fanlusky
a7779d2b05
`vendor/MetasequoiaImeEngine` submodule 改为 `engine/` 目录，导入的是 product-lock.json 原本钉住的提交 c810d201，逐文件核对过 blob，因此这是一次 搬迁而不是引擎升级。引擎自己的嵌套 submodule（googlepinyinime-rev、utfcpp、 miniaudio）按各自 gitlink 一并展开，来源与提交见 engine/UPSTREAM.md。 Windows 专门化裁掉了不属于本产品的部分：whisper.cpp 与本地推理 provider、 macOS 示例宿主、整个 dictionary/ 建库流水线、上游 CI 与 release-please 配置。 词库仍按 product-lock.json 在构建时取发布产物。 引擎不再进 product-lock.json：没有 gitlink 就没有可比对的东西，锁里写一个上游 commit 等于宣称这棵树仍等于上游，而第一个 Windows 专门化改动落地时这句话就假了。 verify-contracts / verify-checkout 随之移除，verify-published 保留。 CI 与工具链：引擎源码并入根 scripts/format.sh（删掉引擎自己的那份，避免两套排除 列表分叉）；引擎 voice.yml 里的 silero_vad.onnx 摘要校验移到本仓 lock job； engine-bump-triage 工作流与其测试删除；gitsubmodule 的 dependabot 频率由 daily 改为 monthly。 README、README.en、AGENTS.md、CONTRIBUTING、docs/product-release.md 及各子目录 文档同步说明新的位置与边界。
Dictionary format v1
format.json is the source of truth for Quanpin table naming and the data format version.
generate.py emits the C++ header; format.py reads the same JSON for dictionary builders.
The engine's query and journal replay use this header, and Server settings writes through the
engine's public quanpin::build_table_name. The out-of-tree dictionary builder consumes the
Python API from this directory, including table enumeration for creation and indexing.
The numbered buckets stop at seven syllables. Eight and more use tbl_others_{initial}.
The shipped table set uses the listed initial letters; query construction may also receive
an unmatched lowercase initial and return an empty result from the dictionary. Malformed
or absent initial characters cannot become SQL identifiers.
Change the format version when changing the on-disk layout. Keep compatibility with the format range declared by dictionary product manifests. Run the Engine input-session tests: they create seven/eight/nine-syllable words, query them and replay their journal into a fresh database. CI also rejects stale generated headers.