52 lines
3.2 KiB
Markdown
52 lines
3.2 KiB
Markdown
# Editor 0.5.1 — ratings and salaries
|
||
|
||
These are actual 1440×900 SDL2/OpenGL/RmlUI captures on Linux under Xvfb
|
||
with software OpenGL. All content is the synthetic `World::demo()` fixture:
|
||
52 invented players, two invented clubs, one league and 50 explicitly synthetic
|
||
per-season contracts (one unknown salary per club). No real-world data, original
|
||
game images, personal paths, settings or credentials are included. This editor
|
||
uses starting worlds, not a career save; reproduce the data with `--demo`.
|
||
|
||
| View | Evidence |
|
||
|---|---|
|
||
| Before, released Editor 0.5.0 | [Player inspector](before-players-0.5.0.png), without a ratings/payroll overview |
|
||
| Players, Midnight Glass / German | [Salary descending](players-midnight-de.png) |
|
||
| Teams, Silver Glass / English | [Player-weighted means, payroll and coverage](teams-silver-en.png) |
|
||
| Leagues, Matchday / German | [League aggregate](leagues-matchday-de.png) |
|
||
| Attribute selection | [Choose columns](columns-midnight-de.png) |
|
||
| Horizontal navigation | [Further attributes](players-scroll-midnight-de.png); header and values remain aligned |
|
||
|
||
The after views use Editor 0.5.1. The before capture uses the released 0.5.0
|
||
Linux executable and its own assets. No screenshot is a browser mockup.
|
||
|
||
Validation:
|
||
|
||
- `editor_core_tests`, `editor_comparison_tests`, `editor_version` pass locally.
|
||
The comparison suite checks 36 outcomes, including actual currency units,
|
||
unknown versus zero salaries, date/team selection, conflicting contracts,
|
||
staff exclusion, complete reserves, duplicate evidence, player-weighted means,
|
||
coverage, empty groups, numeric sorting across more than 200 rows, stable ties,
|
||
nulls last in both directions, different rating systems, overflow, read-only
|
||
session state, cache refresh after editing/undo and the shared MCP tool.
|
||
- `tests/editor_gui_smoke.py` performs real keyboard/mouse operations. It selects
|
||
a rating column via its label, sorts ratings and salaries, opens the first row
|
||
and edits it. MCP verifies the expected stable player/club/league identifier
|
||
changed, then undoes it. The existing inspector/search/JSON/undo/save checks
|
||
still pass. The comparison leaves the saved world clean.
|
||
- All three themes and both languages were inspected. Header arrows do not rely
|
||
on missing font glyphs; horizontal scrolling keeps column alignment.
|
||
- The completed private season was checked independently: 18,920 eligible
|
||
player rows, 545 teams, 32 leagues; 832 source-only duplicate rows excluded.
|
||
Every eligible player has one known current-team salary in that package.
|
||
These are dataset estimates, not verified real wages. No private rows or
|
||
screenshots are distributed in this public directory.
|
||
|
||
Build the targets and run the CTest commands in [the editor guide](../../EDITOR.md).
|
||
For the native input test, run `python3 tests/editor_gui_smoke.py <OpenBMHEditor>`
|
||
inside a graphical session or Xvfb with xdotool. `OpenBMHEditor --demo --page
|
||
comparison` opens the overview; the three buttons choose players, teams or
|
||
leagues. **Choose columns** selects additional attributes.
|
||
|
||
Windows/macOS native build and core tests run in the editor CI. The Linux
|
||
window checks do not establish human acceptance on those operating systems.
|