- fix the control's width (don't have to be explicit, the QQC2 Control
takes care of calculating the implicitWidth/Height automatically based on
contents)
- while at it, fix the metrics to be able to support RTL
languages too (it was off for Arabic)
- some smaller cleanups too
- Created StatusMainLayout to be used as the apps' main
layout
- Created StatusSectionLayout to be used for all sections as
their base
- Created StatusToolBar as part of StatusAppLayout with
standard component the ActivityButton and posibility to
add custom implementation components
Needed for https://github.com/status-im/status-desktop/issues/6635
The culprit here is at the very bottom, in StatusListItem: the toplevel
item was a MouseArea handling the clicks and this causes problems as any
extra buttons placed on top of it get their mouseClick events delivered
in wrong order (after StatusListItem). Fix this by having a MouseArea
handling click events behind the actual toplevel item as a last resort,
catch all handler plus the actual "sensor" being just a MouseArea that
handles merely the hover events
- drop unneeded onPressed/onReleased handlers in StatusChatListCategoryItem
- fix some warnings (typos, unreachable code, shadowed variables)
Fixes: #6733
Scope was to validate why StatusChatListCategoryItem miss behaves when
used in StatusChatListCategory.
On its own all the events works fine, as it can be seen by monitoring
the click events for each item (title, item and buttons).
However, inside StatusChatListCategory the mouse events are overlapped and
button events trigger also for the item. Tried obvious fiexes like
disabling steal of mouse events from buttons but this break the hovering
effect.
Updates: #6733
Fix by disabling layering, leftover from the previous approach of not
using Canvas.
Also investigate TODO regarding artefacts workaround and implement
a proper approach
Fixes: #6640
* fix(StatusAccountSelector): Refactored and reimplemented using `StatusComboBox`
* fix(StatusAssetSelector): Refactored and reimplemented using `StatusComboBox`
* fix(StatusColorSelector): Reimplemented using `StatusComboBox`
* fix(StatusCardPage): Replaced `StatusSelect` with `StatusComboBox`
* fix(StatusSelect): Expose a property to make arrow in the select item visible/invisible
* chore: Remove StatusAssetSelector from StatusQ as its base component StatusSelect is enough to build this further on desktop side.
This component also has dependecy on BE and hence should be moved out.
BREAKING CHANGE:
* chore: Remove StatusAssetSelector from StatusQ as its base component StatusSelect is enough to build this further on desktop side.
This component also has dependecy on BE and hence should be moved out.
BREAKING CHANGE:
* fix(StatusSelect): Expose a property to make arrow in the select item visible/invisible
- Resolved binding loops in StatusInput
- Replaced all instances of StatusBaseinput
with StatusInput
- Cleaned up min/max height properties
- Fixed total height calculation
Needed for https://github.com/status-im/status-desktop/issues/6358
- cmake is superior to qmake (e.g. linking custom libraries)
- maintaining two build system generators is costly
iterates: status-im/status-desktop#6510
1. Fix the issue of floating buttons having incorrect color in Dark theme
2. Added balances param in StatusAssetSelector. ideally should be removed in another task as it is not the correct place for business logic
Similar to what we did in status desktop under
d0a26a326e
- since Qt 5.xy, hover is not enabled by default for QQC2, so enable it
unconditionally as we are a desktop app anyway
- this fixes several hover effects being broken, mostly for builtin
components like MenuItem and some buttons (eg. the leftmost NavBar)
where we haven't enabled those with `hoverEnabled: true` explicitely
since most of the compilation time is spent on creating the QRC bundle
and the resulting binary is quite huge, advice the `rcc` compiler to add
support for "big resources"
(cf https://doc.qt.io/qt-6/resources.html#embedding-large-files)
Also pass the env $LDFLAGS to qmake (to be able to link e.g. with mold,
which again greatly reduces the time to link the binary)
this leads to 6-10x faster time when recompiling
(Linux, gcc 12.1.1, mold 1.3.0)
BEFORE:
Clean build:
real 0m57,256s
user 1m25,869s
sys 0m40,963s
Incremental build, one file touched:
real 0m54,323s
user 0m54,012s
sys 0m2,934s
AFTER:
Clean build:
real 0m7,276s
user 0m21,001s
sys 0m37,256s
Incremental build, one file touched:
real 0m5,569s
user 0m4,850s
sys 0m0,840s