Cover the ConPTY win32-input-mode key-record path #118

New issue

Open

opened 2026-09-05 00:20:08 +09:00 by natsukium · 0 comments

natsukium commented

2026-09-05 00:20:08 +09:00

Owner

Migrated from docs/backlog.md (Deferred / ConPTY win32-input-mode path).

The backend's spawn / read / exit / resize test and the crate's cross-platform PTY tests run green against real conhost on the windows runner, and the known behavioral gap (reader EOF arrives at ClosePseudoConsole, not child exit; recorded in crates/felis-pty/src/windows.rs) is confirmed there. Still uncovered by any test: the win32-input-mode key-record path (PSEUDOCONSOLE_WIN32_INPUT_MODE), which needs an interactive child rather than an echo. Blocked on a test fixture that is that child, not on a runner.

Add a fixture child that exercises PSEUDOCONSOLE_WIN32_INPUT_MODE key records.

Refs: crates/felis-pty/src/windows.rs, docs/reference/protocols/key-encoding.md win32-input-mode.

Migrated from `docs/backlog.md` (Deferred / ConPTY win32-input-mode path). > The backend's spawn / read / exit / resize test and the crate's cross-platform PTY tests run green against real conhost on the `windows` runner, and the known behavioral gap (reader EOF arrives at `ClosePseudoConsole`, not child exit; recorded in `crates/felis-pty/src/windows.rs`) is confirmed there. Still uncovered by any test: the win32-input-mode key-record path (`PSEUDOCONSOLE_WIN32_INPUT_MODE`), which needs an interactive child rather than an echo. Blocked on a test fixture that is that child, not on a runner. Add a fixture child that exercises `PSEUDOCONSOLE_WIN32_INPUT_MODE` key records. Refs: `crates/felis-pty/src/windows.rs`, `docs/reference/protocols/key-encoding.md` win32-input-mode.

natsukium added the

priority/P2

label

2026-09-05 12:11:16 +09:00