fix: guard _parse_cooldown against malformed repeat values; close file handle with context manager#76
Merged
lsdefine merged 1 commit intolsdefine:mainfrom Apr 16, 2026
Conversation
…e handle with context manager (Closes lsdefine#75) Signed-off-by: Cocoon-Break <54054995+kuishou68@users.noreply.github.com>
shaun0927
added a commit
to shaun0927/GenericAgent
that referenced
this pull request
Apr 17, 2026
After PR lsdefine#76 wrapped _parse_cooldown in try/except, parse('every_0h') still returns timedelta(0). The cooldown check at check() then always fires the task on the next 120s scheduler cycle, in effect running a side-effectful task continuously instead of every N hours. Add a positive-interval validation inside the existing try block so the malformed case falls through to the existing 'fallback to 20h cooldown' warning path, the same way 'every_h' or 'every_xyz' do. Builds on the guard introduced by PR lsdefine#76.
This was referenced Apr 17, 2026
shaun0927
added a commit
to shaun0927/GenericAgent
that referenced
this pull request
Apr 17, 2026
Closes lsdefine#96. PR lsdefine#76 already converted the scheduler.py read site to a context manager. Apply the same treatment to the remaining unmanaged open() calls that this audit found across ga.py, agentmain.py, and frontends/wechatapp.py. Sites updated: ga.py - 23 : code_run header copy into the temp script - 387 : prepend-mode read of existing file in do_file_write - 388 : prepend-mode write of new+old back to disk - 424 : _check_plan_completion read of plan.md agentmain.py - 16 : tools_schema*.json load - 24 : global_mem.txt initial seed - 27-28: global_mem_insight.txt initial seed (was nested open(...).write(open(...).read()), so two leaks) - 33 : tmwebdriver cdp_bridge config.js seed - 105 : /session.<key> = <file> case (refactor 47f106c kept the unmanaged read) - 242 : reflect log append frontends/wechatapp.py - 183 : decrypted media bytes write to temp dir These match the pattern in lsdefine#48 ("multiple calls then model unresponsive, needs restart") for long-running adapters where GC of unreferenced file objects is delayed and FD limits matter. The launcher / startup-only sites are still affected because the same handler is reused in reflect-loop code paths. No behavior change: all reads/writes use the same modes, encodings, and bytes; only the file handle release timing changes.
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
2 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.
Description
Fixes two bugs in
reflect/scheduler.py:Fix 1 — Prevent
ValueError/IndexErrorin_parse_cooldownWhen a task JSON has a
repeatvalue starting with"every_"but malformed (e.g."every_","every_h"), the code crashed with an unhandled exception. Added atry/except (ValueError, IndexError)guard around the parsing block so it falls through to the existing warning+fallback instead of crashing the scheduler.Fix 2 — Use context manager to close file handle
Replaced bare
open().read()on line 81 with awithblock to guarantee the file handle is closed even on exception, preventing potential resource leaks.Closes #75
Signed-off-by: Cocoon-Break 54054995+kuishou68@users.noreply.github.com