This patch adds 3 new options to OIDC settings.
Toggle user auto-creation, and select default user/list roles
for auto-created users.
Co-authored-by: Kailash Nadh <kailash@nadh.in>
Pausing a campaign and starting it within the interval where campaign manager
pipe cleanup happens would cause a campaign to be marked as finished. This
patch addresses that by only doing the 'finished' state update if the campaign
wasn't stopped (cancel, pause) manually.
Potentially closes#2278 as this is the only reproducible scenario where a
campaign prematurely gets marked as finished.
- Fix merge conflicts.
- Simply logic in campaign preview handler.
- Remove redundant `GetCampaignForPreviewWithTemplate()` and switch to the old
query that optionally takes a template.
- Fix Vue linting issues.
This permission was never checked for and had an unintended consequence of
allowing a non-superadmin user to execute arbitrary queries (expected), but
getting a superadmin session by joining the `sessions` table.
This patch:
- Introduces a table allowlist that uses the Postgres query plan (JSON)
and validate the referenced tables against the allowed ones on arbitrary
queries issued to the various `/subscribers` APIs.
- Explicitly adds the missing `subscribers:sql_query` permission check to all
handlers that accept `query`.
- Introduces a new `search` parameter on all handlers that accept `query`.
This parameter is an interface over the default name/email substring search
instead of relying on `query`.
- During install, listmonk now accepts the env `LISTMONK_ADMIN_API_USER`
and creates an API user (with username $LISTMONK_ADMIN_API_USER)
with full superadmin permissions. This requires LISTMONK_ADMIN_USER and
LISTMONK_ADMIN_API_PASSWORD to be set so that that there's always a superadmin
user to avoid bad states, mainly: bot superadmin exists, but no admin user
exists, leaving the installation perpetually open with the superadmin user
creation UI on the first login.
The API user's token is printed to stderr in the following format:
`export LISTMONK_ADMIN_API_TOKEN="7I81VSd90UWhKDj5Kq9c6YopToRduyDF"`
This can be redirected to a file with ./listmonk 2> /tmp/token or captured
directly and then source()'d.
- Add new function `core.GetRole(id)`.
- Fix `at least one super admin` query in user deletion.
- Make the beginning of handlers consistent with uniform variable declaration
and grouping.
- Add missing comments.
- Fix staticcheck/vet warnings and idiom issues.
- Move user models from `/models` to `internal/auth`.
- Move and refactor various permission check functions into `User.()`
- Refactor awkward `get, manage bool` function args into `Get|Manage` bitflags.
This patch introduces new `campaigns:get_all` and `campaigns:manage_all`
permissions which alter the behaviour of the the old `campaigns:get` and
`campaigns:manage` permissions. This is a subtle breaking behavioural change.
Old:
- `campaigns:get` -> View all campaigns irrespective of a user's list
permissions.
- `campaigns:manage` -> Manage all campaigns irrespective of a user's list
permissions.
New:
- `campaigns:get_all` -> View all campaigns irrespective of a user's list
permissions.
- `campaigns:manage_all` -> Manage all campaigns irrespective of a user's list
permissions.
- `campaigns:get` -> View only the campaigns that have at least one list to
which which a user has get or manage access.
- `campaigns:manage` -> Manage only the campaigns that have at list one list
to which a user has get or manage access.
In addition, this patch refactors and cleans up certain permission related
logic and functions.
- Fix status/button state management issues when `Send at` was toggled
under various scenarios.
- Allow paused campaigns to be edited and turned into scheduled campaigns.
- Add Cypress UI tests for unscheduling.
This patch introduces a new `Domain allowlist` input in Settings -> Privacy UI
as a new tab alongside domain `Domain blocklist`. If any domains are entered
here, then only subscriptions/imports/additions of e-mails from those particular
domains are accepted. blocklist is mutually exclusive with allowlist when there
are values in the allowlist.
This patch removes the forced suffixing of all media upload filenames with
random strings. The upload handler now checks the `media` table to ensure that
the filename being uploaded doesn't exist before forcing a suffix.
Outright rejecting duplicate filenames cannot be done to maintain backwards
compatibility with the old behaviour.
Closes#2277.
This patch adds a new optional `name` field to SMTP server config on the UI.
When a name is given to an SMTP server, it's initialized as a standalone messenger
which shows up as a sub-group item under the main "email" messenger
on the campaign page.
Co-authored-by: Kailash Nadh <kailash@nadh.in>
This has been a hair-pulling rabbit hole of an issue. #1931 and others.
When the `next-campaign-subscribers` query that fetches $n subscribers
per batch for a campaign returns no results, the manager assumes
that the campaign is done and marks as finished.
Marathon debugging revealed fundamental flaws in qyery's logic that
would incorrectly return 0 rows under certain conditions.
- Based on the "layout" of subscribers for eg: a series of blocklisted
subscribers between confirmed subscribers.
A series of unconfirmed subscribers in a batch belonging to a double
opt-in list.
- Bulk import blocklisting users, but not marking their subscriptions
as 'unsubscribed'.
- Conditions spread across multiple CTEs resulted in returning an
arbitrary number of rows and $N per batch as the selected $N rows
would get filtered out elsewhere, possibly even becoming 0.
After fixing this and testing it on our prod instance that has
15 million subscribers and ~70 million subscriptions in the
`subscriber_lists` table, ended up discovered significant inefficiences
in Postgres query planning. When `subscriber_lists` and campaign list IDs
are joined dynamically (CTE or ANY() or any kind of JOIN that involves)
a query, the Postgres query planner is unable to use the right indexes.
After testing dozens of approaches, discovered that statically passing
the values to join on (hardcoding or passing via parametrized $1 vars),
the query uses the right indexes. The difference is staggering.
For the particular scenario on our large prod DB to pull a batch,
~15 seconds vs. ~50ms, a whopping 300x improvement!
This patch splits `next-campaign-subscribers` into two separate queries,
one which fetches campaign metadata and list_ids, whose values are then
passed statically to the next query to fetch subscribers by batch.
In addition, it fixes and refactors broken filtering and counting logic
in `create-campaign` and `next-campaign` queries.
Closes#1931, #1993, #1986.
