50cf424146
The build no longer self-wraps into `nix develop`: src/tools/xce-build.sh now runs meson+ninja directly on system packages, with a dependency preflight that fails early naming the missing tool/pkg-config package (DEPS.md). Dependencies (see DEPS.md for the full tables): - 11 C libraries from apt (libuv, openssl, libedit, sqlite3, hwloc, libcgroup, libtirpc, libcyaml, libyaml, zlib, libsystemd) with version floors in meson.build mirroring the Ubuntu 24.04 baseline. - sd-bus: default libsystemd, -Duse_elogind=true opt-back for nix-era hosts. - 6 libraries/tools unavailable in apt become committed meson wrap subprojects under src/subprojects/ (source closure, ~500 files): the upstream trees plus packagefiles/ overlays and diff_files patches registering dependency/program overrides the nix .pc files used to provide (incl. the '<name>-raw' variable deps the Go CGO path reads). Fresh-clone reproducibility verified: wiped trees + `meson subprojects download` reproduce the committed state exactly (fixed librtti.wrap's dangling patch_directory and c-capnproto's empty packagefiles overlay). - xceadmin: self-contained venv built from requirements-build.txt (pinned, offline-seedable via PIP_FIND_LINKS) instead of --system-site-packages. Build/packaging: - Makefile facade over xce-build.sh; scripts/ci-ubuntu.sh drives a bare Ubuntu 24.04 container through deps + build + tests. - src/tools/xce-package.sh: release-tree tarball packager. Collects non-base DT_NEEDED libs from the BUILD-tree binaries (staged ldd lies: the dangling build RPATH makes ldd drop the very lines the naive filter needs), sets $ORIGIN RPATHs, rebases the xceadmin venv at install time (kernel does not parse quotes in shebangs), ships systemd unit samples + install.sh. Local smoke test: all 8 binaries resolve, xceadmin works from both bin/ wrapper and venv script. Test fixes found on the way (all environmental, not build bugs): - harness_test.c setUp(): pin XCE_CONTROLLER_PEERS to a bound-then-closed loopback port so the CLI-error tests assert the no-controller failure path even on a dev box running a live cluster on localhost:9990. - test_test_cli.c: setpgid(0,0) in the spawned-xce_test child. libedit's sig_handler re-raises received signals via kill(0, sig) -- process GROUP wide (strace -k evidence) -- and the child previously shared the harness's pgrp, so an interrupted SSL handshake killed the whole suite (exit 130). Also fixed while re-baselining on the de-nix toolchain (pre-existing, now deterministic): raft_init zeroes pre-vote tally state (pre_vote_in_progress/ pre_votes_granted/pre_vote_replies were left at stack garbage); push_config_to_standbys clears stale config acks before fan-out (repl_reset_config_acks) -- SHA-256-derived tags are not monotonic, so a previous push's ack numerically above the new tag fabricated a barrier quorum; capnpc-c's generated encode_* memsets the destination first (skipped Data fields fed garbage into capn_setp). Validation: asan profile 161/161 meson tests green (one known p2_integration -110 timing flake on first run, green on re-run); debug + release full builds OK; tarball smoke-tested on the build host. Cluster rollout on t1/t2/t3 follows. Co-Authored-By: Claude Code <noreply@anthropic.com>
5.2 KiB
5.2 KiB
XCE 第三方依赖清单
建立：2026-10-20（build/de-nix 分支）。背景与决策见 gaps.md "六、构建/分发系统"。
原则：系统库优先（apt/发行版包），发行版没有的用 meson wrap subproject 从源码构建； 运行期需要自带的库由打包脚本 bundle（
$ORIGINRPATH），不引入 /nix/store 或任何 内容寻址二进制闭包。glibc/libm/libpthread/pthread 永不 bundle。
构建工具链
| 工具 | 最低版本 | 获取 | 备注 |
|---|---|---|---|
| gcc | 13 | 系统包 | apt install gcc（glibc 2.39+ 基线） |
| meson | 1.5 | pip install meson / apt install meson(≥1.3 时 pip 补) |
src/meson.build 要求 >=1.5 |
| ninja | 1.11 | apt install ninja-build |
meson 默认 backend |
| pkg-config | 0.29 | apt install pkg-config |
|
| go | 1.22 | 上游 tarball / apt install golang-go |
apiserver/xcecli 已 vendor 全部模块，构建离线 |
| capnp (capnproto) | 1.0.1 | apt install capnproto（含 libcapnp-dev） |
schema 编译器 + C++ 运行时（pycapnp 链接用） |
| packcc | 2.2 | wrap subproject（Ubuntu 24.04 无包） | 3 个 .peg 解析器（lib/section, rusage, xce_resreq, xce_select） |
| python3 | 3.10 | 系统自带 | xceadmin 构建/测试；no-fork rule 检查脚本 |
| swag | latest | go install github.com/swaggo/swag/cmd/swag@latest |
仅 apiserver 文档再生成时需要，非构建必需 |
C 库依赖
apt 可直接安装（11 个）
| 依赖 (meson 名) | apt 包 | 版本下限 | 用途 |
|---|---|---|---|
| libuv | libuv1-dev | >=1.44 | 所有 daemon 事件循环 |
| openssl | libssl-dev | >=3.0 | TLS（controller 互联、apiserver）、SHA256 |
| libedit | libedit-dev | >=3.1 | xce_test REPL（BSD 许可，替代 GPL readline） |
| sqlite3 | libsqlite3-dev | >=3.37 | worker 本地状态 |
| hwloc | libhwloc-dev | >=2.8 | worker 拓扑/槽位 |
| libcgroup | libcgroup-dev | >=2.0 | worker cgroup v2 限额 |
| libtirpc | libtirpc-dev | >=1.3 | rpc 头文件 |
| libcyaml | libcyaml-dev | >=1.2 | YAML 配置读取 |
| yaml-0.1 | libyaml-dev | >=0.2.5 | 同上（libcyaml 依赖） |
| zlib | zlib1g-dev | >=1.2.11 | 压缩 |
| threads/math/rt | (libc) | — | 无需安装 |
apt 没有 → meson wrap subproject（源码随 repo 拉取构建，6 个）
| 依赖 (meson 名) | 上游 | 版本 | 用途 / 备注 |
|---|---|---|---|
| c-capnproto | github:shen390s/c-capnproto（pin e45ff68） | — | Cap'n Proto C runtime（libCapnC_Runtime）+ capnpc-c 代码生成器；32 个源文件 include capnp_c.h |
| librtti | git+ssh gitea rshen/librtti（pin 2149f778） | 0.0.2 | C 运行时类型信息库；上游本身是零依赖 meson 项目；93 个源文件消费 |
| struct2json | xbuild channel 源码 vendor | 1.0.0 | cJSON + s2j 封装（纯静态库）；6 个 meson 消费点 |
| tinylog | xbuild channel 源码 vendor | 1.0.0 | 轻量异步日志（-ltlog + tlog.h）；13 个源文件消费 |
| unity | xbuild channel 源码 vendor（上游 ThrowTheSwitch/Unity） | 2.6.1 | 单元测试框架（libunity.a + unity/unity.h）；仅测试目标链接 |
| packcc | 上游 arithy/packcc | >=2.2 | PEG 解析器生成器；构建工具而非链接库 |
特殊：libelogind → libsystemd
worker 的 systemd transient unit 管理（sd_bus_*，263 处调用）目前链 nix 的 elogind 255.5。
Ubuntu 24.04 无 elogind 包，但有 libsystemd-dev 255.4——sd-bus API 同源兼容。
构建默认 dependency('libsystemd')，保留 -Duse_elogind=true 选项供非 systemd 平台。
（Step 2 实施时验证；这是本迁移唯一有 ABI 行为风险的替换点，worker 单测兜底。）
Python（xceadmin）
src/xceadmin/pyproject.toml 已声明全部运行依赖：click>=8.0, pycapnp>=1.3, PyYAML>=6.0,
jsonschema>=4.0。构建用普通 venv（不要 --system-site-packages——那是 nix 环境的产物）：
python3 -m venv .venv && .venv/bin/pip install .。pycapnp 的 C 扩展链系统的
capnproto C++ 运行时（apt capnproto/libcapnp-dev 提供版本 1.0.1；pycapnp>=1.3 兼容）。
为什么不用 nix store（决策记录）
nix 把依赖闭包做成内容寻址的 /nix/store 路径 + mount namespace，要求闭包出现在每一台
运行机器上，且 patchelf 重定位后 ldd/gdb/stack trace 全部变形（t2/t3 master 进程 argv 是
ld-linux... --library-path /opt/newxce/app/nix/store/...，人读不懂）。EDA 客户环境没有
/nix/store，这条路在产品定义上不可用。本 repo 依赖只有十来个中小型库，"自包含"的正确形态
是源码闭包（wrap subproject）+ 运行期 bundle 清单，不需要二进制 store。此外 nix 的
LD_LIBRARY_PATH 泄漏会导致系统工具崩溃（t2/t3 的 ls/head segfault 实录）。
运行期 bundle 规则（打包脚本用）
- bundle：libuv、libcapnp/libCapnC_Runtime、librtti、libtlog、libstruct2json、 libcyaml、libyaml、libtirpc、libedit、libcgroup、libhwloc、libssl/libcrypto（若目标基线 <3.0）
- 永不 bundle：libc/libm/libpthread/libdl/librt（glibc）、libsystemd（按目标发行版）
- 判定方法：对每个产物
ldd，凡路径不在 {/lib, /usr/lib, /lib64} 的（即 wrap 构建产物） 一律复制进lib/，二进制设$ORIGIN/../libRPATH。