summaryrefslogtreecommitdiff
path: root/docs/next-tests.md
diff options
context:
space:
mode:
Diffstat (limited to 'docs/next-tests.md')
-rw-r--r--docs/next-tests.md13
1 files changed, 13 insertions, 0 deletions
diff --git a/docs/next-tests.md b/docs/next-tests.md
index 7f80fa7a..059faf09 100644
--- a/docs/next-tests.md
+++ b/docs/next-tests.md
@@ -50,6 +50,19 @@ agent's survey. The `e2e` line (kbdin, mousein) is its first step.
agent_session.py tool.
- **Still Python in the build:** gui_monkey.py (monkey-gui, for T4).
+ Each backend's own (2026-11-04, the user: an effect's test is the
+ backend's):
+ - **`zig build process-test -Dplatform=tty` drives `pardes`, `-Dplatform=gui`
+ drives `pardes-gui`,** hidden (PARDES_TEST, SDL_VIDEO_DRIVER=offscreen,
+ no display of the user's) on a pty of its own and a 160x50 grid.
+ `Detached.frame()` waits for the window to say it is idle and reads its
+ capture.
+ - **A case about one frontend's own behaviour says `process.only(.tty)`.**
+ On the window, 31 cases pass and 5 skip.
+ - **Open:** on the window the canary's Newcol is refused after its
+ terminal steps (a fresh window takes it), and one run hung there. The
+ canary is tty-only until that is found.
+
Before the port, the plan was: the Python suites in the build (fs.py, monkey9p.py, gui_monkey.py,
agent_session_test.py) move onto it or onto Session, then
leave the build. The developer tools (ui_review, record_demos,