Reverts the preloading mechanism from PR
https://github.com/status-im/status-mobile/pull/19927, which introduced a
problem I couldn't find any satisfactory solution. The namespaces user and
dev.user were being preloaded by shadow-cljs, but the problem is that preloading
will fail if those namespaces don't exist in your file system.
I couldn't find any solution to conditionally preload namespaces via shadow-cljs
and ClojureScript unfortunately doesn't support conditional requires like
Clojure.
At least we got to keep the code to not lint them and the whole idea to start to
commit dev-only code inside src/dev/.
Preload the user namespace (src/user.cljs and src/dev/user.cljs) for the mobile
target and for dev-only purposes. The files are git-ignored.
Just a reminder that you'll be responsible for making sure your user namespace
is correct. If it's broken in any way (e.g. calling non-existent code) the app
will crash at initialization (dev-only environment obviously).
Why? When the app initializes, it loads namespaces that were required at least
once. If you create a user namespace, it won't be automatically required for
you. And if you, like some Clojure devs, like to use the user namespace as
your safe heaven for experimentation and dev-only utilities, you'll need to
remember to evaluate the namespace at least once.
This is tedious and many times I forgot to do so and the app crashed because the
compiler didn't know where the symbols were coming from.
This commit rewrites utils.re-frame/delay-render to use hooks. The new
implementation renders significantly better than what we have today, at least on
Android.
Why not hiccup instead of a function call to delay-render? The Settings screen
is rendered slightly faster if I use delay-render as a function call instead of
hiccup. My only guess is that this is just less work to be done by Reagent,
since the wrapper function is not creating a wrapper component with its own
lifecycle.
The full analysis can be found here, but it's copied for future reference:
--------------------------------------------------------------------------------
Based on my analysis of individual frames being rendered and having investigated
3 different scenarios:
1. Scenario 1: No delay whatsoever, i.e. not using `delay-render`.
2. Scenario 2: Using `delay-render` like in `develop`, that is, a form-2
component with a local Reagent atom.
3. Scenario 3: Using `delay-render` as in this PR, using hooks.
All 3 scenarios open the Settings screen with all rendered views in the same
amount of time. In terms of raw performance, they are completely identical. The
absolute value doesn't matter, but in my recordings, on average, 10 frames of
video after the first press on the user's profile image.
So how can it be that on Android the new solution is visibly smoother? It's all
about latency and our brains are very picky about it.
Scenario 1 - Not using delay-render: the user notices a longer delay after
pressing on the profile image because all components in the Settings screen are
mounted in one go. This gives the impression to the user of being slower. In
slower Android devices, we've seen a user even press twice because the Settings
screen was taking longer to open. On newer Android devices this is not much of a
problem. There's another problem in `Scenario 1`, on Android, with too many
elements and/or too many heavy elements being mounted, the opening animation is
sometimes completely cut off or very clunky (a similar problem can happen while
opening the Activity Center).
Scenario 2 - Use delay-render with a form-2 component: The Settings items are
always rendered after the opening animation completes. Our brains perceive this
as a slight delay because we can see the empty gray background for 1-3 frames.
This is quite noticeable on my physical Android device, even with a prod build.
Scenario 3 - Use delay-render as a hook: the optimal solution from the user's
perspective, Settings items sometimes can be rendered before the animation
completes. I say sometimes because other times the items are rendered only 1
frame before or right when the animation completes, which would be almost the
same as Scenario 2.
What the hooks solution gave us is a little bit of the Scenario 1 and
Scenario 2 in one package, and because the Settings items can be sometimes
rendered before the opening animation completes, our brains see that as being
faster.
In future performance investigations, we might want to focus on manipulating
latency more aggressively to see where that leads us.
fixes#19802
Contract Tests are failing with missing symbols runtime error on `MacOS` after `golang` version was upgraded to 1.20
A simple hack fix such a case is to use `go 1.19` only to build the status-go library that is used specifically for these integration tests.
Its ugly but it works and unblocks devs from running tests locally.
- execute `make test-contract` on MacOS and it should not fail.
not needed since this impacts only the integration tests.
- macOS
status: ready
Some devs reported invalid schemas in the develop branch during app
initialization. We knew this could happen when Malli was first
introduced, but we wanted to play safe in the beginning due to the
overall inexperience of the team with Malli.
This commit removes all guardrails on instrumented vars, i.e. during app
initialization the app will crash on any invalid schema. A reminder that
instrumentation only take effect when js/goog.DEBUG is true.
## Summary
any `yarn` command calls `node-pre-gyp rebuild` which prints out this warning message :
```
make[1]: Entering directory '/Users/siddarthkumar/code/experiments/status-mobile/build'
CXX(target) Release/obj.target/status_nodejs_addon/modules/react-native-status/nodejs/status.o
../modules/react-native-status/nodejs/status.cpp:1153:11: warning: ignoring return value of function
declared with 'warn_unused_result' attribute [-Wunused-result]
func->Call(isolate->GetCurrentContext(), v8::Null(isolate), argc, argv);
^~~~ ~~~~~~~~~~~~~
1 warning generated.
```
This happens due to return value of `func->Call` here :
52a6f5c17d/modules/react-native-status/nodejs/status.cpp (L1153)
not being used.
Which is not a huge red flag but produces a warning which is not nice when looking at logs and can be sometimes misleading for other devs.
This commit fixes that warning by assigning return value of that function and checking if its empty or not.
## Testing notes
not needed since this impacts only the integration tests.
## Platforms
- macOS
status: ready
* chore: add "key pairs and accounts" label
* chore: feature flag wallet-settings
* tidy: extact navigate-back function into static defn
* wip: add initial keypairs and accounts list view to wallet settings
* tweak: wire-up initial action menu for key-pairs
* tidy: extract key-pair container styles into style namespace
* tweak: fix dark background for key-pair and account settings
* tidy: refactor on-press handler for key-pair options
* fix: move feature-flag usage to settings screen instead of settings items definition
* tidy: remove unneeded key props
* tidy: clean up de-structuring and passing of props
* tidy: use keep with when expressions instead of filter and map expressions
* tidy: rename the wallet-settings feature flag
* tweak: rename and add feature-flags for mobile wallet settings
* tweak: use scrollview for feature-flags and add spacing between feature-flag groups
* tweak: adjust the way feature-flags are displayed in groups
* tidy: remove unneeded prop
* tidy: use bottom-inset for padding key-pair and accounts list
* tidy: change `filterv` to `filter`
* tidy: use subscription for building account-props
* tidy: use subscription to build the entire keypair-account
* tweak: use key-pair type to determine default key-pair
* tidy: rename component to settings-category-view
* tidy: use assoc instead of merge
* tidy: extract function from subscription
* test: add tests for formatting key-pairs and accounts for wallet settings
* tweak: use `match?` instead of `=`
* tidy: use `swap!` without anonymous functions
This PR upgrades clj-kondo (our Clojure linter) from v2023-09-07 to
v2024-03-13, so ~6 months of development updates.
You can check out the changes starting at
https://github.com/clj-kondo/clj-kondo/blob/master/CHANGELOG.md#20231020.
Nothing terribly useful to us this time, but as usual, clj-kondo can catch
more problems and more reliably than before.
This commit improves in-app feature flags to persist what is currently only
stored in a Reagent atom by using RN Async Storage
https://reactnative.dev/docs/asyncstorage. This should make them more convenient
to use, which is a good thing overall for developers.
Additionally, there's now a top-right button in screen Settings > Feature Flags
that will reset the flags to the initial values obtained from environment
variables.
These in-app feature flags are exclusively available in debug builds in
Settings > Feature Flags, and only visible when flag ENABLE_QUO_PREVIEW is
enabled. There's no impact whatsoever in prod builds. A reminder that they are
not meant to be used by users (yet).
It's worth noting that RN has deprecated Async Storage and now recommends other
community solutions, but for a dev-only feature, I think it's fine.
Detailed explanation:
The schema failure is due to utils.image-server/get-initials-avatar-uri being
called with a nil profile customization color right after the user confirms
logout.
Right after logging out, the subscription :profile/profile-with-image is
recomputed. One of its signal inputs is :profile/profile. Right after logout,
the output of sub :profile/profile is always nil (this is correct, nobody is
logged in). This means that the sub :profile/profile-with-image will try to
calculate the multiaccount URI by passing a nil profile. This is wasteful
computation and is also the cause of the schema for
utils.image-server/get-initials-avatar-uri to fail, because it expects the
profile's customization-color to be present.
* tweak: refactor image-uri helpers
* fix: add `:primary` as customization color for profile images
primary is being used as the default color for accounts that were migrated without customization color to a default color
fixes#19770
Devs are often unaware of the supported Xcode version and often run into weird build issues with either higher or lower Xcode versions.
We currently only support Xcode 15.1
This change ensures we convey that change when devs execute `make run-ios`
This commit also attempts to ensure `xcodebuild` command is wrapped with `XcodeWrapper`.
- iOS
- macOS
status: ready
On brand new ubuntu setup `make run-android` fails with :
```
Task :app:mergeExtDexDebug
FAILURE: Build failed with an exception.
What went wrong:
Execution failed for task ':react-native-reanimated:configureCMakeDebug[arm64-v8a]'.
[CXX1300] CMake '3.22.1' was not found in SDK, PATH, or by cmake.dir property.
BUILD FAILED in 2m 30s
673 actionable tasks: 673 executed
error: cannot unlink '/tmp/tmp-status-mobile-a094d218d/
nix-build-status-mobile-build-debug-android.drv-0/tmp.moQHI1tJWE/caches/8.4': Directory not empty
removed '/tmp/tmp-status-mobile-a094d218d/tmp.RpThXxcRZR'
make: *** [Makefile:278: run-android] Error 1
```
we do have `cmake` present in Android Shell and yet the build process is not able to find it.
This commit makes sure that we explicitly provide that reference.
Parvesh Monu
Mohsen
Alexander
mmilad75
Nikolay
Omar Basem
Icaro Motta
Siddarth Kumar
Sean Hagstrom
flexsurfer
Brian Sztamfater
Lungu Cristian
John Ngei
codemaster
frank
Jamie Caprani
Yevheniia Berdnyk