## Description
**Pending testing on preview**
Enable yearly team pricing during checkout.
## Changes Made
- Enable selecting yearly team pricing during checkout
- Update pricing on marketing to $480 for yearly team plan
## Testing Performed
- Stripe simulations
Pending manual testing
<!-- This is an auto-generated comment: release notes by coderabbit.ai
-->
## Summary by CodeRabbit
- **New Features**
- Updated the savings message in the pricing table to reflect new
savings amounts.
- Adjusted yearly pricing from $500 to $480.
- Modified user addition costs based on subscription periods.
- **Improvements**
- Simplified button text in the Create Team Checkout Dialog to
consistently display "Checkout".
<!-- end of auto-generated comment: release notes by coderabbit.ai -->
## Description
Update the direct template signing process and emails to correctly
reflect the role of the recipient who actioned the direct template.
<!-- This is an auto-generated comment: release notes by coderabbit.ai
-->
## Summary by CodeRabbit
- **New Features**
- Dynamic updating of title and description based on recipient role in
the document signing process.
- Enhanced email templates to include recipient roles, providing more
context in email notifications.
- **Improvements**
- More descriptive actions in email templates based on recipient roles,
improving clarity for recipients.
<!-- end of auto-generated comment: release notes by coderabbit.ai -->
## Description
Direct templates links is a feature that provides template owners the
ability to allow users to create documents based of their templates.
## General outline
This works by allowing the template owner to configure a "direct
recipient" in the template.
When a user opens the direct link to the template, it will create a flow
where they sign the fields configured by the template owner for the
direct recipient. After these fields are signed the following will
occur:
- A document will be created where the owner is the template owner
- The direct recipient fields will be signed
- The document will be sent to any other recipients configured in the
template
- If there are none the document will be immediately completed
## Notes
There's a custom prisma migration to migrate all documents to have
'DOCUMENT' as the source, then sets the column to required.
---------
Co-authored-by: Lucas Smith <me@lucasjamessmith.me>
## Description
General enhancements for templates.
## Changes Made
Added the following changes to the template flow:
- Allow adding document meta settings
- Allow adding email settings
- Allow adding document access & action authentication
- Allow adding recipient action authentication
- Save the state between template steps similar to how it works for
documents
Other changes:
- Extract common fields between document and template flows
- Remove the title field from "Use template" since we now have it as
part of the template flow
- Add new API endpoint for generating templates
## Testing Performed
Added E2E tests for templates and creating documents from templates
## Description
Update the template flow to allow for entering recipient placeholder
emails and names
## Changes Made
- General refactoring
- Added advanced recipient settings for future usage
## Description
Currently if you complete a pending encrypted document, it will prevent
the document from being sealed due to the systems inability to decrypt
it.
This PR disables uploading any documents that cannot be loaded as a
temporary measure.
**Note**
This is a client side only check
## Changes Made
- Disable uploading documents that cannot be parsed
- Refactor putFile to putDocumentFile
- Add a flag as a backup incase something goes wrong
## Description
Currently deleting a pending document where you are a recipient off will
delete the document, but will also throw an error.
This is due to the recipient being updated after the document deleted,
which is only supposed to happen for completed documents.
## Description
Feature flags are broken on SSR due to this error
```
TypeError: fetch failed
at Object.fetch (node:internal/deps/undici/undici:11731:11)
at process.processTicksAndRejections (node:internal/process/task_queues:95:5) {
cause: RequestContentLengthMismatchError: Request body length does not match content-length header
at write (node:internal/deps/undici/undici:8590:41)
at _resume (node:internal/deps/undici/undici:8563:33)
at resume (node:internal/deps/undici/undici:8459:7)
at [dispatch] (node:internal/deps/undici/undici:7704:11)
at Client.Intercept (node:internal/deps/undici/undici:7377:20)
at Client.dispatch (node:internal/deps/undici/undici:6023:44)
at [dispatch] (node:internal/deps/undici/undici:6254:32)
at Pool.dispatch (node:internal/deps/undici/undici:6023:44)
at [dispatch] (node:internal/deps/undici/undici:9343:27)
at Agent.Intercept (node:internal/deps/undici/undici:7377:20) {
code: 'UND_ERR_REQ_CONTENT_LENGTH_MISMATCH'
}
}
```
I've removed content-length header since it isn't mandatory to my
knowledge for get requests.
## Changes
- Add fallback local flags when individual flag request fails
- Add error logging
- Remove `content-length` from headers being passed to Posthog
## Description
Currently users can sign and complete draft documents, which will result
in a completed document in an invalid state.
## Changes Made
- Prevent recipients from inserting or uninserting fields for draft
documents
- Prevent recipients from completing draft documents
- Remove ability to copy signing tokens unless document is pending
<!-- This is an auto-generated comment: release notes by coderabbit.ai
-->
## Summary by CodeRabbit
- **New Features**
- Enhanced document status visibility and control across various
components in the application. Users can now see and interact with
document statuses more dynamically in views like `DocumentPageView`,
`DocumentEditPageView`, and `DocumentsDataTable`.
- Improved document signing process with updated status checks, ensuring
actions like signing, completing, and removing fields are only available
under appropriate document statuses.
- **Bug Fixes**
- Adjusted document status validation logic in server-side operations to
prevent actions on incorrectly stated documents, enhancing the overall
security and functionality of document processing.
<!-- end of auto-generated comment: release notes by coderabbit.ai -->
## Description
Fix the issue where duplicate templates can be created
## Changes Made
- Prevent duplicate templates by correctly disabling elements
- Clear the PDF when the form is reset
- General UI changes to new template dialog for consistency
- Add description for new template dialog
- Add close button for new template dialog
## Testing Performed
Manual testing
## Checklist
- [X] I have tested these changes locally and they work as expected.
- [X] I have followed the project's coding style guidelines.
<!-- This is an auto-generated comment: release notes by coderabbit.ai
-->
## Summary by CodeRabbit
- **New Features**
- Introduced new components in the template creation dialog for better
usability: `DialogClose`, `DialogDescription`, and `DialogFooter`.
- Enhanced file upload handling and display.
- Added a cancel button and a create template button to the dialog
footer for improved navigation.
- **Refactor**
- Updated the dialog content and form layout for a more intuitive user
experience.
- Refactored form structure and labels for clarity.
- **Removed Features**
- Removed the `Label` component from the dialog.
<!-- end of auto-generated comment: release notes by coderabbit.ai -->
## Description
When document audit logs were first introduced, we by default set the
`fieldSecurity` to `NONE`
Now that document auth has been added, this is causing issues since we
do not use `NONE` to define field that has no migrations required, but
rather have the `fieldSecurity` field itself be undefined.
To keeps things consistent, this migration replaces `NONE` with
undefined.
There are a few ways to approach this:
- Run a prisma migration on the JSON
- Modify the data before we pass the data to the schema in
`parseDocumentAuditLogData`
- Use `NONE` instead of undefined
If anyone thinks there's a better way to do this, please drop a comment
🙇
## Description
General UI updates
## Changes Made
- Add consistent spacing between document edit/view/log pages
- Add document status to document audit log page
- Update document loading page to reserve space for the document status
below the title
- Update the document audit log page to show full dates in the correct
locale
## Changes Made
- Refactor/optimise tests
- Reduce flakiness
- Add parallel tests (if there's enough CPU capacity)
- Removed explicit worker count when running parallel tests. Defaults to
50% of CPU capacity.
Might want to consider sharding the test across runners in the future as
our tests grows.
## Description
Fixed issue where accounts that were initially created via
email/password, then linked to an SSO account, can bypass the 2FA during
login if they use their email password.
## Testing Performed
Tested locally, and 2FA is now required for linked SSO accounts
## Description
Add the following document action auth options:
- 2FA
- Passkey
If the user does not have the required auth setup, we onboard them
directly.
## Changes made
Note: Added secondaryId to the VerificationToken schema
## Testing Performed
Tested locally, pending preview tests
## Checklist
- [X] I have tested these changes locally and they work as expected.
- [X] I have added/updated tests that prove the effectiveness of these
changes.
- [X] I have followed the project's coding style guidelines.
<!-- This is an auto-generated comment: release notes by coderabbit.ai
-->
## Summary by CodeRabbit
- **New Features**
- Introduced components for 2FA, account, and passkey authentication
during document signing.
- Added "Require passkey" option to document settings and signer
authentication settings.
- Enhanced form submission and loading states for improved user
experience.
- **Refactor**
- Optimized authentication components to efficiently support multiple
authentication methods.
- **Chores**
- Updated and renamed functions and components for clarity and
consistency across the authentication system.
- Refined sorting options and database schema to support new
authentication features.
- **Bug Fixes**
- Adjusted SignInForm to verify browser support for WebAuthn before
proceeding.
<!-- end of auto-generated comment: release notes by coderabbit.ai -->
## Description
Currently opening modals, clicking select boxes or using anything from
radix that overlays the screen in some way will shift the screen.
This can be easily noticeable when changing the document "Period"
selector on the /documents page.
## Changes Made
Undo the gutter change for now. Can find a proper solution another time.
https://github.com/documenso/documenso/assets/20962767/5bcae576-2944-4ae5-a2c3-0589e7f61bdb
## Description
Currently if you try to load the document edit page when fields need to
be rendered, you will not be able to see the fields until you proceed to
the next step.
This is because the fields require the document PDF to be loaded prior
to rendering them.
This PR resolves that issue by only rendering the fields after the PDF
is loaded.
## Changes Made
- Add a state to track whether the PDF is loaded
- Render the fields only after the PDF is loaded
## Testing Performed
Tested document flow manually and the fields are rendered correctly on
load.
## Checklist
- [X] I have tested these changes locally and they work as expected.
- [X] I have updated the documentation to reflect these changes, if
applicable.
## Description
**Fixes issues with mismatching state between document steps.**
For example, editing a recipient and proceeding to the next step may not
display the updated recipient. And going back will display the old
recipient instead of the updated values.
**This PR also improves mutation and query speeds by adding logic to
bypass query invalidation.**
```ts
export const trpc = createTRPCReact<AppRouter>({
unstable_overrides: {
useMutation: {
async onSuccess(opts) {
await opts.originalFn();
// This forces mutations to wait for all the queries on the page to reload, and in
// this case one of the queries is `searchDocument` for the command overlay, which
// on average takes ~500ms. This means that every single mutation must wait for this.
await opts.queryClient.invalidateQueries();
},
},
},
});
```
I've added workarounds to allow us to bypass things such as batching and
invalidating queries. But I think we should instead remove this and
update all the mutations where a query is required for a more optimised
system.
## Example benchmarks
Using stg-app vs this preview there's an average 50% speed increase
across mutations.
**Set signer step:**
Average old speed: ~1100ms
Average new speed: ~550ms
**Set recipient step:**
Average old speed: ~1200ms
Average new speed: ~600ms
**Set fields step:**
Average old speed: ~1200ms
Average new speed: ~600ms
## Related Issue
This will resolve#470
## Changes Made
- Added ability to skip batch queries
- Added a state to store the required document data.
- Refetch the data between steps if/when required
- Optimise mutations and queries
## Checklist
- [X] I have tested these changes locally and they work as expected.
- [X] I have followed the project's coding style guidelines.
---------
Co-authored-by: Lucas Smith <me@lucasjamessmith.me>
## Description
Add support to login with passkeys.
Passkeys can be added via the user security settings page.
Note: Currently left out adding the type of authentication method for
the 'user security audit logs' because we're using the `signIn`
next-auth event which doesn't appear to provide the context. Will look
into it at another time.
## Changes Made
- Add passkeys to login
- Add passkeys feature flag
- Add page to manage passkeys
- Add audit logs relating to passkeys
- Updated prisma schema to support passkeys & anonymous verification
tokens
## Testing Performed
To be done.
MacOS:
- Safari ✅
- Chrome ✅
- Firefox ✅
Windows:
- Chrome [Untested]
- Firefox [Untested]
Linux:
- Chrome [Untested]
- Firefox [Untested]
iOS:
- Safari ✅
## Checklist
<!--- Please check the boxes that apply to this pull request. -->
<!--- You can add or remove items as needed. -->
- [X] I have tested these changes locally and they work as expected.
<!-- This is an auto-generated comment: release notes by coderabbit.ai
-->
## Summary by CodeRabbit
- **New Features**
- Introduced Passkey authentication, including creation, sign-in, and
management of passkeys.
- Added a Passkeys section in Security Settings for managing user
passkeys.
- Implemented UI updates for Passkey authentication, including a new
dialog for creating passkeys and a data table for managing them.
- Enhanced security settings with server-side feature flags to
conditionally display new security features.
- **Bug Fixes**
- Improved UI consistency in the Settings Security Activity Page.
- Updated button styling in the 2FA Recovery Codes component for better
visibility.
- **Refactor**
- Streamlined authentication options to include WebAuthn credentials
provider.
- **Chores**
- Updated database schema to support passkeys and related functionality.
- Added new audit log types for passkey-related activities.
- Enhanced server-only authentication utilities for passkey registration
and management.
<!-- end of auto-generated comment: release notes by coderabbit.ai -->
## Description
Update the add signer form to disable the signers when required.
I assume the actual issue is that `{...field}` was spreading a disabled
prop which was overriding our one.
## Changes Made
- Use fieldset to disable inputs
- Manually disable select since fieldset doesn't work for that select
for some reason
## Description
Fixes the issue with Vercel preview deployments failing.
It appears that the old `PGHOST` environment variable injected by neon
was:
`ep-snowy-snowflake-a2vc5pa2.eu-central-1.aws.neon.tech`
It is now:
`ep-snowy-snowflake-a2vc5pa2.eu-central-1-pooler.aws.neon.tech`
Notice the `-pooler` being attached automatically to the `PGHOST`.
## References
> The following changes were made to the [Neon Vercel
Integration](https://vercel.com/integrations/neon):
>
>To ensure that users accessing a Neon database from a serverless
environment have enough connections, the DATABASE_URL and PGHOST
environment variables added to a Vercel project by the Neon integration
are now set to a pooled Neon connection string by default. Pooled
connections support up to 10,000 simultaneous connections. Previously,
these variables were set to an unpooled connection string supporting
fewer concurrent connections.
https://neon.tech/docs/changeloghttps://neon.tech/docs/guides/vercel#manage-vercel-environment-variables
## Description
Add support for enterprise billing plans.
Enterprise billing plans by default get access to everything early
adopters do:
- Unlimited teams
- Unlimited documents
They will also get additional features in the future.
## Notes
Pending webhook updates to support enterprise onboarding.
Rolled back env changes `NEXT_PUBLIC_PROJECT` since it doesn't seem to
work.
## Description
Currently we're not restricting team transfer and email verification
tokens from flowing into the frontend.
This changes restricts it to only return the required information
instead of the whole data object.
## Description
Currently the document redirect URL feature is preventing documents from
being created unless a redirect URL is provided.
During the document edit flow, the redirect URL is hidden in an advanced
tab with the value of an empty string, which will always fail the
current Zod validation since `optional` requires undefined to pass.
There are multiple ways to fix this, but I think this is the easiest
method where we can assume an empty string is valid.
## Description
Fixed issue where setting a recipient role as CC will prevent any
further changes as it is considered as "sent" and "signed".
## Other changes
- Prevent editing document after completed
- Removed CC and Viewers from the field recipient list since they will
never be filled
- Minor UI issues
## Checklist
- [X] I have tested these changes locally and they work as expected.
- [X] I have added/updated tests that prove the effectiveness of these
changes.
- [X] I have followed the project's coding style guidelines.
## Description
Add the ability to feature flag the teams feature via UI.
Also added minor UI changes
## Checklist
- [X] I have tested these changes locally and they work as expected.
- [X] I have added/updated tests that prove the effectiveness of these
changes.
- [X] I have followed the project's coding style guidelines.
## Description
Add support for teams which will allow users to collaborate on
documents.
Teams features allows users to:
- Create, manage and transfer teams
- Manage team members
- Manage team emails
- Manage a shared team inbox and documents
These changes do NOT include the following, which are planned for a
future release:
- Team templates
- Team API
- Search menu integration
## Testing Performed
- Added E2E tests for general team management
- Added E2E tests to validate document counts
## Checklist
- [X] I have tested these changes locally and they work as expected.
- [X] I have added/updated tests that prove the effectiveness of these
changes.
- [ ] I have updated the documentation to reflect these changes, if
applicable.
- [X] I have followed the project's coding style guidelines.
## Description
Currently the count for the documents table tabs do not display the
correct values when the period filter is applied.
## Changes Made
- Updated `getStats` to support filtering on period
## Testing Performed
- Tested to see if the documents tab count were being filtered based on
the period
## Checklist
- [X] I have tested these changes locally and they work as expected.
- [X] I have followed the project's coding style guidelines.
## Description
Currently we are required to ensure PII data is not passed around in
search parameters and in the open for GDPR reasons.
Allowing us to encrypt and decrypt values with expiry dates will allow
us to ensure this doesn't happen.
## Changes Made
- Added TPRC router for encryption method
## Testing Performed
- Tested encrypting and decrypting data with and without `expiredAt`
- Tested via directly accessing API and also via trpc in react
components
- Tested parsing en email search param in a page and decrypting it
successfully
## Checklist
- [X] I have tested these changes locally and they work as expected.
- [X] I have followed the project's coding style guidelines.
## Description
Currently there are various z-index values that are causing:
- Toasts to be placed behind dialog blur background
- Menu being cropped off by header
## Changes Made
- Revert `z-[1000]` back to `z-50` for the header (not exactly sure why
it was bumped)
- Refactor z-indexes over 9000 to start from 1000
- Ensure z-index of toast is higher than dialog
## Description
Previously we assumed that there can only be 1 subscription per user.
However, that will soon no longer the case with the introduction of the
Teams subscription.
This PR will apply the required migrations to support multiple
subscriptions.
## Changes Made
- Updated the Prisma schema to allow for multiple `Subscriptions` per
`User`
- Added a Stripe `customerId` field to the `User` model
- Updated relevant billing sections to support multiple subscriptions
## Testing Performed
- Tested running the Prisma migration on a demo database created on the
main branch
Will require a lot of additional testing.
## Checklist
- [ ] I have tested these changes locally and they work as expected.
- [ ] I have added/updated tests that prove the effectiveness of these
changes.
- [X] I have followed the project's coding style guidelines.
## Additional Notes
Added the following custom SQL statement to the migration:
> DELETE FROM "Subscription" WHERE "planId" IS NULL OR "priceId" IS
NULL;
Prior to deployment this will require changes to Stripe products:
- Adding `type` meta attribute
---------
Co-authored-by: Lucas Smith <me@lucasjamessmith.me>
David Nguyen
