CI / UI Checks (pull_request) Successful in 11s
CI / Rust Checks (pull_request) Failing after 4m11s
CI / Rust Feature Validation (pull_request) Skipped
CI / Frontend E2E (pull_request) Skipped
CI / Community Image Smoke (pull_request) Skipped
CI / Required Checks (pull_request) Failing after 1s
CI / Deploy (pull_request) Skipped
Story 3.18 (#25) surface: клиент получает только закрытые structured `notifications/message` выбранного уровня, без запросов, секретов и internal causes. - `logging` capability advertising в обеих поддерживаемых revisions; `logging/setLevel` — legacy-only dispatch (modern — `_meta` per request); - закрытый словарь LoggingLevel (RFC 5424), malformed → -32602 с одним safe detail; закрытый набор safe events, task-local request scope; bounded in-memory opt-in, без миграций; - `notifications/message` идут перед ответом в SSE; JSON-режим их не несёт; - ATDD в `scripts/ci-story-checks.sh`: mcp_logging (10), inventory projection (3), protocol::logging unit (12); - проекция capability inventory/frozen oracle переведена на `logging`. Story ATDD и affected проверки пройдены на этом дереве (fmt, clippy, python oracle 56 OK, capability inventory valid). Полный `just verify` — NOT_RUN на story по validation-policy, остаётся на собранной фиче.
42 lines
1.6 KiB
Bash
Executable File
42 lines
1.6 KiB
Bash
Executable File
#!/usr/bin/env bash
|
|
set -euo pipefail
|
|
|
|
active_log=""
|
|
cleanup() {
|
|
if [[ -n "$active_log" ]]; then
|
|
rm -f -- "$active_log"
|
|
fi
|
|
}
|
|
trap cleanup EXIT
|
|
|
|
run_atdd() {
|
|
local description="$1"
|
|
shift
|
|
active_log="$(mktemp)"
|
|
if cargo test "$@" >"$active_log" 2>&1; then
|
|
if ! grep -Eq '^test result: ok\. [1-9][0-9]* passed' "$active_log"; then
|
|
cat "$active_log" >&2
|
|
printf 'ATDD command selected zero passing tests: %s\n' "$description" >&2
|
|
return 1
|
|
fi
|
|
grep '^test result:' "$active_log"
|
|
rm -f -- "$active_log"
|
|
active_log=""
|
|
return 0
|
|
fi
|
|
cat "$active_log" >&2
|
|
printf 'ATDD command failed: %s\n' "$description" >&2
|
|
return 1
|
|
}
|
|
|
|
# Keep Story 3.16 prompt ATDD with Story 3.17 completion and its bound-template regression.
|
|
run_atdd "MCP prompt composition" -p mcp-server --test integration mcp_prompts::composition -- --test-threads=1
|
|
run_atdd "MCP completion" -p mcp-server --test integration mcp_completion -- --test-threads=1
|
|
run_atdd "bound template resources/list regression" -p mcp-server --test integration static_resources::list::list_stays_available_when_a_published_template_is_bound -- --test-threads=1
|
|
|
|
# Story 3.18: safe structured logging — the closed level vocabulary, the endpoint ATDD, and the
|
|
# capability projection the `logging` surface changes.
|
|
run_atdd "MCP log level vocabulary" -p crank-community-mcp --lib protocol::logging -- --test-threads=1
|
|
run_atdd "MCP safe logging" -p mcp-server --test integration mcp_logging -- --test-threads=1
|
|
run_atdd "capability inventory projection" -p mcp-server --test integration inventory_projection -- --test-threads=1
|