diff options
Diffstat (limited to 'docs/next-tests.md')
| -rw-r--r-- | docs/next-tests.md | 39 |
1 files changed, 29 insertions, 10 deletions
diff --git a/docs/next-tests.md b/docs/next-tests.md index 054958b5..838202eb 100644 --- a/docs/next-tests.md +++ b/docs/next-tests.md @@ -22,8 +22,8 @@ agent's survey. The `e2e` line (kbdin, mousein) is its first step. `focused`), plus `signal`, the exit code, pty bytes and captured frames by mode. It is mostly test/snapshot.zig's launching and 9P client moved out: - - **the `.snap` scripts** become one user of it (the interpreter and the - goldens stay in snapshot.zig); + - **the `.snap` scripts** became its tty cases (2026-11-09, below: The + suites), their goldens kept in test/snapshots/; - **process tests (T2)** are plain Zig tests in test/process/<label>.zig, under a `process-test` step; - **the window's tests (T4)** are `Process` in its window mode. @@ -40,8 +40,8 @@ agent's survey. The `e2e` line (kbdin, mousein) is its first step. Ported (2026-11-04, a fan-out of three): - **fs.py's 38 cases:** 36 process tests in test/process/, a few at T1 in test/behaviour/. - - **monkey9p.py:** test/monkey9p.zig, at T1 (`monkey-session`, its smoke - run part of behaviour-test). + - **monkey9p.py:** test/monkey9p.zig, at T1 (the T1 runner: its smoke + run part of behaviour-test, `behaviour-test -- --seed N` the rest). - **The build steps fs-test, fs-discovery-test, monkey-9p, monkey-9p-smoke and agent-session-test are gone.** - **fs.py is kept** as the helpers the developer tools import (session, @@ -112,7 +112,7 @@ agent's survey. The `e2e` line (kbdin, mousein) is its first step. |---|---|---| | T1 | the core through `serveFs`, in process | most behaviour; each case also through a replica that must agree | | T2 | a real `pardes` over its socket, a Zig runner | daemon, attach, mounts, shells, flags | -| T3 | terminal bytes, the `.snap` scripts | tty rendering and keys | +| T3 | terminal bytes: process-test's tty cases, a pardes in a pty of its own (test/process/tty_*.zig) | tty rendering and keys | | T4 | the window: `pardes-gui` hidden (`PARDES_TEST`, offscreen), its input through the same 9P files, frames captured once its status file says idle | what only pixels show | | perf | T1 or T4 gestures, timed | latency gates against baselines | @@ -122,11 +122,9 @@ images, fonts, edges). Built (2026-11-04): test/process/window.zig's monkey (a smoke run in process-test, all 40 Motion x Lift x Bloom with `zig build monkey-gui -Dplatform=gui`) drives the window's own input (Detached.typeKeys and mouse: its pty and OSC 777) and checks it settles, -alive, with no crash record; its goldens (`zig build gui-golden --Dplatform=gui`, PARDES_GUI_GOLDEN=update to rewrite) hash 17 scenes' -frames from Detached.capture. The old notes: test/gui_golden.py's 17 frame hashes -(test/gui-goldens.txt) are an exact gate for one machine, never a CI one, -and move to the Zig runner with T2. gui.zig's own tests: 19 M, 17 B, 47 U. +alive, with no crash record. Its goldens (15 scenes' frame hashes, exact +for one machine) went on 2026-11-09: what they caught, gui.zig's unit +tests and the tty cases catch (The suites, below). gui.zig's own tests: 19 M, 17 B, 47 U. Performance: test/perf.zig times gestures through `update` and `render` with a few internal reads (12), against test/perf-baseline-*.json (the @@ -138,6 +136,27 @@ component stays a U-level bench of that component. U tests (a stable library at its own pub functions with no core: regexp, diff, sam addresses, walk, wire, Text, modal's parse) stay as they are. +## The suites (2026-11-09, step 11 of the clean-room plan) + +Sixteen harnesses folded into the kept set: T1, hxdiff, one process suite, +one soak. Each fold was checked by a break of the code its cases guard, +failing the kept case as it failed the dropped one. + +| was | is | +|---|---| +| `snap`: 103 `.snap` scripts through pardes-snap (test/snapshot.zig) | the tty's own process cases, test/process/tty_{boot,editing,layout,themes,images,terminals}.zig, each a script's steps as `Tty` calls (test/tty.zig), its golden in test/snapshots/ unchanged; `process-test -- --update <filter>` rewrites one | +| `snap-driver-test` (the runner's retry and report) | process_test.zig's pool: a child a case, a core's worth at once, a failure run again alone with the wider settle probe | +| `monkey` (random snapshot scripts) | `behaviour-test -- --monkey <seeds> [--steps M]`: the same input mix at an in-process session (/kbdin, /mousein, Session.resize), a panic or a hang a failure, shrunk | +| `monkey-test` (did not compile) | gone: the generator is test/monkey9p.zig's genInput | +| `replay` (every script's and 20 seeds' tapes) | `behaviour-test -- --replay` (each seed's sessions taped in memory and replayed: three replicas and the dump) and `process-test -- --replay [--replay-with=<pardes>]` (the tty cases' real sessions, their hosts' answers on the tape) | +| `monkey-session`, `monkey-session-smoke` | `behaviour-test` bare runs the smoke; `-- --seed N ...` the full 9P monkey | +| `hxparity` | `hxdiff -- --parity` | +| `gui-golden` (15 frame hashes, this machine's) | the window's unit tests (`unit-test -Dplatform=gui`), the image and kitty harnesses' window arms, the tty cases for what the core draws, `monkey-gui` | + +What stays as it is: the image, kitty and pdf harnesses (end to end +through real programs and the GPU capture), pane-test (pane, output and +namespace integration), replicas, the soak and test/daemon.py, perf. + ## Inventory (2026-11-03, a regex classifier: about ±15%) | class | tests | lines | |
