Summary: as title; we'd like to reuse this for the Litho sections plugin but not having it in OS makes setting up deps more difficult than it should be.
Reviewed By: danielbuechele
Differential Revision: D16052298
fbshipit-source-id: cd965688eff4fedbe57264e6676b6ca09b9deb45
Summary:
Event emitter was typed any. Therefore, Client extended any and thing went south.
removing the explicit any type and fixing flow errors.
Reviewed By: jknoxville
Differential Revision: D15987630
fbshipit-source-id: 9456149d46f839d9579fcc2671ab2a82935fc23a
Summary:
Bug Summary:
When running Flipper on Mac, the Mac device is always first to register with Flipper and thus is always selected as the default connected device when first launching. This blocks the beautiful tutorial screen from appearing for new users.
Fix:
When registering new devices in the redux store, a new check has been added that maintains a blacklist of devices that cannot be selected as default. I have added Mac to start. At the same time, this fix preserves userPreferredDevices so that if the user has selected Mac as a device in the past, then next time Flipper is opened, even blacklisted devices will be displayed.
Making this fix uncovered a race condition, between the redux state being rehydrated and the devices being registered. I have potentially fixed this via a callback function in persistStore.
Reviewed By: passy
Differential Revision: D16048371
fbshipit-source-id: 79580b30e8a3b077dac1ac15131266e59646253f
Summary: Now if you copy one of components in flipper in layout plugin, you will get not only name but also id in your clipboard
Reviewed By: priteshrnandgaonkar
Differential Revision: D16031628
fbshipit-source-id: e2c38182dea28719615609f3aa2c74e58749cce4
Summary:
Summary of Bug:
The devices menu popup does not toggle as expected.
Because of the way the click event is fired in Javascript, that is mousedown, mouseup, click, and the fact that the electron menu closes on mousedown, was why the menu was not toggling as expected.
Fix:
This is a potential fix where I add some state to tell if the menu was recently closed, and disable the popup if it was. This is done through the callback method of menu.popup. Then on the next mouse event I renable the popup.
In addition to this, I moved the popup logic into onMouseUp instead of onClick. This was simply to make the code neater. Before I would have to add a timer to the onMouseUp event that would fire after the click event and renable the popup.
Let me know if there is a simple solution to this problem that I've completely missed.
Reviewed By: jknoxville
Differential Revision: D16028864
fbshipit-source-id: a73202536762cce2ab623176cbda0369ca2b8079
Summary:
Fixes the bug when the crash options had same title. Selection was broken in that case. Look at the video.
{F163450316}
Reviewed By: jknoxville
Differential Revision: D15985919
fbshipit-source-id: 7366c8f5f33bbddc15c058b7d20d78d295161404
Summary: Before this diff "Clear Logs" didn't used to clear the log entries from the redux store. It just used to update the state, so when you switched back from other plugins, all previous log information used to pop up again. This diff fixes that.
Reviewed By: passy
Differential Revision: D16012710
fbshipit-source-id: 4393520ab6f77caa5f286e22446d2cb6d0827d71
Summary: Do not fetch meta data when the selected device is the archived one,
Reviewed By: jknoxville
Differential Revision: D16011742
fbshipit-source-id: cd84ef62094ffd8f90922d304ebd911f3b36f469
Summary: This will make it easier to handle errors, because you know you can get the message out, along with stack trace, without having to deal with potentially any object that could be passed in.
Reviewed By: passy
Differential Revision: D16006180
fbshipit-source-id: f200965d7663c5f8a471603fceb6b4cdbea73381
Summary:
Line 166 is missing a process.exit() causing the process to stay open forever if anything goes wrong.
Making everything use errorAndExit, and making use of the utility function so we don't need to remember when to print e, and when to print e.message.
Reviewed By: passy
Differential Revision: D15989164
fbshipit-source-id: c683fa07d98fe49283a59b5f8108ef54babc97a3
Summary:
Upgrading `tmp`, `dashify` and `fs-extras` to their latest versions in an attempt to break my PSC writer's block.
Changelogs don't indicate any breaking changes that should affect us.
Reviewed By: danielbuechele
Differential Revision: D15983944
fbshipit-source-id: 53e87cf3c05e5742db1073e6ccf7f3bd4412638d
Summary: When an Android device disconnects, the device is marked as offline. In this diff, we want to make sure the client for the disconnected apps is not removed, but converted into a "disconnected" client. This makes sure plugins are still shown for the disconnected app.
Reviewed By: passy
Differential Revision: D15985424
fbshipit-source-id: 650ef1344b8be4411794b0344805cb75ceae4a83
Summary:
Logging to see how many people use this feature in addition to the notification logs
we keep.
Reviewed By: oprisnik
Differential Revision: D15983761
fbshipit-source-id: db64bf97d0643f2bfce25a539ee10147924ee3f8
Summary: Refactors the headless to have a list of closures instead of the `if` conditions. This will make addition of new arguments easy as it will be just adding a new closure to the list. One can change the order of the execution of the closures by inserting it at whatever index the user wants.
Reviewed By: passy
Differential Revision: D15965142
fbshipit-source-id: e9f348fe9d9011adcd47d140713d6187eab3b3d3
Summary:
Overall improvements
* Fixed a memory leak caused by promise chaining
* Fixing multi-row selection using shift in tables
* Option to include column names when copying from a table
* Copying table rows which are currently outside the viewport
* Support regex search
* Making client connection more reliable
Plugins
Layout
* fixed a bug in layout search which caused a crash
* Adding accessibilityIdentifier for iOS
Logs
* Debouncing user input in search
* Fixing “stick to bottom” in logs
* Showing timestamp with logs
* Logs for a disconnected device are kept
Network
* Catching malformed URLs in network plugin
* copy full network request and response
* search network request using regex
* Correctly calculate response size
Fresco
* UI fixes in Fresco plugin
* Add missing events
* Add open-source documentation
Crash Reporter
* Account for long crash titles
* Fixing crash notifications
* Fixing display of long stacktraces
Export
* Fix error display
Other Plugins
* Add analytics logging to export
* Fixing UI glitch in Sections plugin
Infrastructure
* Add experimental support for mixed mode with TypeScript
* Upgraded to Flow 0.101.1
* Fix tracking of request/response types between client and device
* Improved iOS simulator compatibility
Reviewed By: jknoxville
Differential Revision: D15963819
fbshipit-source-id: 8af009381c05b9891d46f05a4b8f018754ae6517
Summary: We catch everything that's thrown during the export, which sadly doesn't guarantee that everything is actually an `Error`. So we'll try to JSONify it instead if that happens. It's gross, but well, JavaScript.
Reviewed By: danielbuechele
Differential Revision: D15943111
fbshipit-source-id: 03e6c18403961c00c5878d093c813573461e637b
Summary:
I'm reverting back to not using range qualifiers, too, because
my VSCode got out of sync with the flowconfig and got kinda upset.
Reviewed By: jknoxville
Differential Revision: D15940128
fbshipit-source-id: ce621e72245b66810a29559788515559740c028d
Summary:
Adding a `archive` method to Android devices, that returns a new ArchivedDevice with the same properties as the Android device. This method is called when an android device disconnects and the new ArchivedDevice is added to the devices list. When the device reconnects again, the archived device is removed.
Currently only logs are persisted. In following diffs we can:
- add support for iOS
- move the persisted pluginStates to the archived device as well
Reviewed By: passy
Differential Revision: D15942904
fbshipit-source-id: 07c5415994594abd630d0c4b458b76d1aac6ef02
Summary:
Apple has changed the xcrun api in some cases, it seems.
Adding support for true instead of YES...
Reviewed By: danielbuechele
Differential Revision: D15940810
fbshipit-source-id: 94ba2733527e005b989fb5a62ffbab2f7a0243b9
Summary:
We didn't start the perf marker with an actual marker, so I don't think we logged the event anywhere.
Searching for `onMessageCallback` doesn't bring anything up in the dev tools.
Reviewed By: danielbuechele
Differential Revision: D15939232
fbshipit-source-id: 95ede47f8a762e33d5532230ba83c2939539ba77
Summary:
This sometimes take *A LONG* time for me, so I want to
know if I'm an outlier or not.
Reviewed By: jknoxville
Differential Revision: D15939147
fbshipit-source-id: 9ebd3914efdd537eeadd49e522397acc97a6ff9c
Summary: Multiple markers might got truncated, now the list expands on hover
Reviewed By: jknoxville
Differential Revision: D15939404
fbshipit-source-id: 2640f200dc17876391e04eeced59bc7a2d925113
Summary:
When the stack trace is too long its difficult to read the stack trace as the scroll bar shows up. This diff makes the stack trace text to wrap onto the next line, inorder to avoid scrolling. Look at the attached image, to understand the bug.
{F162756039}
Reviewed By: passy
Differential Revision: D15921226
fbshipit-source-id: f7320951ccdcf531f4276088e3f3ee0f552ab2fc
Summary:
I received an error like this while rendering:
So we should type the callstack accordingly.
Reviewed By: priteshrnandgaonkar
Differential Revision: D15921517
fbshipit-source-id: 5863c75af142f1c22d5b520db8d7c0283f401e11
Summary: Getting the "definition" of the database was originally on the GetTableStructure endpoint. This diff moves it to a new GetTableInfo endpoint
Reviewed By: jknoxville
Differential Revision: D15902619
fbshipit-source-id: ac136d24ee577711366636801b5d74d83fbc523f
Summary:
For every client, there should be a device, otherwise flipper isn't working properly. However, these two things are created independently and either can be created first.
So make the device a promise, that will get fulfilled with the device if any appears in a reasonable time frame.
Previously, if there wasn't a device at the time needed, we just used '' for the name, I've preserved this functionality, though with a potential small delay.
We should be able to get rid of the NEW_CLIENT_SANITY_CHECK in the connections reducer, because the success rate of this promise achieves the same result.
Reviewed By: passy
Differential Revision: D14242264
fbshipit-source-id: ad21a179160a766304ff90f8c81e0563b990ebac
Summary: This was annoying me so I thought I'd rename it before anyone else starts to use the prop.
Reviewed By: danielbuechele
Differential Revision: D15919159
fbshipit-source-id: c351503797cedc5c11a158cec62515b1eb791a70
Summary:
When the text is too long either for the crash name and the crash reason, user is not able to see scroll and see the text.
Bug:
{F162636692}
Reviewed By: danielbuechele
Differential Revision: D15907908
fbshipit-source-id: a9557074e310728d4cb4825ec18740ca0ea0c12f
Summary:
Fixes the bug which didn't fire the crash notification.
Bug:
{F162618487}
Reviewed By: passy
Differential Revision: D15905154
fbshipit-source-id: 8b34cc567789fbf7f255663a14017cbe300ac387
Summary:
Seemed a bit weird typing a regex without monospace.
Still would be good to have syntax highlighting. I tried using prismjs, but it seems to not work by default on Input elements.
Reviewed By: danielbuechele
Differential Revision: D15918742
fbshipit-source-id: b44c228b30f37f97811d09c80b191b52282720f8
Summary:
This diff fixes the bug which showed the length of characters as the byte length when the response headers didn't have `content-length` in it.
This diff uses `Buffer.bytelength` api to calculate the same.
Reviewed By: danielbuechele
Differential Revision: D15899547
fbshipit-source-id: 29f4e8a741f4dd550b9191cc31bf6930065653b3
Summary:
Now that we finally have something we can screenshot ...
Just for the features page right now, will add some
"installation" instructions next.
Reviewed By: oprisnik
Differential Revision: D15898050
fbshipit-source-id: 9dfd3b1121799e7069336aac845ba2abdb3e42df
Summary: Just a minor version change, no changes expected.
Reviewed By: jknoxville
Differential Revision: D15897958
fbshipit-source-id: b103772159177e47884eb88e47f7b4077f77be03
Summary: Means that attribution is now part of the export, too.
Reviewed By: oprisnik
Differential Revision: D15877937
fbshipit-source-id: e089597269b8977320d06284179f72d40d01ebf2
Summary:
Now that the events tracking works again, make sure that the padding is not all over the place in the sidebar.
Before:
{F162525912}
Reviewed By: priteshrnandgaonkar
Differential Revision: D15875758
fbshipit-source-id: a30f057c8c0dc2dc77fc40f416ff565cc6621109
Summary:
Adds regex support to the logs plugin.
You can now do the same for any other SearchableTable, just by adding the prop `regexSupported={true}`
For other types of searchables, like maybe the layout hierarchy, we'll need to modify the class that does the filtering for that type.
I've done what I can with the UI to make it usable, though it would be awesome to have some proper regex syntax parsing and highlighting going on. It will go red if it's not valid at least.
Reviewed By: danielbuechele
Differential Revision: D15898806
fbshipit-source-id: 425edb1834dcc14ca741ac7fc8d566b4f2763c63
Summary:
deepLinkPayload is passed as initialQuery to Search component and in componentDidMount we perform search if initially query is defined.
This does not handle the case if node is not present in the layout inspector tree.
Reviewed By: danielbuechele
Differential Revision: D15874343
fbshipit-source-id: c604baea16838f07e8f8bfc0f1e67c5e830dfe97
Summary: Rows that were virtualized and currently not rendered weren't copied, because we were using the DOM to get their content. This changes this to use out `textContent` utility to get the value of a row and therefore enables copying rows that are not rendered.
Reviewed By: jknoxville
Differential Revision: D15898935
fbshipit-source-id: c372bc4e77477214860f0513b442374da8f72416
Summary: Copying/Creating a Paste from a network request only copied the domain. Now we are including all request/response data.
Reviewed By: jknoxville
Differential Revision: D15898038
fbshipit-source-id: 570ca5bb76e65b90b2b467856e6d3747682cc20d
Summary:
Where was a bug, where a range selection using the shift key wasn't working, when the start of the selection was outside the viewport.
The problem was caused by `selectInRange` firing twice, the second time with wrong arguments. In this diff, we are not firing the `onMouseEnterRow` handler, when the shift key is pressed.
Reviewed By: jknoxville
Differential Revision: D15897880
fbshipit-source-id: 95402dcbe9f8ce691be16cfd103918b09ff9e75c
Summary:
The previous setup would first check for the presence of a `persistedStateReducer` and then skip the standard subscriptions if this was found. This is a bit of a foot gun because simply by introducing a new function, you'd silently lose your existing subscriptions.
There are still some good reasons to support both as some subscriptions may not require persisting as they're directly tied to the state in the app at a given point.
Reviewed By: danielbuechele
Differential Revision: D15856372
fbshipit-source-id: a36bf40b1ceac431964610571eb70fff687b7607
