Fixes#9803 --the height of the banner was cutting off banner messages spanning multiple lines
Due to the animation of the view surrounding the banner text, calculate the banner height, then re-render the banner with the updated height.
Render a view that wraps not only the message banner itself, but also a hidden (0-opacity and absolutely positioned) view with the banner text that calculates the height of the banner during the `on-layout` event. The banner height calculation causes the reagent atom value change to re-render the banner view with the udpated height.
1. There is an animation on the message banner. The "out" animation (ie when the banner is in the process of disappearing) seems to work correctly. However, there did not seem to be an "in" animation. I'm fairly confident these changes did not break any existing animations, however it's worth noting this so that others who are more familiar with the existing codebase may be in a better place to say otherwise.
2. Testing during changes was done by dispatching the `network-info-changed` event like so:
```
(re-frame/dispatch [::network-info-changed {:isConnected false}])
```
Co-authored-by: Pascal Precht <pascal@status.im>
Co-authored-by: Eric Dvorsak <eric@status.im>
Signed-off-by: andrey <motor4ik@gmail.com>
We didn't check for encoded characters in the URL.
Instead now first we decode the URL components, and then match against
the URLs.
We also make sure that there's no HTML in the link passed.
Signed-off-by: Andrea Maria Piana <andrea.maria.piana@gmail.com>
In order to prevent `nix-store --gc` from removing too much I've:
- Added the `keep-outputs = true` setting to `nix/nix.conf`
- Fixed `nix/scripts/gcroots.sh` to make symlinks in `/nix/var/nix/gcroots`
- Made `nix/scripts/build.sh` and `nix/scripts/shell.sh` use it
This way when running `make nix-gc` most recently used shells and built
derivations won't be removed along with their dependencies.
Signed-off-by: Jakub Sokołowski <jakub@status.im>
Accessibility hide password only input
Do not hide all other elements, as only password should be secured
Signed-off-by: Gheorghe Pinzaru <feross95@gmail.com>
This way we can use different token for e2e, release, and devel builds.
Signed-off-by: Jakub Sokołowski <jakub@status.im>
Signed-off-by: Andrea Maria Piana <andrea.maria.piana@gmail.com>
As reported in #10127 there's an issue that the log level isn't properly
applied when changed through log settings UI.
Turned out that we're setting the log-level only once in the entire
lifetime of the application and that is during bootstrap.
When the log level is changed through the UI, we never actually update
the log level in the logger module.
To fix this, we just need to ensure that, whenever the log-level is changed,
we also call `log/set-level!` with the new level.
It should also be noted that, every time the log level is changed through
the UI, the user is logged out, so one place to update the logger is
after very successful login.
Fixes#10127
Signed-off-by: Andrey Shovkoplyas <motor4ik@gmail.com>
Before errors like this:
```
finder.go:91: error: 'no pom data' for: <Dep ID=com.afollestad.material-dialogs:commons:0.9.6.0 O=false S= >
```
Would be printed to `stderr` but exit code would not be non-zero.
Now the whole `make nix-update-gradle` target should fail if any
dependency resolution fails.
Signed-off-by: Jakub Sokołowski <jakub@status.im>
`multiaccount` was used only by simulated keycard, so it is removed
and will be retreived by accessing app-db directly. This is done because
simulated kk implementation should be compatible with real keycard and
thus can't require extra parameters.
This reverts commit db202f03bf.
Fix leftover get-ref for text input
Use uuid as a unique identifier for input
Generate a random uuid on mount, ensures that no stale ref are kept
Signed-off-by: Gheorghe Pinzaru <feross95@gmail.com>