fix: handle oversized Slack message edits

[cong91]

Summary

• chunk oversized Slack edits before calling chat.update
• reuse or delete overflow thread replies across repeated edits to avoid stale duplicates
• keep overflow replies in the correct parent thread and bound in-memory tracking caches
• add regression tests covering overflow splitting, repeated edits, thread-root handling, and cache pruning

Test Plan

• . .venv/bin/activate && pytest tests/gateway/test_slack.py -q -o 'addopts='

[teknium1]

Thanks for identifying the missing Slack chat.update length handling. The defect remains on current main: plugins/platforms/slack/adapter.py:1487-1501 passes formatted content directly to chat_update despite the 39,000-character limit at :421.

Problems

• The patch targets gateway/platforms/slack.py, but 5600105478ffde29d7566b45421b100eaa29c4ef moved the live adapter to plugins/platforms/slack/adapter.py; this requires a port rather than a clean application.
• A port must honor the current finalize contract. gateway/stream_consumer.py:1630-1661 passes it through, while Telegram and Discord keep one bounded preview during streaming and split only on final delivery (plugins/platforms/telegram/adapter.py:3946-3978, plugins/platforms/discord/adapter.py:2257-2279).
• Preserve Slack's current final Block Kit behavior for single-chunk edits (plugins/platforms/slack/adapter.py:1493-1501) and disable it when overflow chunks are used, as send() already does at :1389-1403.

Suggested changes

• Port the implementation/tests to the bundled Slack plugin and cover both finalize=False and finalize=True overflow behavior.
• Return/track continuation IDs consistently with the stream consumer if final overflow changes the editable target.

Automated hermes-sweeper review.

[teknium1]

Closing as superseded by #69317 (merged): targeted the pre-plugin-migration file; oversized-edit handling landed via #33224's reapply + block fallback.

Thanks for digging into this — the consolidated fix stands on the cluster's collective analysis, and your work is credited in #69317's summary.