PaperMtn/slack-watchman 4.6.0

Slack Watchman - 4.6.0

on GitHub

Changed

• [High] Message and file workers now cache users.info and conversations.info lookups by ID for the lifetime of the worker, eliminating redundant API calls per match. Fixes #117
• [High] find_messages and find_files now use a bounded multiprocessing.Pool (capped at 8) instead of one Process per search string, preventing memory and file-descriptor usage from scaling with signature size. Fixes #113
• [High] Pool workers now construct their own SlackClient via _init_worker_client instead of unpickling the parent client, preserving connection pool and retry config across workers. Fixes #114
• [High] SlackClient.__init__ accepts new keyword-only session_token and cookie_dict parameters, used by pool workers to skip re-running _get_session_token on startup.
• [High] find_messages and find_files now share a single multiprocessing.Manager per call, replacing three per-call managers that leaked until GC. Fixes #116

Fixed

• [Critical] Worker exceptions (rate-limit exhaustion, malformed payloads, AttributeError) are now caught and reported via a shared errors list instead of silently killing the child process. Fixes #108
• [Critical] _multipro_message_worker no longer crashes on messages with no channel field (DMs, tombstoned messages, bot events). Fixes #109
• [Critical] _multipro_file_worker no longer crashes when a file's name or filetype is None (deleted/redacted/tombstoned states). Fixes #110
• [Critical] main() error handler no longer crashes with AttributeError when an exception is raised before init_logger runs; falls back to traceback.print_exc() when the logger is uninitialised. Fixes #111
• [High] SlackClient._make_request now honours the Retry-After header on 429s (falling back to 90s), loops back through _make_request for consecutive 429s, and raises after a configurable retry cap. Fixes #115
• [Medium] Canvas URLs are now built via _build_canvas_url(), which normalises the trailing slash on the workspace URL before concatenation. Fixes #126
• [Medium] after: query timeframes are now computed in UTC (time.gmtime) instead of local time, fixing off-by-one-day windows on CI/cloud hosts; refactored into a _compute_timeframe() helper. Fixes #125
• [Medium] find_auth_information return type widened to Dict[str, Any] | None to reflect the actual mixed-type payload. Fixes #124
• [Medium] find_auth_information now catches RequestException, JSONDecodeError, and AttributeError, returning None with a WARNING instead of aborting the scan on transient scrape failures. Fixes #123
• [Medium] find_messages and find_files now explicitly return [] on no-match and exception paths instead of implicit None. Fixes #118
• [Medium] _multipro_file_worker no longer appends a duplicate result per sig.file_types entry; replaced per-type loop with a single any() check. Fixes #119
• [Medium] Removed dead is_dataclass guard and unused dataclasses import from _resolve_file_user. Fixes #120
• [Medium] _multipro_message_worker now skips messages with no text instead of coercing None to the string 'None', preventing false positives on attachment-only and system messages. Fixes #122
• [Medium] _multipro_message_worker now compiles each pattern once per worker and reuses the Match object, eliminating per-message recompilation and a duplicate search call on the match path. Fixes #121
• [Low] cursor_api_search now updates cursor once per page after items are collected, fixing a potential infinite loop on empty pages with a non-empty next_cursor.
• [Low] OUTPUT_LOGGER annotation in __init__.py widened to StdoutLogger | JSONLogger.
• [Low] Added explicit parentheses to if everything or (not pii and not secrets): for readability; no behaviour change.