fix(qqbot): add .md and other doc extensions to extract_media() regex - #22492
Open
Allonz wants to merge 12 commits into
Open
fix(qqbot): add .md and other doc extensions to extract_media() regex#22492Allonz wants to merge 12 commits into
Allonz wants to merge 12 commits into
Conversation
- Add media file upload support to _send_qqbot function
- Support chat_type detection from target format (c2c:/group:/guild:)
- Upload media via QQ Bot v2 API (/v2/users/{openid}/files, /v2/groups/{group_openid}/files)
- Map file extensions to QQ Bot file_type (1=image, 2=video, 3=voice, 4=file)
- Include media in message payload via 'file_info' field
- Update error messages to include qqbot in supported platforms
- Update schema description with qqbot target format examples
… attempts Fix 4 issues in QQBot adapter reconnect logic: 1. [P0] _listen_loop exit now calls _set_fatal_error() to notify Gateway When reconnect attempts are exhausted, the adapter now sets a retryable fatal error so _platform_reconnect_watcher can take over. Previously the listen loop would silently die, leaving QQBot in a zombie state where the process is alive but no messages are received. 2. [P1] CLOSED/ERROR WS events now raise QQCloseError instead of plain RuntimeError, preserving close code/reason for proper error classification. 3. [P1] Added 15s cooldown after reconnect failure to give QQ server time to clean up the old session before the next attempt. 4. [P2] Moved _heartbeat_interval reset inside _reconnect() try block so it only resets after a successful connection, not on failure.
…a support) Keep both QQBot and Feishu native media delivery blocks. Preserve QQBot's full chat_type routing and base64 file upload logic from HEAD, while incorporating Feishu's media support and thread_id parameter from upstream/main. Update error/warning strings to list both platforms.
…ping/pong Bug 5: _reconnect() opens a new WebSocket but never recreates _heartbeat_task. The old heartbeat task is orphaned on the dead connection — no heartbeat is sent on the new one. QQ server closes the connection after 60s with code=None (no close code because the server drops it cleanly). Reconnect succeeds, but the death loop repeats every 60s: connect → 60s silence → disconnect → reconnect → repeat. Changes: - _open_ws(): add heartbeat=20 to aiohttp ClientWebSocketResponse for WS-level ping/pong, preventing idle disconnects at the transport layer - _reconnect(): cancel old _heartbeat_task before opening new WebSocket - _reconnect(): create_task(_heartbeat_loop()) after successful reconnect
MEDIA: tag extraction silently ignored .md, .json, .xml, .html, .log, .yaml, .yml, and .toml files because the extension allowlist in BasePlatformAdapter.extract_media() didn't include these formats. This prevented the gateway's _deliver_media_tags_streaming() and the send_message tool from delivering document files produce d by agents (e.g. markdown reports, logs, config files) via QQBot and other platforms that use send_document(). Fix: add md|json|xml|html?|log|yaml|yml|toml to the regex extension alternation. All 183 existing tests pass with no regressions. Verified: - MEDIA:/path/to/file.md → extracted correctly - MEDIA:/path/to/file.json → extracted correctly - Existing formats (pdf, docx, png, etc.) → no regression
alt-glitch
added
type/bug
This was referenced May 19, 2026
13 tasks
This was referenced May 22, 2026
Closed
teknium1
reviewed
Jul 13, 2026
teknium1
left a comment
teknium1
left a comment
Contributor
There was a problem hiding this comment.
Thanks for tracing the original MEDIA extraction failure. The main extension-list work was superseded by commit 781604ce4 (fix(gateway): unify MEDIA: extraction extension set + close the unknown-ext black hole), which replaced the inline regex with the shared MEDIA_DELIVERY_EXTS contract at gateway/platforms/base.py:1432-1474 and made extract_media() consume it at gateway/platforms/base.py:3618-3622.
Problems
- Current main still omits
.logand.tomlfromMEDIA_DELIVERY_EXTS(gateway/platforms/base.py:1432-1449), although both are recognized document types atgateway/platforms/base.py:1338-1343. The remaining bug is therefore narrower than the PR's original list. - The PR's literal-regex change cannot be salvaged directly without bypassing the shared source of truth introduced by
781604ce4. - The bundled QQBot media/reconnect work needs separate re-evaluation: current main already has a chunked native-media implementation in
gateway/platforms/qqbot/adapter.py:2840-2914.
Suggested changes
- Re-scope the extraction fix to add
.logand.tomltoMEDIA_DELIVERY_EXTS, with parameterized regression coverage intests/gateway/test_platform_base.py.
Automated hermes-sweeper review.
| # and quoted/backticked paths for LLM-formatted outputs. | ||
| media_pattern = re.compile( | ||
| r'''[`"']?MEDIA:\s*(?P<path>`[^`\n]+`|"[^"\n]+"|'[^'\n]+'|(?:~/|/)\S+(?:[^\S\n]+\S+)*?\.(?:png|jpe?g|gif|webp|mp4|mov|avi|mkv|webm|ogg|opus|mp3|wav|m4a|flac|epub|pdf|zip|rar|7z|docx?|xlsx?|pptx?|txt|csv|apk|ipa)(?=[\s`"',;:)\]}]|$)|\S+)[`"']?''' | ||
| r'''[`"']?MEDIA:\s*(?P<path>`[^`\n]+`|"[^"\n]+"|'[^'\n]+'|(?:~/|/)\S+(?:[^\S\n]+\S+)*?\.(?:png|jpe?g|gif|webp|mp4|mov|avi|mkv|webm|ogg|opus|mp3|wav|m4a|flac|epub|pdf|zip|rar|7z|docx?|xlsx?|pptx?|txt|csv|apk|ipa|md|json|xml|html?|log|yaml|yml|toml)(?=[\s`"',;:)\]}]|$)|\S+)[`"']?''' |
Contributor
There was a problem hiding this comment.
Current main replaced this literal alternation with the shared MEDIA_DELIVERY_EXTS contract (gateway/platforms/base.py:1432-1474, commit 781604ce4). Please re-scope the salvage to add the remaining .log and .toml entries there, rather than restoring a second extension list.
teknium1
added
sweeper:risk-session-state
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
MEDIA: tag extraction silently ignored .md, .json, .xml, .html, .log, .yaml, .yml, and .toml files because the extension allowlist in BasePlatformAdapter.extract_media() didn't include these formats.
This prevented the gateway's _deliver_media_tags_streaming() and the send_message tool from delivering document files produced by agents (e.g. markdown reports, logs, config files) via QQBot and other platforms that use send_document().
Root Cause
The regex in gateway/platforms/base.py:1928 had a hardcoded extension alternation that only matched:
png|jpe?g|gif|webp|mp4|mov|avi|mkv|webm|ogg|opus|mp3|wav|m4a|flac|
epub|pdf|zip|rar|7z|docx?|xlsx?|pptx?|txt|csv|apk|ipa
Common document/code/text formats were missing - most notably .md (markdown), which Hermes agents frequently generate.
Fix
Added md|json|xml|html?|log|yaml|yml|toml to the regex extension alternation (1 line change in gateway/platforms/base.py).
Test Results
Existing tests (no regression)
Manual verification
All new extensions extracted correctly:
Existing extensions still work:
Impact
Low risk - purely additive: no existing extensions removed, no logic changed.