feat(nix): known failures, with unknown_monitor the first

Checks.known_failure records a check that fails for a reason that is
to be followed up on: the suite does not fail, the totals say how many
known failures there are, and one that holds fails the suite so the
marker goes. unknown_monitor takes the reason, and core-e2e runs it
last, because the window it loses stays lost. BEHAVIOUR.md says so.
This commit is contained in:
LGUG2Z committed 2026-10-30 12:03:00 -07:00
1 parent 54d866f83a
commit 8fc7c961d0
6 files changed
+83 -8

No files matched your search

+6 -4
View File
@@ -36,7 +36,7 @@ runs, and goes stale when the code changes.
| `closing` | a closed window leaves the state, and the others take its room | pass | pass |
| `resize` | dragging the edge two tiles share moves it by what the pointer moved | pass | **fails**, cause found |
| `commands` | every layout, and about 216 commands, leave the daemon healthy | pass, none skipped | not run, see below |
| `commands.unknown_monitor` | a window sent to a monitor that does not exist stays managed | **fails** | **fails** |
| `commands.unknown_monitor` | a window sent to a monitor that does not exist stays managed | **known failure** | **known failure** |
| `rules` | initial, permanent and ignore rules by title | see below | not run |
| `hotplug` | an unplugged display takes its windows to the other, a plugged one is a monitor | pass | not run |
@@ -54,9 +54,11 @@ present, with the command exiting 0 and printing nothing. On kfde `core-e2e` fai
check on a real KWin, and the windows after it. kfde's code shows how:
`WindowManager::move_container_to_monitor` takes the focused container out of its workspace with
`remove_focused_container()` and only then looks the monitor up, with an error for one that is not
there, so the container is gone when the error is returned. This is not fixed, on either port. The
scenario is `commands.unknown_monitor`, which a suite can call; it is not in `core-e2e` while it
would fail there. The sweep of `commands` is not affected on kfde, which has two outputs, and on
there, so the container is gone when the error is returned. This is not fixed, on either port, and is a
**known failure to follow up on later**: `commands.unknown_monitor` takes a `known_failure`
reason, with which its checks are recorded by `Checks.known_failure`. They do not fail the suite,
the totals say `known failures: N`, and one that holds fails the suite to say the marker can go.
`core-e2e` runs it last (section 12c), because the window is lost. The sweep of `commands` is not affected on kfde, which has two outputs, and on
kfw the sweep now plugs in a second display first.
**`resize` on kfw.** Dragging the edge between two tiles by 30, 60 and 120 pixels moved it by 27,