* form handler example: Update deps and fix build from dep update
* Ran lint error fixers
* Fixes errors that occur when commit occurs
* Commit linter fixes
The with-google-analytics example had the "routeChangeComplete" event listener set up in components/Page.js, but that the event listener would only be set up if the user visited a page using that component. From the example, it's not clear if google analytics can be used without making every page use a component like components/Page.js. Someone following the example may make pages that don't use components/Page.js and fail to have page views reported, or feel compelled to force a shared component into their design unnecessarily, or might even make a mistake by making multiple different components like Page.js which each add a new "routeChangeComplete" event listener, causing page views to be over-reported when the user navigates between pages using the different components.
This PR moves the "routeChangeComplete" event listener into _app.js, where it's guaranteed to be executed for every page and is more obviously decoupled from page-layout-related components.
This PR also fixes a React warning about the lack of an onChange handler on an input tag, and removes the unnecessary implementation of `getInitialProps` in _document.js (the default implementation is inherited if not present, there's nothing this example needs to do with `getInitialProps` specifically, and the body of the method seems to have been based on an old version of next's internal implementation).
This PR also fixes the url being passed to google tag manager incorrectly. It looks like page_path should be used instead of page_location because the `url` value only has the path, not the full url with the domain name, etc. (https://developers.google.com/analytics/devguides/collection/gtagjs/pages)
Minor changes to examples. Updating major semver updates with only `package.json` changes.
I've done my best to make sure that these packages.json files all have `latest` for the `nextjs` package, `cross-env` for those with `server.js` files, etc.
I also added a `package.json` to `with-dynamic-app-layout` (it was missing one previously)
Made sure to test all of these packages post-upgrade to ensure maintained functionality
wrapping `startClock` in `bindActionsCreator` there's no need to pass `dispatch` in:
```
this.props.startClock(dispatch)
```
Furthermore `bindActionsCreator` is not needed because following already bind actions:
```
const mapDispatchToProps = {
startClock
};
```
* Add support for .amp.js pages and
resolving /page?amp=1 to page.amp.js
* Update amp tests
* Update example and clean up amp page resolving
* Add nested amp test
* page => normalizedPage
* Add type to page options
* Add handling of amp with all pageExtensions
and normalize page
* Make sure findPageFile only falls back to
amp if enabled
I noticed that the nodemon.json file was not watching for all file changes in the correct location for filetype .ts (typescript). After some research I found that the "ext" option in the nodemon.json fixes the issue and should work across all operating systems.
Replicates the behaviour of the `with-mobx` example but implemented using `mobx-react-lite` and React context.
I'm still working out a best practice regarding actions and welcome feedback on anything.
* Typecheck server vs. client code independently, ensuring that each respects its own tsconfig.
* Use nextjs default distDir in tsconfig
* Update packages
* Fix type error in server.listen callback
### changes
#### remove trailing spaces
When I was using example I noticed trailing spaces.
So, this PR removes the trailing spaces of json file, README, and others.
`examples/with-jest-typescript/src/modules/cars/Overview.tsx` also has it, but this time it did not change as tslint error occurs at commit.
Playing with this example, I realized that it was not doing what I expected in case of an error coming back from the API (e.g : throw properly the error and save it in the state).
Adds an example app using [graphql-hooks](https://github.com/nearform/graphql-hooks) that started life as the with-apollo example app. It uses the same graph.cool backend, mostly to demonstrate how similar it is.
Hello,
I have been using next.js for a while in a bunch of projects, so first for all thanks for all the vibrant effort around the project 🖤.
Always I see the server side next.js approach as an advantage, but also a weakness for the extra resources you need to have, specially comparing how cheap is a client side app.
In order to do my things cheaper, I started using the SSR pattern you suggested in your examples, so useful! It saves time and resources.
However, it was *too simple*. In a real production scenario, you need a bit more, specially related with send the right response headers to keep the rest of external network agent updated of your cache state.
I started a tiny script code for doing that; basically, I copy/paste it on my ssr projects.
Now, after a time, I think it's worth it publish it as [cacheable-response](https://github.com/Kikobeats/cacheable-response) module.
The PR is for adding the module leverage into the next.js ssr example.
It's doing the same, plus:
- be possible use a multi storage cache (memory by default; mongodb, redis, mysql, supported).
- sending `cache-control` response headers.
- sending `X-Cache-Expired-At`, just a humanize way to see the expiration time.
- support for forcing invalidation via `force=true` query parameter.
I hope you like it 🙂
The latest version of babel-jest doesn't require `babel-core` with the bridge version anymore (updated in this PR : https://github.com/facebook/jest/pull/7016).
So I'm updating with-jest and with-jest-react-testing-library examples accordingly.
* This is a non-working example of using PHASE_DEVELOPMENT_SERVER and PHASE_PRODUCTION_SERVER in production. I followed @timneutkens gist but was unable to make it work so I've boiled it down into this non-working example which I believe is the same. The README is not updated. Once it is figured out why this is not working, I'll clean up the project and update the pull request to be complete but for now just want to make it work.
* added .eslintrc so that ` eslint . --fix` would work (not sure if that was really necessary). I assume that is same as `yarn lint-fix`.
* fixes for standard style
* fixes for standard style
* Fix example and add some comments
* Updated documentation and small change to logic of prod,dev,staging to work as expected.
Added significantly more doc then I normally would in the hopes that it helps someone avoid the mis-understanding I went through. If it's too much, LMK and I'll reduce.
* removed eslint and updated package.json to get rid of eslint and standardjs
Next.js no longer uses `jsonPageRes`.
This PR removes it from `en-us` and `zh-cn` documentation (formatting picked up some changes on zh-cn readme).
Also updated the Flow type def for `getInitialProps`.
Resolves#6363
[React ESI](https://github.com/dunglas/react-esi) is a brand new cache library for vanilla React and Next.js applications, that can make highly dynamic applications as fast as static sites by leveraging the open Edge Server Include specification.
https://github.com/dunglas/react-esi
Because this spec is widespread, React ESI natively supports most of the well-known cloud cache providers including Cloudflare Workers, Akamai and Fastly. Of course, React ESI also supports the open source Varnish cache server that you can use in your own infrastructure for free (configuration provided).
This PR shows how to integrate React ESI with Next.js.
After talking with @timneutkens it was decided it'd be more streamlined to replace the onDemandEntries WebSocket with an alternative. Using the EventSource connection gives us these benefits over the WebSocket one:
- less code needed
- no extra server running
- no extra config for onDemandEntries
After discussion, I added falling back to fetch based pinging when the WebSocket fails to connect. I also added an example of how to proxy the onDemandEntries WebSocket when using a custom server. Fixes: #6296
* extract request login from auth
* add clarification that the monorepo is for deploy in Now only and fix typo
* Refactor HOC
- add authorization to HOC
- add displayName to HOC
- remove unnecessary `run`s in local routing
I've just thought of a way to improve the initial props page by adding
an example for a list/detail page structure. To do that, I've created a
separate `/detail` page, and a mock API which calls data from the array
we made on the previous PR.
A ListDetail component is created as an example for displaying detail.
Page structure is also cleaned up. Should I go ahead and add an example
on how to style with styled-jsx + its TS declarations? I might decide to
do it within this week anyway.
Update examples with-relay-modern and with-relay-modern-server-express to react-relay 2.0.0
- react-relay has started to use new Context API instead of Legacy Context API
- add `parseInt` because graphql 14.0.0 introduced stricter scalar value coercion
Closes#6157
This reverts the changes made in [this pr](https://github.com/zeit/next.js/pull/6109).
`redux-saga: "1.0.0"` changed the way it handles it's queues. Because of that we're still having trouble to implement the synchronous side-effects flow in `next-redux-saga`. See [this discussion](https://github.com/bbortt/next-redux-saga/pull/1) for more information.
Therefore I would feel more comfortable not to mislead users by giving them a non-working example in the main branch.
I tried to make the example a bit more descriptive. I changed `publicPath` in `now.config.js` to be `/_next/static/`, in place of `./`, and `outputPath` to `static/` in place of `static/css/`. The reason is that the webpack config will still fallback to `file-loader` for any content that is imported by the user and which is bigger that `8192` bytes. I think this content should not land in the css folder, which should probably stay css specific.
Moreover, for user content, like regular images, the former settings will fail.
If you have this:
```javascript
import LargeFile from './LargeFile.png'
```
it would be placed in `static/css/` but its url would resolve to `<base-url>/LargeFile.png`, which will fail. It works for semantic-ui alone, because `@zeit/next-css` will put the styles in `static/css/` and so `publicPath` of `./` would work just fine.
Putting assets in `static/` and setting `publicPath` to '/_next/static/' will resolve correctly for both semantic-ui related assets as well as for regular user assets.
I hope I am not mixing something up. I tested it locally and in serverless deployment, and this looks pretty consistent.
Hi there
I noticed you have not yet included the api breaking changes in `"redux-saga": "1.0.0"`. Therefore I felt free to upgrade the dependencies in `examples/with-redux-saga`.
I do not know anything about apollo nor graphql, that is why I did not upgrade `examples/with-apollo-and-redux-saga`. But, I think you should do this on occasion.
Keep the great work up.
Regards
I've updated the TypeScript dependency to the latest version. Also
removed some dependencies that may not be needed.
I've also fixed tslint errors which may have appeared because of
previous updates to this starter kit, as well as added comments
to explain some parts of the code.
The current `examples/with-typescript` is not using the latest type definitions currently available on DefinitelyTyped project for next.
Added new list page examples that demonstrate how to use the new Types for both stateless functional components and classes. Also modified examples for list to demonstrate typings for `getInitialProps`.
ctx.pathname was set to url including the query of the page to prefetch
therefore the page was cached with the wrong key (article%3Fid=1?id=1),
and that's why the cache didn't work (right key: article?id=1)
RE: https://github.com/zeit/next.js/issues/4587, this pull request improves the with-segment example.
Previously, only SSR page loads were tracked. This pull request adds manual page view logging via `Router.events.on('routeChangeComplete')` in `Page.js`.
There is also a minor bug fix on the textarea to remove a console error.
…aces (home and about pages). This makes it the example more clear and why someone might want to use _app.js in the first place. Also, added a button on the about page that allows for passing and arbitrary value from the about page to the context provider. I disagree with the naming convention of calling the class CounterProvider. It includes both a Provider and Consumer so it should have some name that covers both. Maybe it should be called CounterContext but I did not change that. I've seen other examples of the same naming conversion so figure I'm the odd duck here (still think it's wrong no matter how many people do it).
