Summary:
In React VR we doing automatic retries in cases when request takes too much time. It helps in cases when network quality is poor. As long as this implicit for developer, we would like to highlight this information in insights. It helps to explain why request were so long.
This diff adds one more field to insights which displays amount of time spent on retries, count and limit of them.
https://pxl.cl/CntV
Reviewed By: passy
Differential Revision: D15845598
fbshipit-source-id: 3e0d1baf93b6be5c4d1c6c1d2b64c05852b843b7
Summary:
https://pxl.cl/Cd84
In React VR we got a case to explain developers why their requests took so long. To do this we export timings from Curl and send them in Flipper. This diff added visualization for this data. Any other platforms can use this feature by extending json which they send to Flipper.
Reviewed By: jknoxville
Differential Revision: D15804247
fbshipit-source-id: 9deea5c9f5f7c5416b42c476f7f0ed431c8dea04
Summary: Enable highlight of multiple rows so they can be copied
Reviewed By: passy
Differential Revision: D15822749
fbshipit-source-id: 2302ddd27a57574c1eaf1f1395939c9f006ba301
Summary:
Select multiple rows, and then right click, or use Ctrl-click.
This was broken, it was resetting the selection when this happened.
This fixes it.
Now after selecting multiple, what happens depends where you right-click.
If you right-click within the current selection, it stays selected and you can copy it.
But if you right-click outside the current selection, that row becomes the new selected set, and you will just be able to copy that one. This feels right to me.
Reviewed By: passy
Differential Revision: D15822668
fbshipit-source-id: 6a6cda8a8f16fec55cffa52c1c2ff958c86b3f43
Summary:
When running in headless mode, if SIGINT is received, but there's a failure to export the state the process doesn't exit. Repeated SIGINTs do the same thing.
This changes it to exit whether the export succeeds or not.
Reviewed By: passy
Differential Revision: D15806412
fbshipit-source-id: 1f6a5b4ea1cd65dacb201f8a1cd020531b3976e6
Summary: Keytar binaries weren't updated when upgrading the electron version. This diff updates the binaries to match electron's node version.
Reviewed By: passy
Differential Revision: D15821734
fbshipit-source-id: 04688f1e465d2f7bdefecfdda8e2f3868de18780
Summary: Transfer it from the device, reformat the notification to make use of it.
Reviewed By: danielbuechele
Differential Revision: D15779267
fbshipit-source-id: 747dc7f895528618ff6a07c15b7f72bf6a1adde9
Summary: Need a Fresco release including the previous diff first.
Reviewed By: danielbuechele
Differential Revision: D15762090
fbshipit-source-id: ec30778ffc1cd3e48dcfc4d514a9ba13fb26fd75
Summary:
Trying to detect version mismatches between the running Simulator and the version of Xcode's CLI tools.
What is done in this check:
- The running path of the running simulator is queries from `ps`
- The path of Xcode CLI tools is queried using `xcode-select -p`.
- The two paths are compared if they differ
Reviewed By: jknoxville, passy
Differential Revision: D15803238
fbshipit-source-id: 6638312505fc950ff7869fd55494984bfffb677b
Summary:
This makes it easier to track down leaks as we currently only have their class names. There are some limitations if this is creates on a separate thread, but that can be addressed separately.
This does require temporarily removing the Flipper integration as the interface changes in a backwards-incompatible manner. This will be restored with the next diff in the stack.
Reviewed By: oprisnik
Differential Revision: D15760947
fbshipit-source-id: 3a5c48f6244f1afc8b7e00230bd1ae2f197e6cdf
Summary:
Pull Request resolved: https://github.com/facebook/flipper/pull/463
Was made aware [here](https://github.com/facebook/flipper/issues/461) that
we currently don't build with NDK 20 because of Folly. The above issue exists only for x86 platform and not for other platform. I have updated the folly and RSocket. Although this diff doesn't solve the issue mentioned above for x86, but it updates the dependencies to the latest version. To mitigate temporarily, I have updated the gradle to not build for x86.
Reviewed By: passy
Differential Revision: D15713205
fbshipit-source-id: 79dd5825b0b876c87a1868fde8f3353a76012267
Summary:
This is fairly experimental. The test passes, but I don't know how stable it will be when litho changes for example.
However, it's easy to exclude specific attributes from the hierarchy, so maybe we can iterate if it breaks at first.
It just takes a snapshot of the layout hierarchy from headless flipper, and compares it to it's known one.
To run it:
`cd headless-tests && DEVICE=emulator-5554 yarn test` while you have the sample app running on an emulator.
To update the snapshots, just add `-u` to the end of that command.
Reviewed By: danielbuechele
Differential Revision: D15715674
fbshipit-source-id: 4fe6f83b60f8003d48aceb6468d93c075e6c38b8
Summary:
We in React VR were need newer Chrome DevTools version to make our plugin work properly. Chromium version with required fixes were shipped in electron 5.0.0.
This diff bumps electron version and explicitly allows node integration (accessing require, process and other node related functions from window object in index.html), because it was disabled by default in 5.0.0.
Reviewed By: jknoxville
Differential Revision: D15693539
fbshipit-source-id: 4f71a07b2afaa489aec46940c924f4ad30ad0413
Summary: Adds the capability to customize rows in the elements inspector via a callback.
Reviewed By: danielbuechele
Differential Revision: D15738355
fbshipit-source-id: 27b91a74535736318b7fdb9d2eb44dfa20b4d77b
Summary: No longer necessary, after flow types for `lodash.memoize` were added.
Reviewed By: jknoxville
Differential Revision: D15738356
fbshipit-source-id: 3c911e5eefee347ca69a37051fd5279c5de47873
Summary: The "Conflict in temp dir" message was misleading when there were 0, rather than >1 file. This fixes that error message.
Reviewed By: passy
Differential Revision: D15716247
fbshipit-source-id: 572addf99f623cdcffd7602dd3935bb69a50d047
Summary:
I'm seeing some failures where no client is connected when we do the export.
This increases the wait time before sending the SIGINT.
Reviewed By: passy
Differential Revision: D15716214
fbshipit-source-id: 4646ef9d1c37e491273bd9deff504675b0e4e780
Summary: Add sidebar to the flipper CPU plugin. Sidebar displays detailed information about the selected CPU core
Differential Revision: D15696198
fbshipit-source-id: 47e687b4354653e6bef626cd3f27adcb219dbd28
Summary:
`StatusIndicator` works well as a flex child, but didn’t support usage in inline / text contexts.
Adding `display: inline-block` fixes that.
Reviewed By: danielbuechele
Differential Revision: D15695673
fbshipit-source-id: 30378d2292390b8d86c0a26c11571e5953b841f2
Summary:
Adds a JS test suite for running jest tests against flipper headless binaries. Replaces the jq-based bash script.
Simple sanity test included for all expected top level keys in output JSON.
Reviewed By: passy
Differential Revision: D15625723
fbshipit-source-id: 5ec8e764bcdbb99d9070de704e76a5cd9d8e5dc6
Summary:
`/*` is the standard throughout open source code. For example, Firefox uses single /*: https://hg.mozilla.org/mozilla-central/file/21d22b2f541258d3d1cf96c7ba5ad73e96e616b5/gfx/ipc/CompositorWidgetVsyncObserver.cpp#l3
In addition, Rust considers `/**` to be a doc comment (similar to Javadoc) and having such a comment at the beginning of the file causes `rustc` to barf.
Note that some JavaScript tooling requires `/**`. This is OK since JavaScript files were not covered by the linter in the first place, but it would be good to have that tooling fixed too.
Reviewed By: zertosh
Differential Revision: D15640366
fbshipit-source-id: b4ed4599071516364d6109720750d6a43304c089
Summary: Includes the LeakTracker interface change. Restores the temporarily deleted tracker, wires it back up. Requires Fresco 0.14.2 being released which has yet to happen.
Reviewed By: oprisnik
Differential Revision: D15655667
fbshipit-source-id: ab4361be35101efe7564dfdd7a69cba93135eb7e
Summary:
Ugh, this is the only way we can land the API-breaking next diff. Will
be reverted straight after.
Reviewed By: oprisnik
Differential Revision: D15655666
fbshipit-source-id: c9318e348c187de4562a9a636287cfa7b523356f
Summary:
Pulls out the local `Lamp` component from `PluginDebugge` into `StatusIndicator`.
This will allow to reuse the component in different context, to guarantee a consistent appearance of Flipper throughout.
For good measure, we make diameter and title configurable. Vertical alignment is changed from `margin-top` to `align-self`.
Reviewed By: danielbuechele
Differential Revision: D15692275
fbshipit-source-id: 438b2f5300175565dbf07d07a5d757936b4f7cfe
Summary:
Using `process.exit()` stops the node process without waiting for the event loop to finish, so when using async i/o, which is what happens when piped, if the output is buffered, the process can terminate before it finishes flushing the buffer.
This means you only get some of the output and the JSON is malformed.
This fixes it by calling `process.exit()` inside the flushed callback.
Reviewed By: passy
Differential Revision: D15624806
fbshipit-source-id: ea540ed5a40fb1811e5b705b190da96c8e54730d
Summary: Catch the exception thrown in xcrun, which will occur for linux machines and in the cases when Xcode command line toolchain is not installed.
Reviewed By: passy
Differential Revision: D15621368
fbshipit-source-id: f3044cee6f7a4b5ba927a33003f9690c1fe31586
Summary:
Adds a "Track Leaks" option that will show notifications (even retroactively)
for `CloseableReferences` that were tracked.
Reviewed By: danielbuechele
Differential Revision: D15622596
fbshipit-source-id: ef610379aa96f9a5e541f741af608db30bee74e1
Summary: Should be better than the mock we generated before, I hope.
Reviewed By: danielbuechele
Differential Revision: D15621591
fbshipit-source-id: a153d9233f6715c4c53ccabf2b8d18ce95b37f03
Summary: Added unit tests for the "WASTED_BYTES" for the fresco plugin
Reviewed By: passy
Differential Revision: D15610871
fbshipit-source-id: dc9bd84363e1ba22fc96890b74e217372188a9d0
Summary:
Notification from plugins (especially network) could be very spammy. This diff introduces a throttle, so a plugin can only send a notification every 5 seconds. If a notification is sent earlier, it is not displayed and the counter is reset to 5 seconds from now.
In the notifications panel still all notifications are dispalyed.
Reviewed By: passy
Differential Revision: D15621427
fbshipit-source-id: ea5e3b39fdd5cdedff3fd5234e752800520a423b
Summary:
1. Yargs doesn't like having the same option name as the given alias and will just silently skip those (like --metrics).
2. Having multiple ways of specifying the same argument is not a good practise.
I think we've been misusing `alias` as a way to have more JavaScript-y accessors, but ignoring that yargs already
converts `my-long-argument` to `myLongArgument` without having to expose this.
We haven't rolled out a version with the previous long arguments, so we should
still be able to change this without breaking stuff.
Reviewed By: jknoxville
Differential Revision: D15620636
fbshipit-source-id: 84a8046cf06d696e947719032c4f9c34ac9c0474
Summary: Unrelated to what I was going to work on, so separate diff.
Reviewed By: jknoxville
Differential Revision: D15620643
fbshipit-source-id: a555bbd54adfe44fa7d566cb9cb6d31bd75be28e
Summary: Add a metric for the raw number of leaked closeable references.
Reviewed By: priteshrnandgaonkar
Differential Revision: D15603348
fbshipit-source-id: 03379e180b5822620e8e8347bef354edb3527d19
Summary:
The FlipperArray creation from JSONArray used to create an array of `null`'s instead of the FlipperObject, which used to crash the Flipper. Thus added a check in the flipper to not crash in the case when it receives array of nulls. Also fixed the Fresco plugin in mobile to send the proper FlipperArray.
Bug:
{F160049909}
Reviewed By: passy
Differential Revision: D15608297
fbshipit-source-id: 625e017d3bce4135ac699ee5aee8d871db378749
Summary:
This diff fetches the additional data by calling the hook added in each plugins when the headless tries to export metrics directly out of the store rather than from the trace.
I added `fetchMetadata` function which will be called while exporting the state as well as exporting the metrics. We do not need to fetch metadata in the case when the metrics are exported from the given trace.
Reviewed By: passy
Differential Revision: D15560129
fbshipit-source-id: 9b14340e565ce17d1825bc2d32520d7b0c2b2219
Summary: Actually reduce in the reducer step instead of using mutable state.
Reviewed By: jknoxville
Differential Revision: D15603095
fbshipit-source-id: f603ca548bb760db5423a4c1871d94284ece9521
Summary:
We've got some exit troubles, meaning we have to use process.exit() to stop the headless process. This isn't good because the output isn't necessarily flushed when the process forcefully exits this way.
The tests can check for this, so I've removed the forceExit flag, and fixed the case stopping them from cleanly exiting - the android adb client.
Reviewed By: passy
Differential Revision: D15602919
fbshipit-source-id: b4a9dbabf5d3c874bda6bf938e4acb7b5b7e901b
Summary:
0.100! Just noticed there's a new version and it continues to work.
destruction100
Reviewed By: jknoxville
Differential Revision: D15601865
fbshipit-source-id: 2d9d8a9ffb073468df4b85ed83800bce00ee6cda
