to a screen that no longer exists.
* fix: serialize connects so two popups cannot seat one network twice
The observer's occupiesNetwork() is a check-then-insert with nothing
holding the gap, and the new unique index covers the identity, not the
network. Two tabs finishing OAuth at the same moment for *different*
identities on one network both passed the exists() check and both
inserted, leaving a Cloud workspace with the two accounts the rule
exists to prevent. The same-identity race was already safe - the unique
violation is caught and re-queried.
A database constraint cannot hold this: allow_multiple_social_accounts
is a runtime flag, so the rule is on for Cloud and off for self-hosted,
and an index cannot read config. Lock per workspace and network instead,
the way markAsDisconnected() and ConnectionVerifier already do.
This covers connectIdentity(), which every OAuth flow and the Telegram
action go through. A direct create() still answers to the observer
alone, and a self-hosted install running file cache across several nodes
locks per node.
* fix: handle a busy connect lock on the telegram path
Every other caller funnels LockTimeoutException into its generic
\Exception catch and closes the popup with error_connecting. Telegram
has no such catch, so the new lock could 500 the webhook - and because
the nonce is spent before connectIdentity runs, Telegram's retry of the
same update short-circuits on the consumed code and returns without
dispatching anything. The dialog would spin forever on a code that can
no longer be used.
Also restores coverage the picker removal dropped: the deleted select
tests were the only ones driving a multi-channel response, so nothing
exercised the reconnect narrowing to its own card, or multi-account mode
skipping an already-connected channel. Both are back against the
callback, and removing the narrowing in filterConnectableIdentities
fails them.
* fix: stop the instagram login seating an account already held via facebook
filterConnectableIdentities() drops every identity already connected on the
network, which is what keeps one Instagram account from being seated twice
under its two platforms. Every flow that persists an identity ran it except
the direct Instagram Login callback, so the guard only held in one direction:
InstagramFacebookController refused an account already connected as
`instagram`, but the reverse was allowed through.
With multiple accounts per network enabled the observer's network check is
bypassed and the unique index does not span platforms, so authorizing the
same account through the direct flow created a second row. Both then seed a
post_platform row and the post goes out twice to one account.
* fix: name the real reason when a linkedin profile reconnect switches member
Reconnecting a card narrows the authorized identities to that card's own, so
authorizing a different LinkedIn login empties the pool. selectIdentity()
reported that as "Page not found." for every card, including personal
profiles where no page was ever involved.
A profile reconnect has no page to be missing: an empty pool there can only
mean this login is a different member. Say so with the wrong_account wording
select() already uses for the same condition. Page reconnects keep
page_not_found, where the organization really can be absent from the login.
* fix: surface the busy telegram connect instead of a generic failure
The connect lock timing out dispatches its own 'busy' reason so the dialog
can tell the user to retry, but the dialog only mapped network_taken and
wrong_chat and fell back to error_generic for everything else. The reason
reached the browser and died there, leaving "Could not start the connection"
for a case that just needs another moment.
* test: cover reconnect on every flow that gained it
rememberConnectSession() gave Instagram, TikTok, Threads, Mastodon and
Bluesky a reconnect path they did not have before — TikTok had been actively
clearing social_reconnect_id on connect — and none of them had a test for it.
Facebook, LinkedIn, YouTube, X, Discord, Pinterest and Telegram already did.
Each now covers both halves: authorizing the same identity refreshes the
existing card and reports it as a reconnect, and authorizing a different one
is refused with wrong_account instead of quietly seating a stranger on the
card and every post scheduled against it.
* fix: repair what a reconnect leaves behind when it cannot proceed cleanly
Two things connectIdentity got wrong once the reconnect path existed.
A reconnect through the other variant of a network moves the card to the new
platform — same identity, different API flavor. Post targets carry their own
platform snapshot, and that snapshot picks the publisher, the queue and the
scopes checked before publishing. Left behind, it failed every pending post on
a permission the account no longer needs: an Instagram card moved to the
Facebook variant still demanded instagram_business_content_publish and stopped
with "Missing permissions". Pending targets now follow the card and reset a
content type the new platform cannot publish; published targets keep theirs,
since they record what really went out under a platform_post_id from that API.
The network lock timing out also arrived as a raw LockTimeoutException, which
every OAuth callback filed through its generic catch: an error log and "Error
connecting account" for the exact race the lock exists to absorb. It now
carries a busy messageKey through the branch each flow already handles, the
same way the Telegram path already reported it.
* refactor: resolve the linkedin reconnect card once per select
select() already looked the card up before deciding whether the chosen
identity matches it, then connectPerson() and connectOrganization() looked it
up again on their own — two identical queries per submit, and two places that
could disagree about what is being reconnected. The caller passes what it
already holds.
* test: render the grid's multi-account branch
phpunit.xml forces ALLOW_MULTIPLE_SOCIAL_ACCOUNTS false and no browser test
overrode it, so the card the flag exists to add never rendered anywhere. The
pair pins both sides: a taken network offers no second card when multiples are
off, and offers one when they are on.
* test: pin why the linkedin select guards exist
connectIdentity() already refuses a mismatched reconnect and answers with the
same wrong_account message, so every existing test passes with the two guards
in select() deleted — which is exactly how they would get deleted. What they
actually buy is skipping the avatar download that building the connect payload
runs first.
Both now assert the fetch never happens, so the guards fail loudly instead of
looking redundant.
* fix: carry retrying targets through a variant move, atomically
Two holes in the move added a commit ago.
It only carried pending targets, but a retrying one is not finished either —
the publish job reschedules itself and reads the snapshot fresh on the next
attempt, so leaving it behind meant it retried against the old variant until
it exhausted its budget on a permission the account no longer needs. Failed
and published targets stay put; a publishing one has a job mid-flight already
working from the snapshot it read.
The card and its targets also moved in three separate statements, so a crash
between them left exactly the split this was meant to close. They share a
transaction now.
* chore: drop the dusk selectors nothing reads
Laravel Dusk is not installed — no laravel/dusk requirement, no DuskTestCase,
no browse(). Browser tests run on pest-plugin-browser driving Playwright, and
its @selector resolves to data-testid. The 45 dusk attributes left across 18
components selected nothing.
CLAUDE.md was the reason they kept coming back: it told every agent to add
them. Its browser-testing section now describes the setup that exists —
data-testid targeting, the wait helper these tests need because assertions do
not auto-wait on SPA paint, and why BrowserTestCase keeps Vite real.
Verified before removing: every @selector used in tests/Browser resolves to a
data-testid, seven of them through bound :data-testid, so none depended on a
dusk attribute.
* chore: drop the last one-account-per-network helper
hasConnectedPlatform() has no callers left anywhere — app, tests, views or
routes. It sat directly above getSocialAccount(), which this branch already
removed, and is the same leftover from when a workspace could hold one account
per platform.
---------
Co-authored-by: Paulo Castellano <paulo@castellanos.llc>
Co-authored-by: Cursor <cursoragent@cursor.com>
* Improve asynchronous social publishing reliability
* fix: resume asynchronous social publishes
* fix: preserve publish checkpoints across retries
* fix: harden resumable publish lifecycle
* fix: clean retry resources on terminal failures
* test: cover resumable social publishing edge cases
* feat: add failed post retry command
* chore: remove retry command ai rule
* fix: require confirmation for post retries
* chore: remove ai rules index
* chore: remove ai social rule
* refactor: clarify TikTok derivative path validation
* refactor: simplify social publishing retries
* refactor: further simplify social publishing retries
* refactor: retry all failed post platforms
* style: import throwable in social retries
* refactor: decouple TikTok cleanup from image format
* refactor: extract missing publish scopes
* refactor: encapsulate missing scope failure
* fix: resume failed publishes and treat Instagram rate limits as transient
Keep TikTok/Instagram checkpoints on posts:retry so a manual retry does not
start a duplicate remote post. Classify Meta BUC 400s on Instagram status
polls as retryable via GraphError.
Co-authored-by: Cursor <cursoragent@cursor.com>
* test: cover resume paths and transient Instagram rate limits
Co-authored-by: Cursor <cursoragent@cursor.com>
* fix: resume posts:retry only for in-flight publish failures
Co-authored-by: Cursor <cursoragent@cursor.com>
* fix: resume posts:retry via ErrorCategory instead of string lists
Co-authored-by: Cursor <cursoragent@cursor.com>
* fix: handle Instagram PUBLISHED and EXPIRED container statuses
Treat EXPIRED as a terminal server error so posts:retry starts over, and complete already-published containers without a second media_publish.
Co-authored-by: Cursor <cursoragent@cursor.com>
* fix: recover published Instagram stories from /stories
Stories are not on GET /{ig-user-id}/media. Resume a PUBLISHED story container from the stories edge so we do not bind a feed post id.
Co-authored-by: Cursor <cursoragent@cursor.com>
* test: cover Instagram EXPIRED retry and published recovery paths
Co-authored-by: Cursor <cursoragent@cursor.com>
* fix: stop guessing Instagram media ids from recent /media
Co-authored-by: Cursor <cursoragent@cursor.com>
* fix: checkpoint TikTok publish_id and keep in-flight photo derivatives
Persist publish_id right after /init/ so a crash can resume without a second publish. Keep hosted photos while that id is resumable, including token expiry on status fetch; prune only after success or a confirmed remote failure.
Co-authored-by: Cursor <cursoragent@cursor.com>
* test: cover remaining TikTok in-flight derivative edge cases
Guard the empty publish_id prune path, account guards without a checkpoint, and video status 401 after /init/.
Co-authored-by: Cursor <cursoragent@cursor.com>
* refactor: map TikTok publish statuses with an official enum
Use PublishStatus for status/fetch values from the Content Posting API. Keep only the documented cases, including FAILED as the terminal failure.
Co-authored-by: Cursor <cursoragent@cursor.com>
* refactor: share in-flight publish checkpoint keys
Read TikTok and Instagram resume state through one helper so publishers, posts:retry, and derivative cleanup agree on the same keys.
Co-authored-by: Cursor <cursoragent@cursor.com>
* fix: retry Instagram media_publish after transient Graph failures
A 500/code 2 after Meta already published left the job Failed as unknown.
Treat that as still-processing so resume can confirm PUBLISHED instead of posting again.
Co-authored-by: Cursor <cursoragent@cursor.com>
* fix: resume Instagram publish after dropped Graph connections
A timeout or connection reset after Meta already published was marked unknown.
Treat it as still-processing so resume can confirm PUBLISHED instead of posting again.
Co-authored-by: Cursor <cursoragent@cursor.com>
---------
Co-authored-by: Cursor <cursoragent@cursor.com>
* chore: gitignore .superpowers/ scratch workspace
Holds per-plan subagent-driven-development artifacts (ledger, briefs,
review packages) — scratch state, not part of the shipped codebase.
* feat: add connection_warning_sent_at to post_platforms
* feat: add PostAtRisk notification type and translations
* fix: add user_id to NotificationPreferenceFactory definition for ->create() support
* feat: add PostAtRisk mailable and email template
* feat: add VerifyUpcomingPostConnections job
* fix: guard VerifyUpcomingPostConnections against transient errors and cross-workspace leaks
- Add a generic \Exception catch around ConnectionVerifier::verify() so a
transient error (e.g. ConnectionException) on one account can't abort
processing of every other at-risk account in the workspace run.
- Eager-load socialAccount.workspace so markAsTokenExpired's observer chain
never lazy-loads it — this only ever manifested once 2+ distinct accounts
were hydrated in a single run (Eloquent only sets preventsLazyLoading on
batch hydration of >1 row), which is exactly the multi-account scenario
this job exists to handle.
- Add covering tests: enabled=false posts are excluded, one workspace's
at-risk posts never leak into another workspace's notification, and an
unexpected exception on one account doesn't stop the rest of the run.
* feat: add social:check-upcoming-connections command and schedule it
* fix: add composite index for the 15-minute upcoming-post connection query
post_platforms(status, connection_warning_sent_at) supports the filter both
VerifyUpcomingPostConnections and social:check-upcoming-connections run every
15 minutes; without it, every run does a full table scan that only grows as
posts accumulate.
* fix: localize the PostAtRisk email's per-account line and label times as UTC
The postsLabel line was the only hardcoded-English content in an otherwise
fully-translated email, and it showed scheduled_at times with no timezone
indicator even though the app stores everything in UTC. Add
mail.post_at_risk.posts_label (pluralized, one entry per locale, mirroring
each locale's existing post_at_risk.subject plural-boundary syntax) and use
trans_choice() to build the line, with a literal " UTC" suffix left
untranslated in every locale like a unit abbreviation.
Also document why content() reassigns the public $atRiskGroups property
instead of using a local variable (Mailable::buildViewData() overwrites
with() data with same-named public properties).
* fix: time-box the warning dedup and guard against orphaned/ownerless rows
- Re-arm connection_warning_sent_at after a day instead of permanently
suppressing it, so a post rescheduled back into the risk window after a
stale warning is re-evaluated instead of silently skipped forever.
- Exclude post_platforms with a null social_account_id from the at-risk
query. With tries=1, dereferencing a null socialAccount relation would
abort the whole workspace run, including already-detected broken accounts.
- Resolve and check the workspace owner before stamping
connection_warning_sent_at, so an ownerless workspace's posts are left
un-warned (available to be picked up once it gets an owner) instead of
being marked "warned" with no notification ever sent.
Applied the same dedup time-boxing and null-account guard to the
social:check-upcoming-connections dispatch query for consistency.
* fix: PostAtRisk email is always English — drop the locale translation layer
config('app.locale')/App::setLocale() is only ever set by the SetLocale
web middleware, which reads a cookie off the incoming HTTP request. Every
Mailable in this branch is built inside a queued job (SendNotification),
which runs outside the HTTP request lifecycle entirely — no middleware,
no cookie, nothing sets the locale there. So content() always resolved
'app.locale' to the static APP_LOCALE default ('en') regardless of the
recipient's actual preference: the 16-locale mail.post_at_risk.* keys
were dead weight from the start, matching an existing (pre-existing,
out of scope here) gap in the sibling WorkspaceConnectionsDisconnected/
AccountDisconnected mailables.
Replaces the trans_choice()/__() calls with plain English strings built
directly in PostAtRisk, and removes the now-unused mail.post_at_risk.*
block from all 16 locale files. Also strengthens the mailable test to
assert the full "N post(s) scheduled: ... UTC" string, not just a
fragment of it.
* refactor: consolidate the two post_platforms migrations from this branch into one
connection_warning_sent_at and its supporting index were added in two
separate migrations (the column in the original task, the index during
final review). Both are still unmerged/unshipped on this branch, so
folding the index into the same migration that adds the column is safe
and keeps the schema change to post_platforms as one unit instead of two.
Verified with a full rollback + re-migrate cycle that the consolidated
up()/down() is self-consistent.
* refactor: add PostPlatform::scopeEnabled(), replace ->where('enabled', true) everywhere
The raw where('enabled', true) clause was duplicated across 17 call sites
in 12 files (13 including the 2 this branch added), all expressing the
same rule PublishPost enforces at publish time: only enabled platforms
are eligible. Added a scopeEnabled() to PostPlatform and swapped every
query-builder call site to ->enabled().
Three call sites are intentionally left untouched: they filter an
already-loaded relation Collection (->postPlatforms->where(...), no
parens), which is Collection::where(), not a query scope — a query scope
can't apply to an in-memory collection.
No inverse (enabled = false) query pattern exists anywhere in the
codebase — 'enabled' => false only ever appears as a write when a post
is disabled/synced, never as a read filter — so no scopeDisabled() was
added; nothing would call it.
* test: cover re-armed post_platform where the account was reconnected
The re-arm dedup fix (connection_warning_sent_at older than a day is
treated as null) only had coverage for "still broken, warns again" and
"too recent, stays skipped". Missing: the row gets re-evaluated (verify()
is called, not skipped) but comes back healthy because the user
reconnected in the meantime — nothing should change (no new warning, no
notification, marker stays at its old value).
* fix: dispatch-level uniqueness, index the enabled filter, close markAsTokenExpired race
From a deep review pass on the whole branch:
- VerifyUpcomingPostConnections now implements ShouldBeUnique (keyed on
workspaceId, 300s window). withoutOverlapping() on the schedule only
serializes the fast-dispatching command; a queue backlog could still let
two jobs for the same workspace run concurrently, both mailing the owner
for the same at-risk posts.
- The composite index now covers enabled too (status, enabled,
connection_warning_sent_at) — every query that uses it filters on all
three, so the index previously required a heap fetch per row just to
check enabled.
- markAsTokenExpired() silently no-ops if it loses the account's status
lock to a concurrent process (a publish attempt, the daily check). The
job used to push the account into the at-risk notification regardless
of whether the update actually landed. It now re-checks the account's
status after the call and only warns if the transition is confirmed —
a lost race just defers the account to the next run instead of sending
a misleading "reconnect" email for an account whose status didn't change.
Also includes an unrelated stray Pint fix (inline \Throwable -> imported)
in SendNotification.php that had been sitting uncommitted.
* refactor: centralize account handle/display name, expose to frontend, close review findings
Adds SocialAccount::handle()/accountDisplayName() plus appended
display_label/handle_label JSON fields, replacing duplicated
username/display_name fallback logic scattered across platform
previews, NetworkConnectGrid, PreviewTab, Calendar, and the post
editor pages.
Also closes the remaining findings from the final review on this
branch: escapes the workspace name in PostAtRisk's intro (and drops
the now-unnecessary raw-HTML rendering), fixes the tautological
"dispatches once per workspace" test, adds plural/subject test
coverage for PostAtRisk, raises VerifyUpcomingPostConnections'
uniqueFor to cover the full schedule cadence, and updates a stale
docblock.
* test: cover draft-post exclusion, account status after PlatformUnavailableException
Adds the two coverage gaps left open by the last review: a post still
in Draft status inside the 1-hour window must not trigger a check or
warning, and a PlatformUnavailableException must leave the account
status untouched. Also drops the dedicated PostAtRisk XSS test — the
intro is now plain Blade-escaped text, so the coverage is redundant
with the framework's own escaping.
* fix: close final review findings — i18n notification, empty-string fallback, missed refactor sites
- Localize the in-app "post at risk" notification title in all 16
locales via trans_choice (the email stays English, unchanged)
- Use ?: instead of ?? in handle()/accountDisplayName()/handleLabel()
so an empty-string username/display_name still falls back, matching
the old Vue || behavior
- Migrate the 3 frontend sites the earlier sweep missed (Index.vue,
SocialAccountsGrid.vue, ScheduleTab.vue) to display_label/handle_label
- Fix avatar-initial fallback in the platform preview components to use
display_label instead of raw display_name
- Correct handle_label's TS type to string | null across 10 files to
match the accessor's actual return type
- Add test coverage for the command-level "already warned" dedup path
and the in-app Notification row created alongside PostAtRisk's email
* fix: notification storm, duplicate-email race, and queue payload bloat in upcoming-post checks
Three correctness issues found by review, fixed after discussion:
- An already-broken account could get a fresh PostAtRisk email every
15 minutes for as long as it stayed broken, if new posts kept
entering the 1-hour risk window. Gated with a per-account 60-minute
renotify cooldown.
- Two concurrent jobs (RefreshExpiringTokens and this one) could each
discover the same dead token and send their own email for it
(AccountDisconnected + PostAtRisk) within the same tick. Gated with
a 5-minute grace period, applied only when another process already
transitioned the account before we got to it — not when we're the
one making the transition.
- PostAtRisk carried full SocialAccount/PostPlatform/Post model
graphs on the queue payload, since SerializesModels can't reduce
models nested inside a plain array/Collection to lightweight
identifiers. It now carries only post_platform IDs and rehydrates
at send time, with envelope()/content() sharing one memoized query
so their counts can't disagree.
Also replaces the account-health cache with a persisted
SocialAccount.last_verified_at column, and narrows the actual
platform API calls to only fire once a post's nearest scheduled_at
is within 30 minutes — enough lead time to reconnect, without
spending API budget checking a full hour out.
* fix: replace dead unsubscribe link with notification preferences, finish display_label sweep
The shared mail footer's unsubscribe link was permanently dead code
(unsubscribe_url was never passed by any Mailable). Replaced it with
a fixed "Manage notifications" link to the real settings page,
via route('app.notifications.preferences').
Also closes out the remaining sites still computing the
username/display_name fallback locally instead of reading the
backend-computed display_label: 8 more Vue components (platform
previews, per-platform post-editor settings, the AI post wizard, the
automation Generate node config, and the analytics account selector)
plus two PHP call sites (PostPlatform::getDisplayNameAttribute(),
already fixed on main before this branch, and the template image
generator's rendered footer text).
* fix: only show "Manage notifications" on preference-driven emails
The link doesn't make sense on transactional emails that always send
regardless of notification preferences (password reset, email
verification) or that go to recipients who may not even have an
account yet (workspace invite) — and the settings page it points to
requires login, which is actively broken for the first two.
Split the shared footer into two Maizzle components: footer.html
(plain) for the 3 transactional templates, footer-authenticated.html
(adds the link) for the 6 that go through SendNotification and
respect the recipient's notification preferences.
* fix: lock PostAtRisk's subject to the dispatch-time count, expose handle_label from analytics
PostAtRisk's subject/previewText were recomputed from a fresh DB
query at send time, while the in-app notification's title (built in
VerifyUpcomingPostConnections::notifyOwner()) used the count observed
at dispatch time. If a post_platform row disappeared in between, the
two could disagree. The count is now passed into the mailable
explicitly and reused for both — the body's account/post details
still rehydrate fresh from the DB, preserving the anti-staleness fix
from earlier in this branch.
Also adds handle_label to AnalyticsController's account payload,
matching every other endpoint that serializes a SocialAccount.
* fix: don't abort the whole workspace run if an account is deleted mid-verify
An exception thrown inside a catch block isn't routed to a sibling
catch, so $account->refresh() throwing ModelNotFoundException (the
user disconnected/deleted the account in the brief window between
this job loading it and handling the TokenExpiredException) escaped
handle() entirely. With tries = 1, that killed the run for every
other account in the same workspace, not just the deleted one.
Also fixes an inconsistent placeholder in PlatformPreview.vue
(handle_label: null instead of '', matching display_label).
* fix: guard against deleted accounts, guarantee a non-empty account name
Closes the last 4 findings from the sixth review round:
- VerifyUpcomingPostConnections now skips a group whose account
resolved to null (deleted between the main query and its eager-loaded
relation), instead of an unguarded property access aborting the
whole workspace's run
- the same job's nested exception handler now covers any \Exception
from markAsTokenExpired() (lock/DB failures), not just
ModelNotFoundException
- PostAtRisk drops a rehydrated group whose account no longer exists
instead of crashing the render (verified: fails without the fix,
passes with it)
- AnalyticsController's handle_label field is now actually consumed by
AnalyticsAccountSelector.vue instead of being unused payload
Also closes a real gap: every connector requests enough OAuth scope to
populate at least one of username/display_name (confirmed for TikTok,
whose account.py comment implied otherwise but whose connect() scopes
always include user.info.profile), so accountDisplayName()/handle()/
displayLabel/handleLabel now return a guaranteed non-empty string
(falling back to the platform label only as a last resort) instead of
being nullable. This removes the now-pointless @if guards around
accountDisplayName() in the account-disconnected and post-at-risk
email templates, and lets ~30 frontend files drop the `| null` from
display_label/handle_label and the ?? undefined fallbacks that only
existed to satisfy that type.
* fix: drop the now-pointless ?? '' fallback on display_label in TemplateImageGenerator
display_label is a guaranteed non-empty string (see 950558b4).
* fix: correct social_account's TS type to nullable in Index.vue and Calendar.vue
Both declared social_account as required while their own templates
used optional chaining (pp.social_account?.display_label) — the type
was lying. social_account_id is nullable and the account can be
deleted (FK is nullOnDelete), so the field genuinely can be null.
Swept every other social_account/socialAccount field in resources/js
for the same mismatch; all others already declared it correctly.
* Centralize avatar-initial extraction via getInitials()
Replace hand-rolled .charAt(0)/.charAt(0).toUpperCase() avatar-initial
logic across social account previews, the accounts grid, the analytics
account selector, and the mention picker with the existing
useInitials() composable already used by Avatar.vue.
* Drop pointless display_label fallbacks now that it's always populated
display_label is guaranteed non-empty (falls back to the platform
label server-side), so || 'Channel' / || 'TryPost' / ?? platform were
unreachable.
* Fix cold-review findings: dead handle_label guard, slug leak, wrong post count
- AnalyticsAccountSelector: the "@handle" line's guard/value must read the
raw username (nullable — Facebook Pages and Telegram channels legitimately
have none), not handle_label, which always resolves to something and made
the guard permanently true. Drop the now-orphaned handle_label field from
the analytics payload/type since nothing else in analytics used it.
- PlatformPreview: the no-account-selected fallback now uses
getPlatformLabel() instead of the raw platform slug, matching the
backend's own last-resort label fallback.
- VerifyUpcomingPostConnections: count distinct posts (post_id), not
post_platform rows, so one post spanning multiple broken accounts doesn't
inflate the at-risk count in the email subject and notification title.
* Fix cold-review round 2: silent Telegram/Discord false negative, flaky email ordering, dead display_name
- VerifyUpcomingPostConnections: ConnectionVerifier::verify() reports a
dead Telegram/Discord connection by returning false rather than
throwing. The job discarded that return value, so a bot removed from
a channel/guild was stamped last_verified_at and silently trusted
healthy for the next 40 minutes — no warning, post just fails at
publish time. Route a false return through the same
TokenExpiredException handling used by every other platform.
- PostAtRisk: atRiskGroups() had no ORDER BY, so the per-account
"N posts scheduled: H:i, H:i UTC" line rendered in arbitrary
(physical row) order. Sort by scheduled_at before formatting.
- Drop the orphaned display_name field from the analytics payload/type
(superseded by display_label; nothing in resources/js/components/
analytics or pages/analytics read it).
* Add social icons and copyright to email footers
Icons match the trypost-site footer (outline @tabler/icons style,
converted to PNG since email clients — notably Outlook desktop — don't
render inline SVG). Reordered footer content: tagline, manage-notifications
link, icons as the closing element, copyright line last.
* Standardize connection-verify error classification across all 13 platforms
Every platform now follows one contract: verify() returns true on a
healthy connection, throws TokenExpiredException only on a confirmed
dead connection, and PlatformUnavailableException on anything else
(rate limit, 5xx, unrecognized). Previously most platforms silently
returned false on anything but a 401, so callers (all of which only
react via try/catch) could never distinguish "definitely dead" from
"transient" — and Telegram/Discord never threw at all.
Each platform's "is this confirmed dead" check now lives next to its
existing publish-time error classifier (App\Exceptions\Social\*PublishException)
instead of being re-typed inline in ConnectionVerifier, closing real,
already-drifted gaps between the two paths:
- TikTok and Mastodon both had a bare "status === 401/403" check shared
between publish and verify, but TikTok's scope_not_authorized and
Mastodon's write-scope 403 use the same status for a non-fatal scope
gap, not a dead token — verify's lower-privilege endpoint keeps its
own stricter check on top instead.
- Telegram/Discord authenticate with one bot token shared across every
connected account; a 401 means that shared token is misconfigured
(an operator problem), never that one specific account is broken —
excluded from both platforms' confirmed-dead checks accordingly.
- Facebook/InstagramFacebook/Mastodon/Telegram/Discord have no
per-account refresh flow at all, so a confirmed rejection now skips
the pointless refresh-and-retry (Platform::hasTokenRefreshFlow()).
Also fixes two bugs found while hardening VerifyUpcomingPostConnections:
a post hard-deleted mid-run could crash the whole job for every other
account in the batch (now filtered per group), and two overlapping runs
of the same job could send duplicate PostAtRisk warnings (now a
conditional claim on connection_warning_sent_at).
* Skip paused accounts in upcoming-post connection checks, close claim race
A paused (is_active=false) social account already fails at publish time
before any platform API call, so it shouldn't trigger a proactive
connection check or "reconnect" warning. Guard added at dispatch time
(CheckUpcomingPostConnections) and re-checked fresh mid-run inside
VerifyUpcomingPostConnections's per-account loop, since the job can take
real wall-clock time working through a workspace and an account can be
paused or deleted after the query-time guard already ran.
Also wraps the connection_warning_sent_at claim in a SELECT ... FOR UPDATE
transaction (ordered by id, 3 retries) to close a race between two
overlapping runs of the same job double-claiming and double-emailing about
the same post_platform.
* Clarify "commit" wording in claim-transaction comment
Reads ambiguously as a git commit on a PR diff; it means the DB
transaction commit.
* fix: detect dead Threads/Instagram/Facebook tokens reported under non-190 codes
verifyThreads/verifyInstagram/verifyFacebook only threw TokenExpiredException
for Meta error code 190, silently returning false for every other rejection
(e.g. code 100 "The requested resource does not exist"). The hourly
VerifyWorkspaceConnections check never saw that false, so a genuinely dead
token went unflagged — no reconnect email — until the real scheduled post
tried to publish and failed with the same raw error (#230).
GraphError::isTransient() now isolates the known rate-limit/transient codes
(1, 2, 4, 17); everything else on a failed verify/refresh is a confirmed
rejection and raises TokenExpiredException, while transient/5xx/429 raises
PlatformUnavailableException so the account isn't disconnected on a throttle.
Also drops the unused $errorType variable from the three *PublishException
classes.
* fix: treat unparseable Meta failure bodies as transient, drop dead code
Code review on #254 found two issues in the original fix:
- The inverted classifier (`! GraphError::isTransient($body)`) treated a
response body that fails to parse as JSON (WAF block page, truncated
response, gateway hiccup) as a confirmed dead token, since isTransient()
returns false for a body it can't recognize. That flipped a null/unparseable
body from "retry later" (PlatformUnavailableException, the pre-fix behavior)
to "disconnect now" (TokenExpiredException) for both the Threads/Instagram
refresh classifiers and the verify path's classifyMetaVerifyFailure. Fixed
by treating a null body as transient at both call sites — we have no
confirmed rejection from Meta to act on.
- GraphError::indicatesInvalidToken() had no remaining production callers
after the refresh classifiers switched to isTransient() — removed it and
its tests instead of leaving dead code behind.
* test: symmetric Facebook/Instagram coverage for the shared verify classifier
verifyInstagram/verifyFacebook/verifyThreads all delegate to the same
classifyMetaVerifyFailure(), so the non-190 dead-token, rate-limit,
5xx, and non-JSON-body cases were only exercised end-to-end for
Threads. Adds the missing Facebook (rate-limit, 5xx, non-JSON) and
Instagram (non-190 dead token, 5xx) cases so each platform has direct
proof, not just shared-code inference.
* fix: recognize Business Use Case (BUC) rate-limit codes for Page-token accounts
Verified the transient-code list against Meta's official docs. Confirmed:
codes 1, 2, 4, 17, 190 match what's documented at
developers.facebook.com/docs/graph-api/guides/error-handling/. But Meta runs
a SECOND, separately-coded rate-limit system (Business Use Case / BUC) for
Page and system-user tokens — which is exactly what our Facebook and
InstagramFacebook accounts use. BUC rejections come back as a plain HTTP 400
(not 429) with codes in the 80000 range (80001 Pages API, 80005 Instagram
Platform), which GraphError::isTransient() didn't recognize — meaning a
throttled Facebook/InstagramFacebook Page token would have been misclassified
as a confirmed dead token and disconnected.
- Added 80001/80005 to GraphError::TRANSIENT_CODES, with sources.
- Added GraphError::isTransientFailure(Response) to fold the status-based
checks (5xx, 429) and body-based checks together into one documented
method, replacing the ad-hoc multi-condition `if` that lived inline in
ConnectionVerifier::classifyMetaVerifyFailure().
- isTransient() now treats a null (unparseable) body as transient directly,
so the refresh-path classifiers no longer need a separate null guard.
- Documented the full code table, sources, and per-platform token-type
notes (Page token vs. user token, which rate-limit system applies to
which platform) in GraphError's class docblock and in CLAUDE.md, so
future changes here start from verified sources instead of guessing.
* refactor: move Meta verify-failure classification into GraphError
classifyMetaVerifyFailure() lived in ConnectionVerifier but never touched
$this, SocialAccount, or the cache lock — it was a pure (Response, label) ->
Exception translation, same shape as what TokenRefreshClient already owns
for the refresh side. Keeping it in ConnectionVerifier broke that symmetry
and split Meta error interpretation across two classes instead of the one
(GraphError) whose docblock already says that's its job.
Moved as GraphError::classifyVerifyFailure(), dropped the now-unused
Response import from ConnectionVerifier, and added direct unit tests for
the new public method alongside the existing ConnectionVerifierTest
coverage that exercises it through verify().
* fix: correct Instagram Platform BUC code from 80005 to 80002
My earlier WebFetch of Meta's rate-limiting page mis-parsed the BUC code
table and mapped 80005 to Instagram Platform. It's actually Lead Generation
(Marketing API, which this app never calls) — Instagram Platform is 80002.
Verified against a raw, unsummarized reproduction of the same official page
(developers.facebook.com/docs/graph-api/overview/rate-limiting/) plus
independent third-party corroboration, both pointing to 80002.
Also closes a test-coverage gap flagged in review: GraphError::isTransient()
now intentionally treats a parseable body with no "error" key (e.g.
{"data": {...}}) as a confirmed rejection, not transient — a real behavior
change from the pre-#254 code, which silently ignored that shape. Added
explicit unit + integration coverage for it so the decision is asserted,
not implicit.
* Fix Facebook and Instagram-via-Facebook Page connect pagination.
Follow Graph API paging.next on /me/accounts so authorized non-first Pages are found and multi-Page accounts get the picker instead of silently connecting the first result.
Co-authored-by: Cursor <cursoragent@cursor.com>
* Paginate Meta accounts until paging.next is exhausted.
Drop the artificial 50-page cap and stop only when there is no next URL, or the same request URL repeats (broken pagination loop).
Co-authored-by: Cursor <cursoragent@cursor.com>
* Redact tokens in Graph pagination logs and harden test coverage.
Cover happy-path and failure cases for Meta /me/accounts pagination, including mid-loop failures, invalid paging.next, and Instagram pages without a linked IG account.
Co-authored-by: Cursor <cursoragent@cursor.com>
* Fail closed on incomplete Meta accounts pagination.
If a later /me/accounts page fails after earlier pages succeeded, throw instead of returning a truncated list that could auto-connect the wrong Page. Also revert the IG detail timeout that could wipe the whole connect list.
Co-authored-by: Cursor <cursoragent@cursor.com>
* Simplify Graph pagination helpers and page fetchers.
Bake the first request query into the URL, drop requestKey, and let IncompleteGraphPaginationException bubble from the controllers without catch/rethrow noise.
Co-authored-by: Cursor <cursoragent@cursor.com>
* Move incomplete pagination exception under Social\Meta.
Colocate it with GraphPaginator so the Meta scope is clear from the namespace instead of a generic Social exception name.
Co-authored-by: Cursor <cursoragent@cursor.com>
* Rename pagination exception to IncompleteMetaGraphPaginationException.
Keep it under Exceptions/Social with Meta in the class name instead of moving it into Services.
Co-authored-by: Cursor <cursoragent@cursor.com>
* Make GraphPaginator results explicit before mapping pages.
Assign the paginated accounts to a variable first so the Facebook and Instagram-via-Facebook fetchers read more clearly.
Co-authored-by: Cursor <cursoragent@cursor.com>
* Build Meta Graph pagination URLs with Laravel Uri.
Replace manual http_build_query concatenation with Uri::of()->withQuery().
Co-authored-by: Cursor <cursoragent@cursor.com>
* Use Laravel HTTP and Uri helpers in Meta Graph pagination.
Prefer response collect/json key access, filled(), and Uri path parsing over manual array and parse_url handling.
Co-authored-by: Cursor <cursoragent@cursor.com>
* Simplify graphVersion using Uri path and str().
Drop basename and native string casts; Uri::path() already yields the Graph API version segment.
Co-authored-by: Cursor <cursoragent@cursor.com>
* Drop unnecessary str() around graph API config.
Uri: :of() already accepts the string returned by config().
Co-authored-by: Cursor <cursoragent@cursor.com>
* Simplify GraphPaginator with Laravel helpers.
Consolidate failure handling via abort(), and use collect, when, throw_if, and Uri::value() for a shorter pagination loop.
Co-authored-by: Cursor <cursoragent@cursor.com>
* Refactor social OAuth page/channel selection handling. Update selectPage and selectChannel methods in Facebook, Instagram, and YouTube controllers to return popup callbacks instead of redirecting on session expiration or workspace not found. Enhance HandleInertiaRequests middleware to prevent deferring onboarding progress on social OAuth popup routes. Add tests to verify behavior for expired sessions and onboarding progress.
* Unify Instagram connect behind one card with a method picker.
Hide the Instagram-via-Facebook grid card and offer Instagram Login vs Facebook Pages from a single network entry, matching LinkedIn.
Co-authored-by: Cursor <cursoragent@cursor.com>
* Move social popup onboarding assertions into connection tests.
Cover the deferred-prop popup regression on Facebook, Instagram, and YouTube select routes instead of a synthetic onboarding share check.
Co-authored-by: Cursor <cursoragent@cursor.com>
* Stop suppressing onboarding defer on all social routes.
Override onboardingProgress only in popupCallback so picker pages stay deferred and the close page does not re-hit select after session clear.
Co-authored-by: Cursor <cursoragent@cursor.com>
* Always open the Instagram method dialog on connect.
Drop connectMethods and the single-method OAuth shortcut; the picker always offers both Login and Facebook Pages.
Co-authored-by: Cursor <cursoragent@cursor.com>
* Filter Instagram dialog options by enabled platforms.
Keep always opening the method picker, but only list OAuth entry points that are turned on.
Co-authored-by: Cursor <cursoragent@cursor.com>
* Extract Instagram connect methods into a dedicated helper.
Keep connectableOptions focused on shaping grid options while the enabled OAuth list lives in instagramConnectMethods().
Co-authored-by: Cursor <cursoragent@cursor.com>
* Harden Meta Graph pagination and localize Instagram connect copy.
Fail closed on Graph request errors and pathological paging, keep Instagram connect going when profile detail lookups time out, and translate the Instagram method dialog strings.
Co-authored-by: Cursor <cursoragent@cursor.com>
---------
Co-authored-by: Cursor <cursoragent@cursor.com>
Fail hard when media processing fails, map media-specific X invalid-request
errors clearly, and cover GIF/large-image/amplify/finalize/append failure paths.
Co-authored-by: Cursor <cursoragent@cursor.com>
Pricing
- Bill per workspace ($12/mo or $120/yr each); Stripe quantity tracks the
workspace count and syncs on workspace create/delete.
- 2,500 AI credits per workspace, pooled at the account level; monthly reset
on the billing anniversary, annual granted upfront (no rollover).
- One social account per network per workspace; remove all count-based limits
(workspace/social/member) and the legacy plan tiers (single Workspace plan).
Onboarding (cloud only: SELF_HOSTED=false + PostHog)
- Replace the /subscribe plan picker with /onboarding persona selection
(Creator/Freelancer/Startup/Agency/Small business/Other), saved on the user
(users.persona) and mirrored to PostHog, then Stripe Checkout on the monthly
price. 8-day trial so Stripe displays 7.
Billing screen
- Remove the Change Plan dialog (dead with a single plan); add an annual-upgrade
banner for monthly subscribers (swapToYearly).
- Current-plan card shows the workspace count instead of the plan name.
System AI
- Brand analyzer / workspace autofill is always allowed and never debits credits
(system feature, not the user's usage).
Self-hosted (SELF_HOSTED=true) bypasses all billing, credit, limit, network,
and onboarding logic.
Connect a Discord server via OAuth (bot authorization) and schedule/publish
messages to its channels, with mentions and rich embeds.
- Connect: custom Socialite Discord provider (bot scope) maps the authorized
guild to a SocialAccount; throws if no server was authorized.
- Publish: DiscordPublisher posts via the global bot token, validates the chosen
channel belongs to the connected guild (anti cross-guild), optimizes media,
builds allowed_mentions only from explicit mention chips (no accidental pings),
and renders rich embeds.
- Compose: per-post channel picker (live lookup), mention autocomplete and an
embed editor, gated by a required-channel compliance rule; Discord post preview.
- Enum/config/content-type wiring, ConnectionVerifier health check, throttled
lookup endpoints, i18n (en/es/pt-BR), and tests.
Operators must create a Discord application and set DISCORD_CLIENT_ID,
DISCORD_CLIENT_SECRET, DISCORD_BOT_TOKEN and DISCORD_CLIENT_REDIRECT.
Register Telegram as a platform: Platform/ContentType enum cases, a
platforms.telegram config block (shared bot token via env), TelegramPublisher
(sendMessage / sendPhoto|Video|Document / sendMediaGroup over the Bot API, HTML
parse mode, 4096 limit with long text split off a 1024 caption), wired into the
publisher dispatch. Add a Telegram ContentSanitizer branch (Telegram-allowed
HTML + ampersand escaping), MediaOptimizer/profile-url/factory support, and the
TelegramPublishException. Tests cover text, single media, album, long-text split,
overflow, API errors, private-channel URLs, and sanitization.
Same shape of bug as the refreshToken duplication: the regex that strips
access_token / Bearer headers from logged HTTP bodies existed in three
near-identical copies (HasSocialHttpClient trait, TokenRefreshClient,
SocialPublishException), drifting subtly — SocialPublishException was
missing the JSON "token" pattern.
Extracts a TokenRedactor::redact(string) helper and routes all callers
through it. Adding a new token format now means one regex in one file.
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.
Production was returning HTTP 413 with empty body on the first APPEND
segment of every video upload to X v2 — surfacing to users as 'An
unknown X error occurred.'
Empty-body 413 is the classic signature of an edge/CDN rejection: the
X gateway is denying the request before X's application code sees it.
The X v2 reference docs say 'max chunk size: 5MB', but every canonical
reference uses 1MB:
- X's official Python quickstart: `chunk_size = 1024 * 1024`
- X's official JavaScript quickstart: `const chunkSize = 1024 * 1024`
- twitter-api-v2 (the most-used Node SDK, used by Postiz et al.):
`chunkSize: number = 1024 * 1024`
5MB plus multipart-form overhead apparently exceeds an undocumented
edge limit. 1MB is the empirically safe size everyone converges on.
Changes:
- XPublisher chunked APPEND now uses 1MB chunks. An 8MB video uploads
as 8 segments instead of 2; more roundtrips but actually succeeds.
- Set explicit Content-Type on each chunk attach (matches the simple
upload path in the same file).
- XPublishException::fromApiResponse maps HTTP 413 to
ErrorCategory::MediaFormat with the message 'Media chunk rejected
by X (payload too large).' so we don't surface 413 as 'unknown' if
it ever recurs.
Test: unit test covering the 413→MediaFormat mapping. Full suite green
(1503 passed, 2 skipped).
Mentions in post comments
- @mention autocomplete (workspace members, current user excluded) with
marker syntax @[uuid] persisted, display names rendered via CommentBody
chips; live edit replaces markers with names and converts back on save.
- NotifyMentions action with workspace-scoped membership check, dedupes
same user, only newly-added mentions on update.
- Email + in-app via SendNotification job, respecting per-user
notification_preferences.mentioned_in_comment.
- Heartbeat-based presence (Cache, 60s TTL, 30s ping) so online recipients
get only the in-app notification — no email noise.
- Real-time bell on workspace.{id}.user.{id} private channel
(NotificationCreated event), scoped channel name avoids client-side
filtering and lays out a convention for future workspace channels.
- Mailable localized via lang/{en,es,pt-BR}/mail.php; Maizzle source
template for the email is committed and built into resources/views/mail.
AI generation refactor (Action layer + MCP)
- Extracted Actions/Ai/Generate{Image,Video} with QuotaExhaustedException
so agent tools and MCP tools share a single domain entry point.
- Mcp/Tools/Ai/Generate{Image,Video}Tool registered in TryPostServer; both
return MediaResource payloads.
- Orientation::imageApiSize maps non-OpenAI ratios to 1:1/2:3/3:2.
- config/ai.php is now the single source of truth driven by env, removing
the trypost.ai shim. Default text/image providers flipped to OpenAI.
Settings/UX
- /settings/workspace split into shadcn Tabs (Workspace / Brand / Users)
with three components.
- /assets and the in-editor MediaPicker open the ImagePreviewDialog
lightbox on image click while preserving action button behaviour.
- Comments tab landed via ?tab=comments&comment=<id> from notification
click (scroll-to + temporary highlight).
- Mention autocomplete popover flips above when near the viewport bottom.
- Real social platform PNGs replace Tabler brand glyphs in schedule
pills and post list, with hover tooltip carrying display_name + handle.
Bug fixes
- AcceptInvite: controller now passes workspace + role payload that the
Vue page expects; login/register CTAs preselect the invite email.
- WorkspaceInvite mailable: stopped referencing nonexistent
$invite->workspace and $invite->role; column added to the migration,
Invite model casts role to WorkspaceRole, CreateInvite persists it.
- PostCommentCreated: added broadcastAs so .PostCommentCreated actually
matches the Echo listener; payload now includes mentioned_users so
receivers render the chip correctly without a refetch.
- Preview components for X/Pinterest/Threads/Bluesky/LinkedIn/Mastodon/
TikTok/YouTube switched from item.type === 'image' to
!isVideoMedia(item) so media without a persisted type still renders.
- UpdatePostRequest now accepts media.*.{type,mime_type,size,...} so the
posts.media JSON keeps the metadata that the previews need.
- Removed throttle:6,1 from social connect routes (was 429ing legitimate
OAuth retries).
- Used MediaType enum cases instead of literal 'image'/'video' strings
when creating media rows.
Tests
- MentionParser unit tests, NotifyMentions feature tests including
online/offline channel selection and preference gating, MCP AI tool
happy paths, MentionedInComment mailable rendering, AcceptInvite +
search-members + index mentioned_users path. 1229 passing.
Settings pages:
- Redesign layout to match Sendkit (max-w-4xl, space-y-12, Separator sections)
- Merge Members page into Workspace settings with Table, invite Dialog, ConfirmDeleteModal
- Add workspace logo upload/delete routes and controller methods
- Translate all hardcoded strings in Workspace.vue modals
Language system:
- Drop languages table, replace language_id FK with locale string column on users
- Create config/languages.php for available languages and default locale
- Add Spanish (es) translations (13 files)
- Simplify HandleInertiaRequests, ProfileController, RegisteredUserController
Code quality:
- Add declare(strict_types=1) to all PHP files
- Fix MastodonPublisher using wrong attribute (filename -> original_filename)
- Fix HasMediaTest for new has_photo/photo_url accessors
- Fix PublishToSocialPlatformTest type error revealed by strict_types
- Remove orphaned Language model from AppServiceProvider morph map
- Update User TypeScript interface (has_photo, photo_url, locale)
- Eager load media relation on workspaces to prevent N+1
- Add 8 new tests for workspace logo upload/delete
- Update workspace settings test to assert members/invitations props
All 710 tests passing.
Paulo Castellano