Summary:
Pull Request resolved: https://github.com/facebook/flipper/pull/317
This could save our collective sanity. There's now
built-in support for HTTP retries: 0904f9d136
Reviewed By: jknoxville
Differential Revision: D12906015
fbshipit-source-id: 6030bfb6b914cfda186d92d20f3b15e693b1710c
Summary: Arcanist is deprecated and slow. Let's not use it.
Reviewed By: jknoxville
Differential Revision: D12905244
fbshipit-source-id: c966df5e58ac8942bdb6b66f71e1eb7652477348
Summary: This diff should make HMR work again with the latest version of React
Reviewed By: yungsters
Differential Revision: D12900009
fbshipit-source-id: f9d4dad3ccb4e236f0545983534c167c8fcfbf31
Summary:
Just a quick startup test that can save you the time of manually
opening the app.
I don't have any plans of hooking this up in Circle, but it's
still kinda useful to have.
Reviewed By: jknoxville
Differential Revision: D12901590
fbshipit-source-id: a16a1814574d9f27770dcf98c6b238517c589930
Summary: Including a fix for OkHttp issues, contributed by NightlyNexus: https://github.com/facebook/flipper/pull/313
Reviewed By: jknoxville
Differential Revision: D12883196
fbshipit-source-id: e17b43b33694f46a2bd949eafc008a772777f83e
Summary: Added NTSubDescriptor file which implements the SKSubDescriptor interface to pass all NT specific data to the Layout Component Descriptor.
Reviewed By: jknoxville
Differential Revision: D12840392
fbshipit-source-id: bff295e13e8531456428424b7e073aefe7e1ced4
Summary:
We're seeing some crashes when attempting to send payloads larger than are supported by rsocket.
Instead of crashing, skip the message, but log it so we can see what it's containing.
This is where the failure occurs: https://github.com/rsocket/rsocket-cpp/blob/master/rsocket/framing/FramedDuplexConnection.cpp#L64
Reviewed By: passy
Differential Revision: D12857616
fbshipit-source-id: 2b02d7f5dd6499ba81783d3f8aefcbb64d9d408a
Summary:
The bash file was running adb reverse on the oneworld devices.
Switching to use the real code to do this instead.
--forceExit is needed because the dispatcher keeps an open handle on the adb connection, and it doesn't expose a close() method to stop it. I'm not sure of the best way to fix that, but it's out of scope for this diff anyway.
Reviewed By: passy
Differential Revision: D10868563
fbshipit-source-id: fb44ff1353d5050bb7967cb6c4fa18c2e5d0d149
Summary:
It's only used in `JavascriptSessionTest`, and seems fine to keep it in tests
Resolves#172
Pull Request resolved: https://github.com/facebook/flipper/pull/309
Reviewed By: passy
Differential Revision: D12840362
Pulled By: jknoxville
fbshipit-source-id: 4ead5a4bf0c08d792abcadef713f907930a3e4e7
Summary: Added separate subsections for ntData in objc - json string will be parsed on ios into correct subsections
Reviewed By: priteshrnandgaonkar
Differential Revision: D12817695
fbshipit-source-id: e1480303e121ceb537c47e6c30d328654ab91b8a
Summary: Currently we don't have a flag in iOS setup which lets us know if the code is going to be compiled in Open source or internally. Till now we didn't have a use case for it. But NT team is hacking a componentkit descriptor which is open sourced. This flag will come in handy for us to keep the open source setup build successfully.
Reviewed By: jknoxville
Differential Revision: D10854300
fbshipit-source-id: ac5da6f239b2066b6d7f0680cbfc1b7062ae0f36
Summary: Completes D9667222 to allow force rendering of accessibility in litho when an accessibility service is not actually running.
Reviewed By: ikenwoo
Differential Revision: D9720987
fbshipit-source-id: 000a7413fcbe775a525ca3aca64443c3669cdf9e
Summary:
Making the runner apply the necessary oneworld fix patch before running the tests.
This can't be a long term solution, but at least means not getting that breaking change landed doesn't stop our tests from running.
Since that change is only necessary for the tests, maybe we'll get away with fixing the update process before having to land it.
Reviewed By: priteshrnandgaonkar
Differential Revision: D10852110
fbshipit-source-id: b53a6d0ae9cb51e62cafc0c0f2cfe359ee6aff46
Summary: Was set to view previously and blocked the ability to edit view attributes.
Reviewed By: sjkirby
Differential Revision: D10865495
fbshipit-source-id: 8db043bdc8725b2b56f037d7d8467233db524f5e
Summary:
`flipper` does a lot of stuff at import time. I copied the `styled` reexport
to the UI module, so it's possible to just depend on that.
`remote` is another one of those pot holes that can easily ruin your journey
if you aren't careful navigating around it. Added some defensive fallbacks here
which ultimately allows the lint test to run (and pass) without Electron.
Reviewed By: danielbuechele
Differential Revision: D10507655
fbshipit-source-id: 6dad5665ba15b5cb73caa4ea46262cf337173eb8
Summary:
`INITIAL_STATE` is defined at import-time and relies on `remote.getCurrentWindow()`, which means that any import from the `flipper` module will require a display server + electron running.
Happy to take alternative suggestions to make this look a bit less intrusive.
Reviewed By: danielbuechele
Differential Revision: D10507539
fbshipit-source-id: 62b6bff0b4fd40638e54328613de4d3821b69dd6
Summary:
This won't affect runtime, but is useful when debugging issues with flipper itself.
Setting this flag means all state changes will be output in logcat.
Reviewed By: passy
Differential Revision: D10231148
fbshipit-source-id: 7ddd490290ad973fc339b2f5f93a6f9a1fab4577
Summary:
The code here is quite complicated so here's whats happening:
There's a disconnect hook, that waits and reconnects, so flipper never stops, except for the case where the desktop itself fails.
This makes some sense, but not total sense, as the client should really keep connecting, in case there's a different desktop that won't fail, or even if it was a transient failure.
Explicitly disconnecting, by setting the client to null, will trigger the disconnect hook and carry on the connect loop.
Reviewed By: passy
Differential Revision: D10231241
fbshipit-source-id: 90d823dce2221b1fb6501bd2797871dc1fb9770a
Summary:
Flipper tracks devices on an adb connection. If the adb server gets killed, it deletes all devices, but doesn't attempt to reconnect.
This gets it to rety after 500ms.
Reviewed By: priteshrnandgaonkar
Differential Revision: D10526974
fbshipit-source-id: a2101067f2245b728f458fc5e06dc68833c5e772
Summary: This class already has an adb client. Instead of creating new ones, should just use that.
Reviewed By: passy
Differential Revision: D10231152
fbshipit-source-id: c9426a12f0f6baf94026da51a79131d07e58053c
Summary:
Added a basic communication flows and a button to trigger notification
{F142016937}
Reviewed By: jknoxville
Differential Revision: D10492428
fbshipit-source-id: b65fc46b3be695852f9197771a253d9e8596f328
Summary: Added example plugin for iOS, similar to android
Reviewed By: jknoxville
Differential Revision: D10492429
fbshipit-source-id: d639edc7a47ab240bb172516c2f38b8e2a9f285c
Summary:
This responder was added to android but not iOS, and was always being called during layout init.
Fixing by implementing it in iOS too.
Reviewed By: priteshrnandgaonkar
Differential Revision: D10489411
fbshipit-source-id: ef711dc5f54f52dd76db1245d07f9007fc846532
Summary: This is actually the new minimum required version for the gradle plugin we use.
Reviewed By: jknoxville
Differential Revision: D10505346
fbshipit-source-id: ce331851b74997369325f27f050efae07e4bf00c
Summary:
This masks this fact that `mConnection` must be `non-null` for
very little convenience in return. Let's keep this explicit.
Reviewed By: danielbuechele
Differential Revision: D10505205
fbshipit-source-id: d8187cc8e79d4508babe281ff88b3ba75c0baa1f
Summary:
Let's the notification render any React node.
Particularly useful when including links or wanting to
format messages.
Reviewed By: jknoxville
Differential Revision: D10488157
fbshipit-source-id: 06d3c6b4477c78e8bef78e3a3dc0e05bc67a2ef4
Summary:
Gets a oneworld device (just android for now), installs an app. Checks it connects and stays connected for 5s.
The app installed is in an array, we just need to add more targets for further tests.
Reviewed By: danielbuechele
Differential Revision: D10231083
fbshipit-source-id: c33fb091c1b7914a33690398a79fd700c445c06c
Summary:
I feel like doing async stuff here isn't a good idea in general.
But more pressingly, it means you can't immediately call server.close() after new Server(), because the things to close haven't been created yet.
Reviewed By: danielbuechele
Differential Revision: D10488301
fbshipit-source-id: 76ebe91e0c09f353e0bdb9f2e4116757e757abb2
Summary:
We were using `fill={true}` as an attribute to make flexbox containers fill the entire available space.
However, `fill` is an HTML attribute (see: https://developer.mozilla.org/en-US/docs/Web/SVG/Attribute/fill) This caused warnings printed to the console.
This diff renames the attribute to `grow` with is also more in line with the Flexbox terminology.
Reviewed By: priteshrnandgaonkar
Differential Revision: D10488389
fbshipit-source-id: ed8553c6203cdf6df94d26c731164ecec4c9fbd2
Summary:
I tried switching Flipper to use the node.js module: https://nodejs.org/api/perf_hooks.html instead of electrons window.performance. But the module couldn't be found.
So instead, I'm just replacing it in the pure node tests.
This means tests without node can work without issues.
Reviewed By: danielbuechele
Differential Revision: D10302114
fbshipit-source-id: b3b8aa5fb350a8246fcdfc2a80cc72daec2448d7
Summary:
The idea behind this is to enable tests running in pure node, without electron, because electron requires display drivers and complicated things.
I've been lazy and only added the bare minimum electron api's for the connection test to pass.
It would be awesome to auto generate it properly.
Reviewed By: danielbuechele
Differential Revision: D10238132
fbshipit-source-id: d4a2da2baf5f8964184b532a581c3a75ee33ef06
Summary: Clicking on a notifications links to the Notification Hub highlighting the selected notification.
Reviewed By: jknoxville
Differential Revision: D10487822
fbshipit-source-id: ed907ec244bef970d1b30ddb719856949229d0c4
Summary: React has a built in API for computing the state from props called `getDerivedStateFromProps`. So let's use this instead of our custom implementation.
Reviewed By: jknoxville
Differential Revision: D10484213
fbshipit-source-id: 4c434c5252dabfc2f6015cb6a50719b985c60446
Summary:
Allow linking from one plugin to another. Adds a prop `selectPlugin` to the plugin that can be passed a pluginID and an optional `deepLinkPayload`.
The return value tells you if switching the plugin was successful.
```
selectPlugin: (pluginID: string, deepLinkPayload: ?string) => boolean,
```
Reviewed By: passy
Differential Revision: D10483925
fbshipit-source-id: 6f821277150b2db185b7d545c310214a11432eac
Summary:
This will make sure your local pods repo is up to date, so you won't be building against old dependencies.
Pull Request resolved: https://github.com/facebook/flipper/pull/296
Reviewed By: passy
Differential Revision: D10446899
Pulled By: jknoxville
fbshipit-source-id: 61c8832e8d5f7c54a43fc3e419347687756f4883
