emrg: improve tool error handling (glob + edit) - #1
Merged
Conversation
- glob_tool: error on nonexistent workdir instead of silent CWD fallback - glob_tool: remove unused import os - edit_tool: add OSError handling for write_text, matching write_tool
argszero
commented
Jul 18, 2026
argszero
commented
Jul 18, 2026
argszero
commented
Jul 18, 2026
This was referenced Jul 18, 2026
Merged
This was referenced Jul 27, 2026
argszero added a commit
that referenced
this pull request
Jul 28, 2026
Remove 3 set_title call sites (tool_start, first-delta, streaming), unify idle states.
This was referenced Jul 29, 2026
argszero added a commit
that referenced
this pull request
Aug 6, 2026
…07-27） (#429) 对照 .emrg/memory/paper-prompt-remaining-gaps.md 的 5 项缺陷修复 4 项 （#1 记忆路径已由状态文件机制解决，无需改动）： 1. **LaTeX 编译必然失败** → 编译前 which latexmk 检查，不可用则跳过编译 只做文本级交叉引用检查 2. **文献检索无去重** → 检索前先 ls literature/ 已有文献，笔记已存在则跳过 3. **arXiv 关键词与研究方向脱节** → 禁止通用大词 all:large+language+model， 改为从 Agent.md/摘要/状态文件派生方向术语构造靶向查询 4. **seq 重置误导** → 当前状态区注明 daemon 重启后序号重置，判断首次运行 以状态文件为准 新增 tests/test_scheduler.py 模板渲染回归测试（paper_prompt.md 可正常渲染 且含状态文件/latexmk/literature 指引）。 验证：pytest 473 绿（+1 新测试） Co-authored-by: EMRG Evolution <emrg@argszero.dev>
This was referenced Aug 7, 2026
argszero added a commit
that referenced
this pull request
Aug 14, 2026
…795) Rant 2026-08-14T22:13:57: promote_prompt.md §0 only read the state file + project config — the promoter had zero knowledge of the actual project source/latest progress, while open-source tasks (#790) must study the full codebase. Asymmetric: same EMRG contributes after deep code study but promotes from a one-line description → stale or wrong answers to community questions (violating red lines #1/#7). Changes: - promote_prompt.md new §0.4 'Learn the project's latest state (MUST every round)': git fetch + last 10 commits, read repo root, scan key modules, refresh cognition; any capability claim must come from just-verified latest state (mirrors open_source B.2b) - §3 Step 2: Good-example note — feature descriptions must come from §0.4-verified state, not stale memory/description guesses - §4 state file: new 'last learned' field (time + commit HEAD) for knowledge freshness - §5 reflection Q3: must record what project info was learned this round (commit range / modules read) - +1 template render test (7 positive assertions); Agent.md 810->811 Co-authored-by: EMRG Evolution <emrg@argszero.dev>
This was referenced Aug 20, 2026
argszero added a commit
that referenced
this pull request
Aug 26, 2026
…ds/copywriting) + preload API-surface guard (#992) Renderer migration Batch 1 (design doc v0.1 §5, rant 2026-08-26T12:16:48, D1-D7 decided). Follows #991 (Batch 0) now merged. - renderer/src/lib/commands.ts: vanilla renderer/js/commands.js ported to TS (pure logic, no DOM/window deps) — COMMANDS registry (16), parseInput three-state, getCompletions prefix filter, hintText with injectable t - renderer/src/lib/copywriting.ts: vanilla renderer/js/copywriting.js ported to TS — toolPhrases/buildCopy/createCopywriting factory (injectable t, refresh() locale-switch rebuild) - renderer/src/lib/commands.test.ts + copywriting.test.ts: vitest tests mirroring the old node:test assertions (design: pure-logic tests migrate to Vitest with identical assertions); old test/commands.test.js stays until Batch 5 switch (D3/D6) - emrg/gui/test/preload-api.test.js: window.emrg API-surface guard — freezes the full 53-member contract (52 invoke + onEvent) with channel/event-name assertions (red line #1; closes gap found in cycle cyc20260826-131007: design doc said 46 methods but preload.js actually exposes 52 + onEvent, only spot checks existed before) - Agent.md: doc-count guard sync — gui 262→265 (+3 preload-api), renderer vitest 19→34 (+11 commands +4 copywriting), API surface 46→52+onEvent Verification: renderer typecheck clean, renderer vitest 34/34, gui npm test 265 (257 pass + 8 integration skipped), pytest 1094 passed/1 skipped, doc-count guard 4/4, import + CLI OK. Co-authored-by: EMRG Evolution <emrg@argszero.dev>
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 freeto join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
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.
Changes
glob_tool
import os: cleanup.edit_tool
path.write_text()now wrapped in try/except OSError to return clean error (disk full, permissions), consistent with write_tool.