DeepSeek/deepseek-harness
1
0
Fork 0
mirror of https://github.com/deepseek-ai/deepseek-harness synced 2026-08-15 21:04:50 +00:00
Code Issues Packages Projects Releases Wiki Activity

Merge branch 'worktree/ci-native-windows-20260808' into worktree/ci-native-windows-coverage-20260808

Browse Source 
# Conflicts:
#	packages/subagent/subagent-codex/README.i18n.yaml
#	packages/subagent/subagent-codex/README.md
#	packages/subagent/subagent-codex/README.zh.md
This commit is contained in:
Tianyi Cui
2026-08-09 21:52:02 +08:00
parent 3fb06728be bf38526354
commit 525ea43f0b
783 changed files with 1803 additions and 1784 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
packages/session-query/session-query-sqlite/README.i18n.yaml
View File

@@ -2,5 +2,5 @@
# side as of the last confirmed-consistent state. Both languages carry equal authority;
# after editing either side, bring the other along and re-record with:
# pnpm run verify-translation-pairing --write packages/session-query/session-query-sqlite/README.md
README.md: dbbb63e9dd18f3c68cc21b84b560403f4a508709
README.zh.md: 0b183f63ad2f3cfa670babca80d6e4abee6a4bcb
README.md: 8784f1b2c2a3c9291b761c0a6e68201286b9e306
README.zh.md: 5b69c054a0e17a49e9c709b27e497b006a024a91

2
packages/session-query/session-query-sqlite/README.md
View File

@@ -37,7 +37,7 @@ The database is disposable but reset is guarded: every recognized schema version
## Tokenizer and limits
The index uses FTS5 `unicode61`. In the implementation experiment it supported the two-character query `AI` and produced an index about 2.1× smaller than the trigram alternative. The trade-off is token/phrase recall rather than arbitrary substring recall: `AI` does not match the token `BRAID`. Use `ctx.sessionQuery.filterEvents()` with a `text` clause when a literal whitespace-flexible substring scan is required. NUL is rejected in queries; reserved highlight markers and NUL in documents are normalized before indexing so presentation markers cannot collide with source text.
The index uses FTS5 `unicode61`. The trade-off is token/phrase recall rather than arbitrary substring recall: `AI` does not match the token `BRAID`. Use `ctx.sessionQuery.filterEvents()` with a `text` clause when a literal whitespace-flexible substring scan is required. NUL is rejected in queries; reserved highlight markers and NUL in documents are normalized before indexing so presentation markers cannot collide with source text.
Abort signals stop queued work and flow unchanged through snapshot listing and non-mutating inspection. Once source work starts, the serialized state machine awaits that backend promise itself—even when a backend ignores cancellation—then checks the signal before starting any further listing, inspection, reconciliation, or query work. The caller therefore observes cancellation only after started backend work is quiescent, and a later search cannot enter the serializer while that cleanup is pending. Node's synchronous `DatabaseSync` API cannot interrupt a metadata or MATCH statement already executing on the JavaScript thread; signals are checked immediately before and after those non-preemptible calls.

2
packages/session-query/session-query-sqlite/README.zh.md
View File

@@ -37,7 +37,7 @@
## 分词器与限制
该索引使用 FTS5 `unicode61`。在实现实验中，它支持双字符查询 `AI`，产生的索引比 trigram 备选方案小约 2.1 倍。取舍是 token/短语召回而非任意子字符串召回：`AI` 不匹配 token `BRAID`。需要执行字面的空白弹性子字符串扫描时，使用 `ctx.sessionQuery.filterEvents()` 并传入 `text` 子句。查询会拒绝 NUL；文档中的保留高亮标记和 NUL 会在索引前被规范化，使展示标记无法与源文本冲突。
该索引使用 FTS5 `unicode61`。取舍是 token/短语召回而非任意子字符串召回：`AI` 不匹配 token `BRAID`。需要执行字面的空白弹性子字符串扫描时，使用 `ctx.sessionQuery.filterEvents()` 并传入 `text` 子句。查询会拒绝 NUL；文档中的保留高亮标记和 NUL 会在索引前被规范化，使展示标记无法与源文本冲突。
中止信号会停止已排队工作，并原样流经快照枚举和非修改式检查。来源工作一旦开始，串行化状态机会自行等待该后端 promise，即使后端忽略取消，之后也会在启动任何进一步的枚举、检查、对账或查询工作前检查信号。因此，调用方只会在已启动后端工作完全停稳后观察到取消，而后续搜索在该清理尚未完成时无法进入 serializer。Node 的同步 `DatabaseSync` API 无法中断已在 JavaScript 线程上执行的元数据或 MATCH 语句；系统会在这些不可抢占调用前后立即检查信号。