GlobalScope has numerous problems[1] that make it unfit for
use in most applications and making it behave correctly requires
an excessive amount of verbosity that's alleviated simply by using
any other scope. Since we run multiple operations in the context
of the application's lifecycle, introduce a new scope that is created
when our application is, and cancelled upon its termination.
While at it, make the scope default to Dispatchers.IO to reduce pressure
on the UI event loop. Tasks requiring access to the UI thread appropriately
switch context making the change completely safe.
1: https://medium.com/@elizarov/the-reason-to-avoid-globalscope-835337445abc
Signed-off-by: Harsh Shandilya <me@msfjarvis.dev>
- Mark WRITE_EXTERNAL_STORAGE as unused above API 29 since we defer to Storage Access Framework on P and above
- Adds a <queries> tag to allow listing apps for exclusion/inclusion dialog
Signed-off-by: Harsh Shandilya <me@msfjarvis.dev>
Setting the correct value for the allowedAuthenticators field lets the library correctly detect this by itself as verified on an API 21 emulator
Signed-off-by: Harsh Shandilya <me@msfjarvis.dev>
- buildscript must always be the first block in a Gradle build
- ConstraintLayout, Kotlin and bintray plugin are updated to their latest stable revisions
- Biometrics is updated to the latest alpha release to make use of multiple memory leak fixes that plague the 1.0.x implementations
Signed-off-by: Harsh Shandilya <me@msfjarvis.dev>
Google recommends all dependencies with -ktx variants depend on them directly since they transitively pull in the main artifacts and offer extensions for better usage from Kotlin
Signed-off-by: Harsh Shandilya <me@msfjarvis.dev>
We do not support building on Windows for now and deleting this file on every wrapper update is just lame
Signed-off-by: Harsh Shandilya <me@msfjarvis.dev>
Layouts in differing configurations must agree on their root tags, so we give
both copies of main_activity the same root ID, and adjust the check for
two-pane layout to simply test for nullability. This also changes the inset
dispatch code to use ViewCompat and WindowInsetsCompat since they will adjust
insets based on the SDK level allowing us to abstract away that concern.
Signed-off-by: Harsh Shandilya <me@msfjarvis.dev>
The status bar color had a subtle transparency applied which caused
inconsistencies, and the navigation bar color was the wrong shade of
gray.
Reported-by: Danny Lin <danny@kdrag0n.dev>
Signed-off-by: Harsh Shandilya <me@msfjarvis.dev>
/data/adb/modules has been the location for modules since v18.0 and nobody
should reasonably be on any older version anymore. This has continued to
work for this long because Magisk created symlinks for backwards compat. However,
these symlinks are not created anymore on Android 11, which is where this
problem first surfaced.
Signed-off-by: Harsh Shandilya <me@msfjarvis.dev>
While Play Console correctly manages to deobfuscate stacktraces using the
ProGuard mapping, it makes user recorded logs useless which is rather
undesirable.
Signed-off-by: Harsh Shandilya <me@msfjarvis.dev>
Should have been part of the patch that disabled this for ui/
Fixes: 8d128cf2e9 ("ui: disable LongLogTag lint")
Signed-off-by: Harsh Shandilya <me@msfjarvis.dev>
When tunnel deletion is triggered we don't bother with animation theatrics
because the resulting Snackbar needs this fab to be its anchor, which it can't
do if its outside the screen or busy animating.
Signed-off-by: Harsh Shandilya <me@msfjarvis.dev>
We don't (and shouldn't) use BuildConfig values, but the class was
polluting our public API regardless which is undesirable.
Signed-off-by: Harsh Shandilya <me@msfjarvis.dev>
A set will cause identical peers to be dropped during save
Reported-by: Jason A. Donenfeld <Jason@zx2c4.com>
Signed-off-by: Harsh Shandilya <me@msfjarvis.dev>
When biometric hardware is available, it will be used to authenticate
the user before private keys are shown on screen or when zip exports
are executed.
Signed-off-by: Harsh Shandilya <me@msfjarvis.dev>
The correct way to retrieve the value inside a getter/setter is to use `field` to ensure
you don't invoke the getter while inside the getter and trigger a stack overflow
Signed-off-by: Harsh Shandilya <me@msfjarvis.dev>
This contains a share button and a save button, the former using a
custom content provider.
Co-authored-by: Jason A. Donenfeld <Jason@zx2c4.com>
Signed-off-by: Harsh Shandilya <me@msfjarvis.dev>
Signed-off-by: Jason A. Donenfeld <Jason@zx2c4.com>
Since the signing config is also stored in the project root it makes sense to have storeFile
path be relative to it.
Signed-off-by: Harsh Shandilya <me@msfjarvis.dev>
Includes a control set of broken configuration files that we attempt to parse and
verify that the parser fails in a predictable and consistent manner.
Signed-off-by: Harsh Shandilya <me@msfjarvis.dev>
This is covered under SYNTAX_ERROR which feels like a reasonable choice,
so get rid of this unused field and associated resource string.
Signed-off-by: Harsh Shandilya <me@msfjarvis.dev>
