52 lines
3.2 KiB
Markdown
52 lines
3.2 KiB
Markdown
# Desktop and advertising mouse input fixes
|
|
|
|
The optional desktop frame bit controls decoration, not whether the visible
|
|
widget contents accept input. The corrections reuse the Original renderer,
|
|
coordinates, layout records and action dispatchers.
|
|
|
|
- The date widget activates the same owned advance action as Anstoß, with or
|
|
without a frame. In a multiplayer turn this first hands control to the next
|
|
manager. The separate calendar/rescheduling menu keeps its own behavior.
|
|
- Stored icons remain clickable and draggable from unframed folders. Extraction
|
|
restores the actual icon record to the visible chain; the folder stays put.
|
|
Hidden or unlinked owners remain invalid. Removal corners keep their separate
|
|
transaction and mouse-release barrier. The same frame-origin correction also
|
|
covers the existing admission scale and five automation controls.
|
|
- A true double-click on an existing save opens the shared Original-style
|
|
overwrite question. A separate single click on Ja confirms, Nein returns to
|
|
the picker, and repeated double-click releases cannot write. Enter and cancel
|
|
paths remain. FBM, EIN and MEN use the same confirmation boundary.
|
|
- Advertising follows all nineteen original hit rectangles with palette0x27
|
|
outlines and the installed edition's detail/help text. Motion alone does not
|
|
change the career, revision or either random stream. Offers/calculators retain
|
|
their input owner; returning from them reevaluates the actual pointer, avoiding
|
|
a stale highlight without requiring another mouse movement.
|
|
|
|
## Validation
|
|
|
|
The focused configured-data tests are `app_desktop_input_tests`,
|
|
`app_advertising_input_tests`, `ui_desktop_input_tests`, `ui_save_load_tests` and
|
|
`ui_advertising_tests`. They cover frame variants, exact field boundaries,
|
|
canonical actions, stale/hidden-owner refusal, cancellation, input release,
|
|
overwrite confirmation and unchanged game/random state for presentation-only
|
|
input. The desktop input test isolates these checks from unrelated private
|
|
text/badge inventories; the existing full test retains its additional checks.
|
|
|
|
Private Linux SDL before/after captures and actual saved games are maintained in
|
|
`bmh_4k_ui/docs/qa/classic-extended-ui-input`. The fixtures use actual club/player
|
|
data with a controlled desktop layout. No results are fabricated. A successful
|
|
compile or CI run alone is not native Windows/macOS UI verification.
|
|
|
|
Tracked reports: [#51](https://github.com/Weyoune/openbmh/issues/51), [#52](https://github.com/Weyoune/openbmh/issues/52), [#53](https://github.com/Weyoune/openbmh/issues/53), [#54](https://github.com/Weyoune/openbmh/issues/54).
|
|
|
|
The primary regression run uses the supplied Diamond 3.7 installation. Unrelated
|
|
historical rendering fixtures retain their declared reference requirements.
|
|
Two existing Classic test fixtures were aligned with the already implemented
|
|
stadium-capacity policy and verified edition-specific crest gate; product
|
|
validation was not relaxed to satisfy those tests.
|
|
|
|
The overwrite prompt also maps the actual Football Limited English pointer.
|
|
The translation copies exactly that four-byte pointer and preserves adjacent
|
|
empty slots. The launcher translation-cache revision is advanced, so an existing
|
|
cached English installation receives the newly supported prompt as well.
|