This PR makes it so that `modifyResponseHeaders` is supported in our
adapters when set on the collection config. Previously it would be
ignored.
This means that users can now modify or append new headers to what's
returned by each service.
```ts
import type { CollectionConfig } from 'payload'
export const Media: CollectionConfig = {
slug: 'media',
upload: {
modifyResponseHeaders: ({ headers }) => {
const newHeaders = new Headers(headers) // Copy existing headers
newHeaders.set('X-Frame-Options', 'DENY') // Set new header
return newHeaders
},
},
}
```
Also adds support for `void` return on the `modifyResponseHeaders`
function in the case where the user just wants to use existing headers
and doesn't need more control.
eg:
```ts
import type { CollectionConfig } from 'payload'
export const Media: CollectionConfig = {
slug: 'media',
upload: {
modifyResponseHeaders: ({ headers }) => {
headers.set('X-Frame-Options', 'DENY') // You can directly set headers without returning
},
},
}
```
Manual testing checklist (no CI e2es setup for these envs yet):
- [x] GCS
- [x] S3
- [x] Azure
- [x] UploadThing
- [x] Vercel Blob
---------
Co-authored-by: James <james@trbl.design>
I think it's easier to review this PR commit by commit, so I'll explain
it this way:
## Commits
1. [parallelize eslint script (still showing logs results in
serial)](c9ac49c12d):
Previously, `--concurrency 1` was added to the script to make the logs
more readable. However, turborepo has an option specifically for these
use cases: `--log-order=grouped` runs the tasks in parallel but outputs
them serially. As a result, the lint script is now significantly faster.
2. [run pnpm
lint:fix](9c128c276a)
The auto-fix was run, which resolved some eslint errors that were
slipped in due to the use of `no-verify`. Most of these were
`perfectionist` fixes (property ordering) and the removal of unnecessary
assertions. Starting with this PR, this won't happen again in the
future, as we'll be verifying the linter in every PR across the entire
codebase (see commit 7).
3. [fix eslint non-autofixable
errors](700f412a33)
All manual errors have been resolved except for the configuration errors
addressed in commit 5. Most were React compiler violations, which have
been disabled and commented out "TODO" for now. There's also an unused
`use no memo` and a couple of `require` errors.
4. [move react-compiler linter to eslint-config
package](4f7cb4d63a)
To simplify the eslint configuration. My concern was that there would be
a performance regression when used in non-react related packages, but
none was experienced. This is probably because it only runs on .tsx
files.
5. [remove redundant eslint config files and fix
allowDefaultProject](a94347995a)
The main feature introduced by `typescript-eslint` v8 was
`projectService`, which automatically searches each file for the closest
`tsconfig`, greatly simplifying configuration in monorepos
([source](https://typescript-eslint.io/blog/announcing-typescript-eslint-v8#project-service)).
Once I moved `projectService` to `packages/eslint-config`, all the other
configuration files could be easily removed.
I confirmed that pnpm lint still works on individual packages.
The other important change was that the pending eslint errors from
commits 2 and 3 were resolved. That is, some files were giving the
error: "[File] was not found by the project service. Consider either
including it in the tsconfig.json or including it in
allowDefaultProject." Below I copy the explanatory comment I left in the
code:
```ts
// This is necessary because `tsconfig.base.json` defines `"rootDir": "${configDir}/src"`,
// And the following files aren't in src because they aren't transpiled.
// This is typescript-eslint's way of adding files that aren't included in tsconfig.
// See: https://typescript-eslint.io/troubleshooting/typed-linting/#i-get-errors-telling-me--was-not-found-by-the-project-service-consider-either-including-it-in-the-tsconfigjson-or-including-it-in-allowdefaultproject
// The best practice is to have a tsconfig.json that covers ALL files and is used for
// typechecking (with noEmit), and a `tsconfig.build.json` that is used for the build
// (or alternatively, swc, tsup or tsdown). That's what we should ideally do, in which case
// this hardcoded list wouldn't be necessary. Note that these files don't currently go
// through ts, only through eslint.
```
6. [Differentiate errors from warnings in VScode ESLint
Rules](5914d2f48d)
There's no reason to do that. If an eslint rule isn't an error, it
should be disabled or converted to a warning.
7. [Disable skip lint, and lint over the entire repo now that it's
faster](e4b28f1360)
The GitHub action linted only the files that had changed in the PR.
While this seems like a good idea, once exceptions were introduced with
[skip lint], they opened the door to propagating more and more errors.
Often, the linter was skipped, not because someone introduced new
errors, but because they were trying to avoid those that had already
crept in, sometimes accidentally introducing new ones.
On the other hand, `pnpm lint` now runs in parallel (commit 1), so it's
not that slow. Additionally, it runs in parallel with other GitHub
actions like e2e tests, which take much longer, so it can't represent a
bottleneck in CI.
8. [fix lint in next
package](4506595f91)
Small fix missing from commit 5
9. [Merge remote-tracking branch 'origin/main' into
fix-eslint](563d4909c1)
10. [add again eslint.config.js in payload
package](78f6ffcae7)
The comment in the code explains it. Basically, after the merge from
main, the payload package runs out of memory when linting, probably
because it grew in recent PRs. That package will sooner or later
collapse for our tooling, so we may have to split it. It's already too
big.
## Future Actions
- Resolve React compiler violations, as mentioned in commit 3.
- Decouple the `tsconfig` used for typechecking and build across the
entire monorepo (as explained in point 5) to ensure ts coverage even for
files that aren't transpiled (such as scripts).
- Remove the few remaining `eslint.config.js`. I had to leave the
`richtext-lexical` and `next` ones for now. They could be moved to the
root config and scoped to their packages, as we do for example with
`templates/vercel-postgres/**`. However, I couldn't get it to work, I
don't know why.
- Make eslint in the test folder usable. Not only are we not linting
`test` in CI, but now the `pnpm eslint .` command is so large that my
computer freezes. If each suite were its own package, this would be
solved, and dynamic codegen + git hooks to modify tsconfig.base.json
wouldn't be necessary
([related](https://github.com/payloadcms/payload/pull/11984)).
Previously with `@payloadcms/plugin-storage-blob`, if token was not set,
the plugin would throw an error. This caused a less-than-ideal developer
experience.
With this change, if the `token` value is undefined:
- Local storage will be used as a fallback
- The error will no longer be thrown.
This PR makes three major changes to the codebase:
1. [Component Paths](#component-paths)
Instead of importing custom components into your config directly, they
are now defined as file paths and rendered only when needed. That way
the Payload config will be significantly more lightweight, and ensures
that the Payload config is 100% server-only and Node-safe. Related
discussion: https://github.com/payloadcms/payload/discussions/6938
2. [Client Config](#client-config)
Deprecates the component map by merging its logic into the client
config. The main goal of this change is for performance and
simplification. There was no need to deeply iterate over the Payload
config twice, once for the component map, and another for the client
config. Instead, we can do everything in the client config one time.
This has also dramatically simplified the client side prop drilling
through the UI library. Now, all components can share the same client
config which matches the exact shape of their Payload config (with the
exception of non-serializable props and mapped custom components).
3. [Custom client component are no longer
server-rendered](#custom-client-components-are-no-longer-server-rendered)
Previously, custom components would be server-rendered, no matter if
they are server or client components. Now, only server components are
rendered on the server. Client components are automatically detected,
and simply get passed through as `MappedComponent` to be rendered fully
client-side.
## Component Paths
Instead of importing custom components into your config directly, they
are now defined as file paths and rendered only when needed. That way
the Payload config will be significantly more lightweight, and ensures
that the Payload config is 100% server-only and Node-safe. Related
discussion: https://github.com/payloadcms/payload/discussions/6938
In order to reference any custom components in the Payload config, you
now have to specify a string path to the component instead of importing
it.
Old:
```ts
import { MyComponent2} from './MyComponent2.js'
admin: {
components: {
Label: MyComponent2
},
},
```
New:
```ts
admin: {
components: {
Label: '/collections/Posts/MyComponent2.js#MyComponent2', // <= has to be a relative path based on a baseDir configured in the Payload config - NOT relative based on the importing file
},
},
```
### Local API within Next.js routes
Previously, if you used the Payload Local API within Next.js pages, all
the client-side modules are being added to the bundle for that specific
page, even if you only need server-side functionality.
This `/test` route, which uses the Payload local API, was previously 460
kb. It is now down to 91 kb and does not bundle the Payload client-side
admin panel anymore.
All tests done
[here](https://github.com/payloadcms/payload-3.0-demo/tree/feat/path-test)
with beta.67/PR, db-mongodb and default richtext-lexical:
**dev /admin before:**
**dev /admin after:**
---
**dev /test before:**
**dev /test after:**
---
**build before:**
**build after::**
### Usage of the Payload Local API / config outside of Next.js
This will make it a lot easier to use the Payload config / local API in
other, server-side contexts. Previously, you might encounter errors due
to client files (like .scss files) not being allowed to be imported.
## Client Config
Deprecates the component map by merging its logic into the client
config. The main goal of this change is for performance and
simplification. There was no need to deeply iterate over the Payload
config twice, once for the component map, and another for the client
config. Instead, we can do everything in the client config one time.
This has also dramatically simplified the client side prop drilling
through the UI library. Now, all components can share the same client
config which matches the exact shape of their Payload config (with the
exception of non-serializable props and mapped custom components).
This is breaking change. The `useComponentMap` hook no longer exists,
and most component props have changed (for the better):
```ts
const { componentMap } = useComponentMap() // old
const { config } = useConfig() // new
```
The `useConfig` hook has also changed in shape, `config` is now a
property _within_ the context obj:
```ts
const config = useConfig() // old
const { config } = useConfig() // new
```
## Custom Client Components are no longer server rendered
Previously, custom components would be server-rendered, no matter if
they are server or client components. Now, only server components are
rendered on the server. Client components are automatically detected,
and simply get passed through as `MappedComponent` to be rendered fully
client-side.
The benefit of this change:
Custom client components can now receive props. Previously, the only way
for them to receive dynamic props from a parent client component was to
use hooks, e.g. `useFieldProps()`. Now, we do have the option of passing
in props to the custom components directly, if they are client
components. This will be simpler than having to look for the correct
hook.
This makes rendering them on the client a little bit more complex, as
you now have to check if that component is a server component (=>
already has been rendered) or a client component (=> not rendered yet,
has to be rendered here). However, this added complexity has been
alleviated through the easy-to-use `<RenderMappedComponent />` helper.
This helper now also handles rendering arrays of custom components (e.g.
beforeList, beforeLogin ...), which actually makes rendering custom
components easier in some cases.
## Misc improvements
This PR includes misc, breaking changes. For example, we previously
allowed unions between components and config object for the same
property. E.g. for the custom view property, you were allowed to pass in
a custom component or an object with other properties, alongside a
custom component.
Those union types are now gone. You can now either pass an object, or a
component. The previous `{ View: MyViewComponent}` is now `{ View: {
Component: MyViewComponent} }` or `{ View: { Default: { Component:
MyViewComponent} } }`.
This dramatically simplifies the way we read & process those properties,
especially in buildComponentMap. We can now simply check for the
existence of one specific property, which always has to be a component,
instead of running cursed runtime checks on a shared union property
which could contain a component, but could also contain functions or
objects.
- [x] I have read and understand the
[CONTRIBUTING.md](https://github.com/payloadcms/payload/blob/main/CONTRIBUTING.md)
document in this repository.
---------
Co-authored-by: PatrikKozak <patrik@payloadcms.com>
Co-authored-by: Paul <paul@payloadcms.com>
Co-authored-by: Paul Popus <paul@nouance.io>
Co-authored-by: Jacob Fletcher <jacobsfletch@gmail.com>
Co-authored-by: James <james@trbl.design>
- Upgrades eslint from v8 to v9
- Upgrades all other eslint packages. We will have to do a new
full-project lint, as new rules have been added
- Upgrades husky from v8 to v9
- Upgrades lint-staged from v14 to v15
- Moves the old .eslintrc.cjs file format to the new eslint.config.js
flat file format.
Previously, we were very specific regarding which rules are applied to
which files. Now that `extends` is no longer a thing, I have to use
deepMerge & imports instead.
This is rather uncommon and is not a documented pattern - e.g.
typescript-eslint docs want us to add the default typescript-eslint
rules to the top-level & then disable it in files using the
disable-typechecked config.
However, I hate this opt-out approach. The way I did it here adds a lot
of clarity as to which rules are applied to which files, and is pretty
easy to read. Much less black magic
## .eslintignore
These files are no longer supported (see
https://eslint.org/docs/latest/use/configure/migration-guide#ignoring-files).
I moved the entries to the ignores property in the eslint config. => one
less file in each package folder!
# Breaking Changes
### New file import locations
Exports from the `payload` package have been _significantly_ cleaned up.
Now, just about everything is able to be imported from `payload`
directly, rather than an assortment of subpath exports. This means that
things like `import { buildConfig } from 'payload/config'` are now just
imported via `import { buildConfig } from 'payload'`. The mental model
is significantly simpler for developers, but you might need to update
some of your imports.
Payload now exposes only three exports:
1. `payload` - all types and server-only Payload code
2. `payload/shared` - utilities that can be used in either the browser
or in Node environments
3. `payload/node` - heavy utilities that should only be imported in Node
scripts and never be imported into bundled code like Next.js
### UI library pre-bundling
With this release, we've dramatically sped up the compile time for
Payload by pre-bundling our entire UI package for use inside of the
Payload admin itself. There are new exports that should be used within
Payload custom components:
1. `@payloadcms/ui/client` - all client components
2. `@payloadcms/ui/server` - all server components
For all of your custom Payload admin UI components, you should be
importing from one of these two pre-compiled barrel files rather than
importing from the more deeply nested exports directly. That will keep
compile times nice and speedy, and will also make sure that the bundled
JS for your admin UI is kept small.
For example, whereas before, if you imported the Payload `Button`, you
would have imported it like this:
```ts
import { Button } from '@payloadcms/ui/elements/Button'
```
Now, you would import it like this:
```ts
import { Button } from '@payloadcms/ui/client'
```
This is a significant DX / performance optimization that we're pretty
pumped about.
However, if you are importing or re-using Payload UI components
_outside_ of the Payload admin UI, for example in your own frontend
apps, you can import from the individual component exports which will
make sure that the bundled JS is kept to a minimum in your frontend
apps. So in your own frontend, you can continue to import directly to
the components that you want to consume rather than importing from the
pre-compiled barrel files.
Individual component exports will now come with their corresponding CSS
and everything will work perfectly as-expected.
### Specific exports have changed
- `'@payloadcms/ui/templates/Default'` and
`'@payloadcms/ui/templates/Minimal`' are now exported from
`'@payloadcms/next/templates'`
- Old: `import { LogOut } from '@payloadcms/ui/icons/LogOut'` new:
`import { LogOutIcon } from '@payloadcms/ui/icons/LogOut'`
## Background info
In effort to make local dev as fast as possible, we need to import as
few files as possible so that the compiler has less to process. One way
we've achieved this in the Admin Panel was to _remove_ all .scss imports
from all components in the `@payloadcms/ui` module using a build
process. This stripped all `import './index.scss'` statements out of
each component before injecting them into `dist`. Instead, it bundles
all of the CSS into a single `main.css` file, and we import _that_ at
the root of the app.
While this concept is _still_ the right solution to the problem, this
particular approach is not viable when using these components outside
the Admin Panel, where not only does this root stylesheet not exist, but
where it would also bloat your app with unused styles. Instead, we need
to _keep_ these .scss imports in place so they are imported directly
alongside your components, as expected. Then, we need create a _new_
build step that _separately_ compiles the components _without_ their
stylesheets—this way your app can consume either as needed from the new
`client` and `server` barrel files within `@payloadcms/ui`, i.e. from
within `@payloadcms/next` and all other admin-specific packages and
plugins.
This way, all other applications will simply import using the direct
file paths, just as they did before. Except now they come with
stylesheets.
And we've gotten a pretty awesome initial compilation performance boost.
---------
Co-authored-by: James <james@trbl.design>
Co-authored-by: Alessio Gravili <alessio@gravili.de>
Types are now auto-generated by default.
You can opt-out of this behavior by setting:
```ts
buildConfig({
// Rest of config
typescript: {
autoGenerate: false
},
})
```
