feat: add 4 deep-dive comparisons (instruction/MDM/telemetry/token)#33
Merged
feat: add 4 deep-dive comparisons (instruction/MDM/telemetry/token)#33
Conversation
Four new comparison articles based on Claude Code v2.1.89 and Qwen Code v0.15.0: 1. instruction-loading-deep-dive.md (274 lines) - CLAUDE.md 6-layer hierarchy vs QWEN.md 3-layer - @include directives (both support, 5-level nesting) - Frontmatter path filtering (Claude only) - Conditional rules, HTML comment stripping, auto-memory 2. mdm-enterprise-deep-dive.md (166 lines) - Claude: macOS plist + Windows Registry + drop-in dir + remote API - Qwen: file-based only, no MDM - 5-level First-Source-Wins policy precedence 3. telemetry-architecture-deep-dive.md (181 lines) - Claude: 505 tengu_ events + 1P Logger + Perfetto tracing - Qwen: 50 events + QwenLogger RUM + OTLP - Privacy controls comparison 4. token-estimation-deep-dive.md (231 lines) - Claude: API real-time counting + VCR cache + thinking budget - Qwen: static pattern matching (82 models) + effort levels - Natural language budget parsing ("+500k", "spend 2M") Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Round 1 (source verification): - claudemd.ts: ~2,300 → 1,479 lines (actual wc -l) - memoryDiscovery.ts: ~217 → 357 lines - memoryTool.ts: ~113 → 543 lines - tengu_ event count: ~505 → ~656 (grep -roh unique count) Round 3 (cross-doc consistency): - Reconcile telemetry count with 03-architecture.md (782 = binary extraction including dynamic events; 656 = source grep unique strings) - Added note explaining both counting methods Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
There was a problem hiding this comment.
Pull request overview
This PR adds four new deep-dive comparison articles to the documentation, focusing on how Claude Code vs Qwen Code differ across instruction loading, enterprise/MDM configuration, telemetry architecture, and token estimation/thinking support.
Changes:
- Added a deep-dive on token estimation strategies and “thinking” budget handling across providers.
- Added a deep-dive on telemetry architecture, event taxonomy, and privacy controls.
- Added deep-dives on enterprise policy management (MDM) and instruction file discovery/include mechanisms.
Reviewed changes
Copilot reviewed 4 out of 4 changed files in this pull request and generated 6 comments.
| File | Description |
|---|---|
| docs/comparison/token-estimation-deep-dive.md | New article comparing token counting/estimation, thinking configuration, and provider adaptations. |
| docs/comparison/telemetry-architecture-deep-dive.md | New article comparing telemetry pipelines, event volume, sampling, and privacy controls. |
| docs/comparison/mdm-enterprise-deep-dive.md | New article comparing enterprise policy/MDM configuration hierarchies and remote settings handling. |
| docs/comparison/instruction-loading-deep-dive.md | New article comparing instruction file discovery layers, @include, frontmatter path filtering, and trust models. |
💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.
Comment on lines
+70
to
+72
| | `@/path` | 绝对路径 | | ||
| | `@path#fragment` | 自动剥离 `#` 后缀 | | ||
| | `@path\ with\ spaces` | 反斜杠转义空格 | |
There was a problem hiding this comment.
表格里的路径示例 @path\\ with\\ spaces 在 Markdown inline code 中不需要转义反斜杠，当前会渲染成双反斜杠，导致示例与实际语法不一致。建议改为单反斜杠形式。
| - `@include` 不在以下区域内解析： | ||
| - HTML 注释 `<!-- ... -->` | ||
| - 围栏代码块 ` ```...``` ` | ||
| - 行内代码 `` `...` `` |
There was a problem hiding this comment.
“围栏代码块 ```...```” 这里用单反引号包裹包含反引号的内容，容易导致 Markdown 解析失败/渲染错乱。建议改用更长的 inline code 定界符（例如双反引号包裹）或用 HTML <code> 来展示三反引号示例。
Suggested change
| - 行内代码 `` `...` `` | |
| - 行内代码 <code>`...`</code> |
Comment on lines
+11
to
+13
| | **macOS plist** | ✅ `com.anthropic.claudecode` domain | ❌ | | ||
| | **Windows Registry** | ✅ `HKLM\SOFTWARE\Policies\ClaudeCode` | ❌ | | ||
| | **Linux 文件策略** | ✅ `/etc/claude-code/managed-settings.json` | ❌ | |
There was a problem hiding this comment.
Windows Registry 路径在 inline code 中写成了 HKLM\\SOFTWARE\\Policies\\ClaudeCode。Markdown 的 code span 不需要对反斜杠做转义，这会实际显示为双反斜杠，影响可读性/准确性；建议改为单反斜杠形式。
Comment on lines
+61
to
+64
| // 源码: utils/settings/mdm/constants.ts#L23-L26 | ||
| // HKLM\SOFTWARE\Policies\ClaudeCode （管理员策略，最高） | ||
| // HKCU\SOFTWARE\Policies\ClaudeCode （用户级策略，最低） | ||
| // 注: SOFTWARE\Policies 在 WOW64 共享——32/64 位无重定向 |
There was a problem hiding this comment.
本节代码块同样使用了双反斜杠（例如 HKLM\\SOFTWARE\\Policies\\ClaudeCode、SOFTWARE\\Policies）。在 fenced code block 中也无需转义反斜杠，建议统一改为单反斜杠，避免读者复制时出错。
Comment on lines
+72
to
+75
| ``` | ||
| macOS: /Library/Application Support/ClaudeCode/managed-settings.d/ | ||
| Windows: C:\Program Files\ClaudeCode\managed-settings.d/ | ||
| Linux: /etc/claude-code/managed-settings.d/ |
There was a problem hiding this comment.
Drop-in 目录示例里的 Windows 路径写成了 C:\\Program Files\\ClaudeCode\\...。在普通代码块里反斜杠不需要转义，当前会显示为双反斜杠；建议改为单反斜杠以便直接复制使用。
Comment on lines
+11
to
+15
| | **遥测框架** | 自定义 1P Logger + OpenTelemetry | QwenLogger（RUM）+ OpenTelemetry | | ||
| | **上报端点** | Anthropic 内部 API + Datadog | 阿里云 RUM + OTLP（可配置） | | ||
| | **事件数量** | ~656 个 `tengu_*` 事件 | ~50 个事件类型 | | ||
| | **采样策略** | 按事件类型动态采样（GrowthBook） | 批量刷新（1,000 条 / 60 秒） | | ||
| | **调试追踪** | Perfetto（Chrome Trace，ant-only） | 标准 OTLP Span | |
There was a problem hiding this comment.
这里的 tengu_* 事件数量写为 ~656，但 PR 描述的摘要表里写的是 “505 tengu_ events”。建议统一 PR 描述与文档中的统计口径/数字（例如是否包含动态构造事件名），避免读者困惑。
wenshao
added a commit
that referenced
this pull request
Apr 5, 2026
Closed PR #72 (qwen3.6-plus) — 22 of 26 items duplicated existing 183. Extracted 4 new items to p2-tools: - #31 /plan (plan mode) - #32 /rename (session rename) - #33 /upgrade (version upgrade + changelog) - #34 Plugin system (aggregated container for commands+skills+hooks+MCP) Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
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
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
Four new comparison articles (852 lines total) based on source code analysis:
Test plan
🤖 Generated with Claude Code