Files
msime-windows/engine/UPSTREAM.md
T
Lu Fan(SonnyCalcr) d912a926b7 fix(engine): 修复谷歌整句解码器去重越界写导致双拼 zjle 时 Server 崩溃 (#1064)
zan'l 的末尾半音节让「赞了」经 zan'le / zan'liao 各出一次。get_lpis 多字分支
先跳过等于整句候选的第一份,再遇到分数更好的重复项时 remain_num 仍为 0,
写到 lma_buf[-1]。改为先丢掉所有等于整句候选的项再去重。

Fixes #1063
2026-10-18 21:34:56 +08:00

8.1 KiB

Upstream provenance

This directory was vendor/MetasequoiaImeEngine, a submodule of https://github.com/metasequoiaime/MSIME-Engine. It is now a component of this repository, vendored in-tree at engine/ alongside server/, ui/, windows/ and installer/.

Upstream repository metasequoiaime/MSIME-Engine
Imported commit c810d201f549b337ae0c4a65a9d694103f1c1754
Engine version at import 0.9.0 (version.txt)

That commit is the one product-lock.json pinned before the import, so this is a relocation and not an engine upgrade: every source file here was verified against that commit's blob after the move.

The engine's own nested submodules were expanded in place at the commits it pinned:

Path Upstream Commit
googlepinyinime-rev/ metasequoiaime/Google-PinyinIME-Rev 12db5237adfcecb79b8ac602d80f3576639ea219
utfcpp/ nemtrif/utfcpp 2d8e20b22dcb3e9b3c4f52103182ebda949c6089
ngram/kenlm/ kpu/kenlm 4cb443e60b7bf2c0ddf3c745378f76cb59e254e5
ngram/octagram/darts.h s-yata/darts-clone 87b71afd6cf784953e3c08f24c64203397f3b724 (tag v0.32h)
voice/third_party/miniaudio/ mackron/miniaudio 9634bedb5b5a2ca38c1ee7108a9358a4e233f14d

ngram/octagram/darts.h is the only file of that copy in the tree — upstream ships it as include/darts.h next to COPYING.md, both of which are renamed on import to sit beside the octagram_gram reader that is ours. It is a header-only double-array trie: octagram_gram uses set_array to adopt the mmap of a .gram file, then traverse and commonPrefixSearch. None of that is registered with darts-clone upstream, so it is tracked by tag rather than by branch.

This copy is patched, not verbatim. Replaying an upgrade means re-applying all of these against the new tag:

  • DoubleArrayImpl::open() drops upstream's 256-unit header check — the size < 256 || (size & 0xFF) guard, the units[0] validation, and the two-stage read that copies those units into buf before reading the remainder — and reads the array in a single fread.
  • DoubleArrayImpl::save() drops its offset parameter and the fseek that used it, becoming save(const char *, const char *, std::size_t) const.
  • commonLongestPrefixSearch is deleted: both overloads declared on the class and the out-of-line template definition. Only commonPrefixSearch and traverse are called by octagram_gram.
  • DoubleArrayBuilder's BLOCK_SIZE / NUM_EXTRA_BLOCKS / NUM_EXTRAS change from an unnamed enum to static const std::size_t.
  • A label() comment typo is fixed (associted → associated).

The open() and save() rewrites follow from the same decision as the rest of this reader: loading goes through Win32 mmap, so the file-opening and file-writing entry points are dead here. Verify the current state against the tag before assuming any of them still applies.

ngram/kenlm/ did not arrive with the engine import; it was added later, at the commit libime pins for its own src/libime/core/kenlm submodule, so the query code and the shipped sc.lm come from the same generation of kenlm. Only the query subset was copied: lm/*.{hh,cc}, util/ and util/double-conversion/. lm/builder, lm/filter, lm/interpolate and lm/wrappers are not here, and no file that is here includes them. The sources are unmodified except for the one local patch below; the two Windows adaptations it needs — a wide-character util::_open and the _HAS_AUTO_PTR_ETC definition — live outside the subtree, in ngram/kenlm_file_io.h and ngram/CMakeLists.txt.

Local patch: util/string_stream.hh StringStream::AdvanceTo — &*out_.end() and &*out_.begin() dereference string iterators, which the MSVC checked iterators reject (a "Debug Error!" assert under any _DEBUG build, on every formatted kenlm message). out_.data() + out_.size() computes the same past-the-end pointer without touching iterators; release behaviour is unchanged because the asserts were compiled out there anyway.

Two of those files, googlepinyinime-rev/src/{include/userdict.h,share/userdict.cpp}, are committed upstream with CRLF and arrive here with LF. That is this repository's .gitattributes (* text=auto eol=lf) doing what it does to every other file; the content is unchanged.

googlepinyinime-rev/ is no longer tracked against its upstream and is edited here directly, in its original formatting and under its original licence. Local changes so far: a per-syllable Hanzi constraint for the shuangpin mid-sentence helpcode — MatrixSearch::set_char_constraints, MatrixSearch::filter_lpis_by_constraints (called from add_char_qwerty before extend_mtrx_nd), im_set_char_constraints and the CharConstraintFn typedef in dictdef.h; and in the multi-syllable branch of MatrixSearch::get_lpis, items equal to the full sentence candidate are dropped before deduplication, so a better-scored duplicate of a dropped item no longer writes lma_buf[-1] (crashed the Server on shuangpin zjle, i.e. zan'l).

Why this file and not product-lock.json

The engine used to be pinned by a gitlink, and product-lock.json recorded the commit so CI could check that the checked-out submodule matched the reviewed one. There is no gitlink any more, so there is nothing left for CI to compare against: the first Windows-specific fix to engine/ makes "this tree is upstream commit X" false, and nothing would notice. Recording the provenance in prose says what is true — where the code came from — without claiming an equality that no longer holds.

Treat engine/ as first-party code from here on. Changes to it are reviewed, formatted and tested like any other directory of this repository (scripts/format.sh covers it), not merged from upstream. Upstream is history, not a remote to track.

What was removed during the Windows specialization

Everything below was deleted from the imported tree. No engine source file was otherwise modified. voice/CMakeLists.txt lost the option blocks belonging to the deleted sources; the rest of the changes are documentation (README*.md, AGENTS.md, NOTICE.md, the contracts/ and docs/ prose) rewritten to describe a directory rather than a repository.

Non-Windows platform code

  • voice/third_party/whisper.cpp/ and the whisper provider (worker, header, test) — roughly 36 MB of local-inference ASR. This product's voice input uses cloud ASR; nothing here built or shipped it.
  • voice/examples/macos/ — a standalone macOS host application.

Portable if(APPLE) / elseif(WIN32) / else() guards inside the surviving sources were left alone. Stripping them would be a large, untested rewrite of working code for no build-output gain.

The dictionary build pipeline

  • dictionary/ in full — about 228 MB of raw corpora (cn, en, source, makecikudb) and the scripts that turn them into a dictionary release. No CMake target in this repository consumes them, and the workflow that built them lived in the engine's own CI. The built dictionaries are still fetched at build time from their release, as product-lock.json records.
  • build_assets.py and build_profile.py, which import dictionary.build_profile.

contracts/assets/ was kept: assets.h is a generated header included by roughly twenty engine translation units, and it is unrelated to the corpora despite the name.

Upstream governance and tooling

  • .github/ (the engine's own CI), .vscode/, .gitmodules, .mailmap, .git-blame-ignore-revs, CHANGELOG.md, release-please-config.json, .release-please-manifest.json — all of these describe a standalone repository with its own release train, which this directory no longer is.
  • scripts/format.sh — this repository formats the engine from scripts/format.sh at the root, so a second entry point with its own exclusion list would only be a way for the two to disagree. The exclusions it carried were folded into the root script.
  • Unused portions of the vendored miniaudio, googlepinyinime-rev and utfcpp copies. What remains of those three is still third-party and still under its own licence; see NOTICE.md.