validate.yml

name: validate

on:
  pull_request:
    # mcpp.toml and index.toml carry the workspace member list, the inherited
    # [indices] redirect and the client version floor — a change to any of them
    # can break every member, so they gate the run like the descriptors do.
    paths: ["pkgs/**/*.lua", "tests/**", "README.md", "README.zh-CN.md", "mcpp.toml", "index.toml", ".github/workflows/validate.yml"]
  push:
    branches: [main]
  schedule:
    # WEEKLY full regression — every workspace member, regardless of diff.
    #
    # This is the safety net for the one thing selective testing structurally
    # cannot see: two changes that are each green on their own and break each
    # other once both have landed. Neither PR's diff names the member that
    # breaks, so neither PR selects it.
    #
    # Weekly rather than nightly because a full run is no longer cheap — with
    # the second linux toolchain leg it is ~11 hours of runner time — and the
    # net catches the same interactions whether it is cast every day or every
    # seven. Sunday 06:00 UTC.
    - cron: "0 6 * * 0"
  workflow_dispatch:
    inputs:
      update_pr:
        description: "Bot update PR number; dispatch its exact head with selective regression"
        type: string
        required: false
      cache:
        description: "Package build cache — 'local' rebuilds every dependency per member, which is what the timing table should be read against when comparing"
        type: choice
        options: [global, local]
        default: global

env:
  # 2026.8.5.4 carries two things this workflow depends on:
  #   .5.1  `tools = [...]` — how a consumer asks for a dependency's
  #         `kind = "bin"` target, which tests/examples/protobuf-protoc is built
  #         on. Before it: "tools must be a string, inline dep table, or nested
  #         table".
  #   .5.4  windows links with lld. link.exe caps a response-file LINE at
  #         128 KiB and opencv-module / opencv-module-dnn went past it —
  #             fatal error LNK1170: line in command file contains 135135 or
  #             more characters
  #         after 795s / 1166s of compiling. .5.3 newline-separated OUR response
  #         file, which was necessary but not sufficient: clang, acting as the
  #         driver, writes a SECOND one for the linker that we do not control.
  #         lld's response-file parser has no per-line limit at all.
  #         Together with re-enabling the global package cache below, this is
  #         what makes a green FULL run possible again: .5.3 removes the
  #         windows link failure, the cache removes the 150-minute timeout.
  #
  # Neither of them moves index.toml's min_mcpp: exposing compat.protobuf's `protoc`
  # target is additive, and 2026.8.3.3 still parses that descriptor with an
  # empty unknown_keys. The floor an index publishes decides whether older
  # clients keep working at all (mcpp#349), so it moves only when a descriptor
  # genuinely stops being readable — which is not the case here.
  # 2026.8.3.1: on macOS, a global object that touches std::cout during static
  # init crashes on sight (mcpp#336). Mach-O has no priority-ordered init
  # section and libc++'s <iostream> carries no ios_base::Init guard of its own,
  # so the streams are still all-zero when an archive member's initializer
  # runs. mcpp now links a generated object FIRST whose constructor brings them
  # up. It is not fixable package-side — std::ios_base::Init is only
  # forward-declared in libc++'s <ios> — so boost-ext.ut genuinely requires
  # this floor on macOS, and min_mcpp/latest_mcpp move with the pin as they
  # always have (a consumer below the floor would get a segfault with no
  # diagnostic, which is worse than E0006).
  # 2026.8.3.3 is the pin rather than .3.1: .3.2/.3.3 are cross-compilation
  # fixes (PE artifact naming, -static host-vs-target) that no leg of this
  # matrix exercises, so taking the newest of the train costs nothing.
  # 0.0.109: a bare dependency's wire address takes BOTH halves from the
  # descriptor the identity gate accepted (mcpp#286). This is the client-side
  # other half of the SPEC-001 migration below: mcpp used to take the NAME from
  # the descriptor and the NAMESPACE from the request, so a bare `gtest =
  # "1.15.2"` addressed `mcpplibs:gtest` — a key no index has. It only ever
  # worked because the pre-migration literal `package.name` read
  # "compat.gtest", which the hardcoded `compat.<short>` retry then caught.
  # Short names removed that coincidence and left every bare request against
  # this index broken on 0.0.108, which is why min_mcpp moves to 0.0.109.
  # Note this index cannot cover that spelling itself: the `[indices]` redirect
  # is keyed by the REQUEST's namespace, so a bare dependency resolves from the
  # published remote index rather than the checkout under test. Bare-name
  # resolution against a short-name index is upstream's e2e 165; what moving
  # the floor buys here is that consumers of THIS index get a client that can
  # address it.
  # 0.0.106: SPEC-001 package identity (mcpp#280). `package.name` is a SINGLE
  # ATOMIC SEGMENT — all hierarchy lives in `package.namespace` — and mcpp
  # addresses a package by the LITERAL name it read, so descriptors no longer
  # repeat their namespace inside `name`. This index is migrated to the short
  # form, which is why min_mcpp/latest_mcpp move in lock-step: an older client
  # re-derives `<ns>.<short>`, misses, and reports a bare E_NOT_FOUND. Bundles
  # xlings 0.4.69, which keys its index by (namespace, name) so two packages
  # sharing a short name in one index are both addressable (xlings#381) — this
  # index now has three such pairs (imgui / ffmpeg / lua under compat vs the
  # default namespace).
  # 0.0.102: windows command-line ceiling (mcpp#261 — the clang scan rule got
  # its P1689 JSON through shell redirection, which forced a `cmd /c` wrapper
  # and with it cmd.exe's 8191-char limit; clang-scan-deps -o removes both, and
  # $local_includes-carrying rules now fall back to response files). The pin
  # matters here because a package consumed FROM the registry sits under a
  # ~124-char xpkgs path instead of its own ~23-char checkout, which is what
  # pushed the vendored-opencv scan command over the line. Also: purview-include
  # depfile tracking extended to Clang (#257 — stale BMI reuse), OS-conditional
  # `[build].flags` (#258), and per-OS splices keyed on the resolved target
  # rather than the host (#254).
  # 0.0.101: per-feature per-glob flags + per-OS features (mcpp#253) — what
  # lets opencv select its dnn gemm backend per platform.
  # 0.0.99: feature dep/feat forwarding (mcpp#243 — a feature can open a
  # feature OF a dependency, e.g. opencv `dnn` forwarding compat.opencv/dnn);
  # vendored xlings 0.4.67 for the >=2 index_repo install fix (mcpp#238 /
  # openxlings/xlings#374); build.mcpp compiled program named `.exe` on
  # Windows (mcpp#230 secondary surface, after the 0.0.96 scanner crash fix).
  # 0.0.98: closes the obj-path disambiguation follow-ups that gated the
  # source-build compat.opencv unification — #240 (link inputs now follow
  # the disambiguated object names, so a dependency + consumer sharing a
  # source basename like `src/main.cpp` no longer 'obj/main.o missing') and
  # #239 (absolute/`..` dep-generated source paths sanitized component-wise
  # so objects stay under obj/). Also: `MCPP_DEP_<NAME>_DIR` build.mcpp
  # contract (#241), consumer-side `default-features = false` (#242), and a
  # loud unknown-mcpp-key warning with did-you-mean (#237, replaces the
  # silent-ignore at build time). Carried from 0.0.97: default-namespace
  # index redirect (`[indices] default = { path }`), which turned the public
  # module packages (imgui/ffmpeg/opencv/tinyhttps) into ordinary workspace
  # members and retired the per-package reseeding smoke shells + their
  # dedicated jobs; synchronous nasm bootstrap (mcpp#232 — the `mcpp index
  # update` pre-step is gone), obj-path disambiguation (#233), spacey-defines
  # quoting (#234), purview-include depfile tracking (#235). Older floors of
  # note: 0.0.96 fixed the windows scanner symlink-escape crash (mcpp#230);
  # 0.0.94 fixed feature-gated `sources` under `mcpp test` (mcpp#218); 0.0.91
  # added standard = "c++fly" to the resolver grammar, so c++fly descriptors
  # get the lint WARN below, not a hard grammar-parse rejection.
  # 2026.8.6.1 是本次身份迁移的**前置**,不是顺手升级。
  #
  # `Fetcher::install_path(ns, shortName, version)` 的 legacy 扫描此前匹配任何以
  # `-x-<shortName>` 结尾的目录、不看命名空间,于是查 `ocornut:imgui@1.92.8` 会拿到
  # `compat-x-imgui/1.92.8` —— 另一个仅仅短名相同的包。它一直够不到,是因为 module
  # 层用打包计数(imgui@0.0.6)而 compat 用上游版本(compat.imgui@1.92.8),版本永远
  # 不撞;本次把 module 层对齐到上游之后它们重合了。mcpp#364 修掉了它。
  #
  # index.toml 的 min_mcpp **不动**。两个会撞的 compat 邻居都是 Form B,所以旧客户端
  # 撞上时是响亮报错而不是静默用错包;下限是一道让整个索引对旧客户端失效的闸门
  # (mcpp#349),只该在描述符真的读不动时抬。这里读得动,差的是解析得对。
  #
  # 2026.8.8.2 曾经是本 pin 的**前置**,不是顺手升级:
  #
  # `compat.glx-runtime` 依赖 mesa,mesa 声明 `xim:glibc@>=2.38`;下限被任何更高
  # 版本满足,于是安装图形栈会在既有 glibc 旁边**再装一个**。2026.8.8.2 之前的
  # mcpp 用 `readdir` 的第一项来解析「那个 glibc payload」,编译侧与产物的
  # interpreter 因此可以指向不同版本 —— 这正是 #179 落地后 `asio-module` 和
  # `core` 变红、并导致整份改动被 #180 revert 的原因。用早于 2026.8.8.2 的 mcpp
  # 重新落地这份改动,就是在仍会犯这个错的引擎上复现事故条件;
  # tests/check_graphics_install_side_effects.sh 也需要它 —— 该测试在更旧的 mcpp
  # 上失败是**有意的**。
  #
  # 2026.8.10.1 又加了一层:2026.8.8.2 的 registry 无法解析
  # `xim:libglvnd@>=1.7.0.1` —— compat.glx-runtime@2026.08.08 在 linux 上的硬依赖,
  # 任何经 compat.glfw 的 X11 profile 触及图形栈的包(eui-neo 等)都会踩到,导致
  # linux 上安装即失败(main 的 graphics 检查因此持续变红)。
  #
  # 补充根因:这不是索引的数据缺陷,是**内带 xlings 的版本**。四段版本参与范围比较
  # 由 xlings `2026.8.9.2` 修好(N 段 semver 重写);mcpp `2026.8.8.2` 内带的是
  # `2026.8.8.1`,正好落在修复之前,所以它把一个存在的包报成 `not found`。
  # `2026.8.10.1` 起内带的 xlings 已含该修复。**`xim-pkgindex` 一个字都不用改。**
  #
  # 取 `2026.8.10.3` 而不是 `.1`,是因为它另外带两个直接影响本仓的修复:
  #   - mcpp#405 依赖 BMI 缓存命中时,被恢复的包传递依赖的 `std` 没进构建图 ——
  #     **每个图形项目的第二次构建必挂**(消费方自己不 `import std` 时现形)。
  #   - mcpp#407 `mcpp build` 会重放 `mcpp test` 留下的构建图。
  #
  # ── 2026-09-06: raised 2026.8.27.2 -> 2026.9.6.3 ──────────────────────────
  #
  # The pin had drifted about ten releases behind, so this index was validated
  # against an engine no user runs. `llamacpp@b10069.2` made the drift visible
  # rather than causing it: its build program calls `mcpp::toolchain_sysroot()`
  # and `mcpp::toolchain_binutils_dir()` (2026.9.5.2+) to hand the ecosystem C
  # library to the second compiler it drives, and `mcpp::cxx_stdlib()`
  # (2026.9.6.3) to refuse a libc++ toolchain by name. Under the old pin the
  # workspace members could not compile it at all. A package whose build program
  # uses a current engine API is not a defect; a CI that cannot run current
  # engines is.
  #
  # `index.toml` min_mcpp deliberately does NOT move with this. The floor states
  # the oldest mcpp able to RESOLVE every descriptor -- descriptor grammar --
  # and every descriptor here parses under both versions. A client on
  # 2026.8.27.2 keeps the whole index; only `llamacpp@b10069.1` and `.2`, whose
  # build program needs the newer accessors, are out of reach for it, and
  # `b10069` remains published for exactly that case. Raising the floor instead
  # would refuse the index to that client entirely, which is the failure mode
  # `index-floor-must-degrade` names: the index is data, mcpp is the program,
  # and publishing data must not invalidate the program.
  #
  # THE COST IS PAID ONCE PER MOVE, AND IT IS LARGE: the members' caches key on
  # `MCPP_VERSION`, so the first run after this line changes rebuilds every
  # member on every platform. Measured: twelve workspace shards, all still cold
  # after fifty minutes. Batch descriptor changes that need a newer engine into
  # ONE pin move rather than several.
  # ── 2026-09-11: raised 2026.9.6.3 -> 2026.9.11.2 ──────────────────────────
  #
  # Same shape as the move above. `huxerui.huxerui` is a Form A package whose
  # build program calls `mcpp::package_name()` and `mcpp::package_namespace()`
  # (mcpp#587, first released v2026.9.7.1), so under 2026.9.6.3 its host module
  # does not compile at all:
  #
  #     rules.cppm:274:61: error: 'package_name' is not a member of 'mcpp'
  #
  # This landed on 2026.9.7.1 first, deliberately taking the FLOOR rather than
  # the tip, because 2026.9.9.1 introduced a scanner regression that broke
  # `mysql-connector-cpp` on linux and macOS -- mcpp read inside a block comment
  # and choked on libmysqlclient's `module (exe or dll) …`. Reduced to four
  # lines and reported as mcpp-community/mcpp#606; fixed in 2026.9.11.2, which
  # is why the pin can now sit at the current release instead of four behind.
  # Verified at 2026.9.11.2, all three:
  #
  #     four-line scanner repro           builds clean
  #     mcpp test -p huxerui-module       test result ok. 1 passed
  #     mcpp test -p mysql-connector-cpp  test result ok. 1 passed
  #
  # THE PIN IS ALSO THE ONLY WAY TO REFRESH THE INDEX, and that cost a full
  # round of wrong conclusions before it was noticed. The registry cache below
  # holds `~/.mcpp/registry`, which contains `data/xim-pkgindex` -- the resolved
  # xim index -- and its `restore-keys` prefix stops at MCPP_VERSION. So while
  # this line does not move, every run restores the same index snapshot no
  # matter what landed in xim-pkgindex meanwhile.
  #
  # That is how two successive fixes to `xim:wix` (xim-pkgindex#808, then #809)
  # both appeared to fail here: neither was ever loaded. The tell was that
  # `7zip` -- which #809 declares -- appeared zero times in a log that was
  # supposedly testing it. Moving this line evicts that cache, which is the
  # only reason the windows leg can now see the fixed recipe.
  #
  # `index.toml` min_mcpp still does NOT move: the floor is about descriptor
  # GRAMMAR, and `mcpp xpkg parse` accepts pkgs/h/huxerui.huxerui.lua under
  # 2026.8.27.2, 2026.9.6.3, 2026.9.7.1 and 2026.9.11.2 alike.
  # ── 2026-09-15: raised 2026.9.12.3 -> 2026.9.15.2 ─────────────────────────
  #
  # `llvm.libcxx` 22.1.8.3 and `openkal-llvm-runtime` 0.9.7 state
  # `[package] standard = "c++23"`, which from mcpp 2026.9.15.2 compiles a
  # C++-layer provider's own sources at that level whatever the consumer's
  # (mcpp-community/mcpp#641). Neither needs the newer engine to build here --
  # an older engine reads a dependency's `standard` as it always did -- but the
  # same release moves the object addresses of sources outside their declaring
  # package, the host-tool scratch directory, and when a dependency's link form
  # is decided, and this index is where those meet five hundred real
  # manifests. One pin move, batched with the two descriptors.
  #
  # `index.toml` min_mcpp does not move: no descriptor grammar changed.
  # ── 2026-09-16: raised 2026.9.15.2 -> 2026.9.16.1 ─────────────────────────
  #
  # `mcpp.plugins` 0.12.0 is added below, and it reads the resolved graph the
  # root build program receives from mcpp 2026.9.16.1 (`MCPP_GRAPH_FILE`). It
  # does not need that engine to build here -- a member reads the variable and
  # finds it absent under an older one, which is the pre-0.12.0 behaviour -- so
  # the descriptor is not what moves the pin.
  #
  # What moves it is that 2026.9.16.1 changes how images and runtimes are
  # decided, and this index is where that meets five hundred real manifests.
  # On ELF every C++ image a plan loads into one process now shares one C++
  # runtime, so a program over a plan-built C++ shared library takes the
  # library's contract instead of a static one (mcpp-community/mcpp#646 F3a); a
  # static package reached from several shared images is refused where the link
  # or the load cannot succeed (F1); and a package whose targets are all
  # programs no longer contributes its dependencies to a consumer's target
  # graph (#649 E6). Each of those decides what a real manifest links, and the
  # only place they are exercised at this scale is here.
  #
  # `index.toml` min_mcpp does not move: no descriptor grammar changed.
  # ── 2026-09-17: raised 2026.9.16.1 -> 2026.9.17.2 ─────────────────────────
  #
  # 2026.9.17.2 installs the C runtime the default SubOS declares before the
  # toolchain fixup consumes it, and looks the payload up by its exact version
  # (mcpp-community/mcpp#660). Before it, the declared glibc was present only
  # when xlings had installed it as a toolchain dependency, so a runner whose
  # cache restored a toolchain beside another glibc revision failed at the
  # fixup once xim-pkgindex moved glibc to 2.44.3. Every Linux member here goes
  # through that fixup, and this index is where it meets the most manifests.
  # The release in between (2026.9.17.1, compile-flag words, #655) is included.
  #
  # `index.toml` min_mcpp does not move: no descriptor grammar changed.
  # ── 2026-09-17: raised 2026.9.17.2 -> 2026.9.17.3 ─────────────────────────
  #
  # 2026.9.17.3 closes the compiler's own search of the host's C library when
  # the dependency graph supplies it (mcpp-community/mcpp#662): an openkal
  # graph compiled for Windows on a host with mingw headers mixed them with
  # musl's. compat.zlib and compat.mbedtls adapt to that graph by c-abi, and
  # tests/openkal measures it with the host's mingw headers installed.
  #
  # `index.toml` min_mcpp does not move: no descriptor grammar changed.
  # ── 2026-09-18: raised 2026.9.17.3 -> 2026.9.18.1, WITH index.toml's floor ─
  #
  # This is the mcpp release that understands `[c-abi]` (design:
  # openkal/.agents/docs/2026-09-18-openkal-c-environment-and-personalities-
  # design.md §3.2; landing order: the sibling
  # 2026-09-18-c-environment-execution-plan.md §5, step 1). It is the reason
  # `index.toml`'s min_mcpp moves in the SAME commit as this line, unlike the
  # two raises above it: this one is a grammar the index is about to gain a
  # user of. openkal-musl 0.15.0 and openkal-llvm-runtime 0.11.0, the two
  # packages that will declare and consume `[c-abi] presents = "posix"`, are
  # not registered yet (no version, no sha256 exist) — see the versions in
  # pkgs/o/openkal-musl.lua and pkgs/o/openkal-llvm-runtime.lua and the note
  # beside them — so this raise alone changes nothing observable: no
  # descriptor in this index writes `[c-abi]` today, and mcpp's own release
  # notes state that a package which does not declare the block gets the same
  # command line, byte for byte, as before. The first build after this bump
  # is a cold one (mcpp's build-cache epoch moved from 2 to 3 in this same
  # release, because the old cache key did not cover the realised [c-abi]
  # environment).
  # ── 2026-09-25: raised 2026.9.20.1 -> 2026.9.25.1 ─────────────────────────
  #
  # 2026.9.25.1 changes what reaches a dependency's compile lines
  # (mcpp-community/mcpp#690): a consumer's `include_dirs` no longer reach the
  # units of its dependencies, `[build] defines` is a set keyed by macro name,
  # and a member inside an index archive inherits its archive's workspace. No
  # member here declares `include_dirs` and no archive here declares an
  # inheritable workspace table, so no descriptor needs a change; the raise is
  # the measurement of that statement on every member and platform. The build
  # cache epoch moves from 3 to 4 in the same release, so the first run is cold.
  #
  # `index.toml` min_mcpp does not move: no descriptor grammar changed.
  # ── 2026-09-26: raised 2026.9.25.1 -> 2026.9.26.1 ─────────────────────────
  #
  # 2026.9.26.1 changes three things a member's build can see
  # (mcpp-community/mcpp#693-#696): a package's `c_standard` reaches that
  # package's own C units and no one else's; a link over a graph-supplied C
  # library searches no host library directory; and paths are held as UTF-8,
  # so a project in a directory whose name is not ASCII builds on Windows. The
  # three descriptors that spell `-D_GNU_SOURCE` because a dependency's
  # `gnu11` used to be dropped keep the define, which works under every engine
  # this index admits. The raise is the measurement of that on every member
  # and platform.
  #
  # `index.toml` min_mcpp does not move: no descriptor grammar changed.
  # ── 2026-09-26: raised 2026.9.26.1 -> 2026.9.26.2 ─────────────────────────
  #
  # 2026.9.26.2 changes what a member's build writes and reads
  # (mcpp-community/mcpp#699, #701, #703): the compile database is kept per
  # configuration with `directory` naming the output directory; link flags are
  # read as words, so a `$ORIGIN` reaches the linker as written; build programs
  # gain `runtime_search_dir` and the `prepare` role, and an action role
  # outside source, check, object, artifact and prepare is refused. No
  # descriptor here escapes a link flag for the shell or ninja by hand, and the
  # build programs here use only `source`, so no descriptor needs a change; the
  # raise is the measurement of that on every member and platform. The
  # build-program cache epoch moves from 2 to 3, so each build program runs
  # once more after the raise.
  #
  # `index.toml` min_mcpp does not move: no descriptor grammar changed.
  # ── 2026-09-28: raised 2026.9.26.2 -> 2026.9.28.1 ─────────────────────────
  #
  # 2026.9.28.1 changes four things a member's build can see
  # (mcpp-community/mcpp#717-#726): the lib root of a host-module package is
  # ordered among its own package's units by what it imports; a workspace
  # root with its own [package] passes its workspace tables to a member it
  # reaches through a path dependency; two packages may deploy byte-identical
  # files to one name; and on Windows a program built on an LLVM row uses the
  # dynamic CRT with the toolset's redistributable placed beside it, the model
  # the cl row already used, which is also the CRT of vcpkg's default
  # `x64-windows` triplet and of CMake's default. No descriptor here spells a
  # CRT flag (a hand-written one that contradicts the model is now refused),
  # so no descriptor needs a change; the raise is the measurement of that on
  # every member and platform.
  #
  # `index.toml` min_mcpp does not move: no descriptor grammar changed.
  # ── 2026-09-28: raised 2026.9.28.1 -> 2026.9.28.2 ─────────────────────────
  #
  # 2026.9.28.2 (mcpp-community/mcpp#730) changes four things a member's build
  # can see: on Windows the files beside a program are decided by one runtime
  # placement resolver, which places the toolset's C++ runtime over an older
  # copy a dependency ships and states that copy once as a packaging fault; on
  # Windows every action runs with the toolset's runtime directory first on
  # PATH, so each action re-runs once after the raise; a GNU-dialect compile
  # on Windows writes a depfile, so those members rebuild once; and matching
  # conditional tables apply in order of selector specificity. The one
  # descriptor with two `target_cfg` predicates, compat.libarchive, names two
  # that never match one target (linux, macos), so no descriptor's flags move.
  #
  # `index.toml` min_mcpp does not move: no descriptor grammar changed.
  #
  # 2026.9.28.3 (mcpp-community/mcpp#734): build programs speak protocol 14,
  # so every member's build program re-runs once; mcpp.plugins 0.17.0 states it
  # as its floor, and tests/examples/mcpp-plugins builds against it. Engine
  # warnings the members may newly print: W1 to W3 of SPEC-008 (the library
  # interface), warnings only.
  #
  # 2026.9.29.1 (mcpp-community/mcpp#738) builds a workspace as one graph per
  # configuration: `mcpp test -p <member>` plans that member and what it
  # reaches in the workspace's build directory, `target/<triple>/<key>/` at
  # this repository's root, with the member's products in `bin/<name>/`. The
  # members' own `tests/examples/*/target/` directories are no longer written.
  # The directory name no longer hashes a package's own flags, so a member
  # whose flags differ from another's shares the directory and ninja rebuilds
  # the commands that changed.
  #
  # 2026.9.29.2 (mcpp-community/mcpp#739) reads what a member states about
  # itself from the member: a member's own relative [indices].path is anchored
  # at the member, a member whose tests import std is tested with the std
  # module built, and a member's product directory holds every shared library
  # of its closure. 2026.9.29.1 is not used here: its sweep refused the 47
  # members that declare their own [indices].
  #
  # 2026.9.29.3 (mcpp-community/mcpp#740): a program a member ships through
  # `artifacts` waits for no plan-level runtime file, and its own runtime
  # files are placed beside it in the member's product directory.
  #
  # 2026.9.29.4 (mcpp-community/mcpp#741): a program a member ships through
  # `artifacts` links with its own closure's line, and `${mcpp.bin_dir}` in an
  # action a member's build program declares is that member's product
  # directory.
  #
  # 2026.9.29.5 (mcpp-community/mcpp#742): a member's build program is reused
  # across selections and runs after those of the members it depends on, the
  # build database plans by configuration, and a build reports each step once
  # with its outcome while a status line states the build.
  #
  # 2026.9.30.1 (mcpp-community/mcpp#743): a build names each package when it
  # does work, the live display is drawn in one write, a path or git
  # dependency's bare key states only its short name, and the lock file keeps
  # one entry per identity.
  #
  # 2026.9.30.2 (mcpp-community/mcpp#745): a build exits after ninja, the
  # status row counts the work of the build, planning walks each package tree
  # once, a vendored xlings is replaced from the newest source (#744), a
  # module name is unique within one program rather than one build (#732),
  # a BMI served from the global cache waits for the modules it imports that
  # the build compiles, a selected workspace member declares as the root
  # (mcpp-community/mcpp#747), and the clang MSVC row states its compiler
  # version (#746).
  #
  # 2026.10.1.2 (mcpp-community/mcpp#754): every package's BMIs lie below its
  # own directory in the build directory except the root's, and a unit reads
  # one module map of what it reaches; every unit of an ELF target is compiled
  # position-independent; `mcpp pack` states its build; `[build] jobs` bounds
  # `test` and `pack` (#751, #753).
  MCPP_VERSION: "2026.10.1.2"

jobs:
  lint:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v4
      - name: Install lua
        run: sudo apt-get install -y --no-install-recommends lua5.4
      - name: Lint package descriptors
        run: |
          fail=0
          for f in pkgs/*/*.lua; do
            # 1. Lua syntax check — load (= compile) without executing.
            #    `loadfile(name, 't')` rejects bytecode and parses text only.
            if ! lua5.4 -e "assert(loadfile('$f', 't'))" >/dev/null 2>&1; then
              echo "::error file=$f::lua syntax error"
              fail=1
            fi
            # 2. xpkg V1 baseline: the file has to populate `package = { ... }`
            #    with at least `spec`, `name`, and an `xpm` table. Form A vs
            #    Form B (mcpp = "<path>" / mcpp = { ... }) is descriptor-author
            #    choice and not enforced here.
            for needle in 'spec *=' 'name *=' 'xpm *='; do
              if ! grep -q "$needle" "$f"; then
                echo "::error file=$f::missing required field ($needle)"
                fail=1
              fi
            done
            # 3. Package version identifiers and dependency versions should be
            #    bare versions ("1.2.3"), not upstream tag names ("v1.2.3").
            #    Download URLs may still contain refs/tags/v* when upstream
            #    uses that tag spelling.
            if grep -nE '\["v[0-9]+|\["[^"]+"\][[:space:]]*=[[:space:]]*"v[0-9]+' "$f"; then
              echo "::error file=$f::version identifiers must not use a leading v"
              fail=1
            fi
            # 4. Mirror table sanity: when a download `url` is written as a
            #    { GLOBAL=..., CN=... } table, both regions must be present and
            #    the CN entry must point at the gitcode mcpp-res mirror.
            if ! lua5.4 tests/check_mirror_urls.lua "$f"; then
              fail=1
            fi
            # 5. `name` must be a SINGLE ATOMIC SEGMENT; hierarchy belongs in
            #    `namespace` (mcpp SPEC-001 §3.2). The legacy fully-qualified
            #    spelling stays accepted. Cheap second gate: it runs before the
            #    pinned mcpp is even downloaded, and mcpp >= 0.0.106 enforces
            #    the same rule inside `mcpp xpkg parse`.
            if ! lua5.4 tests/check_package_name.lua "$f"; then
              fail=1
            fi
            # 5b. The build-plugin naming rule (mcpp#734): a namespace may not be
            #    one of the reserved second segments of `mcpp.` (core, plugins,
            #    deps, rules, dist, tools), which name the mcpp project's own
            #    modules. The module names themselves are checked where members
            #    build (tests/run_members.sh), by the engine's own warning.
            if ! lua5.4 tests/check_reserved_namespace.lua "$f"; then
              fail=1
            fi
            # 6. c++fly admission policy (mcpp design 2026-07-14 §11-Q2, v1):
            #    c++fly means "toolchain's latest level + every experimental
            #    gate" — deliberately toolchain-dependent, so a published
            #    package built with it is not reproducible for consumers.
            #    Policy: WARN (never fail) and observe ecosystem usage before
            #    deciding whether to tighten. Two spellings: `language = ` is
            #    the descriptor's inline mcpp-segment key; `standard = ` covers
            #    mcpp.toml content embedded in heredoc/generated_files blocks.
            if grep -nE '\b(language|standard)[[:space:]]*=[[:space:]]*"c\+\+fly"' "$f" >/dev/null; then
              echo "::warning file=$f::declares C++ standard \"c++fly\" (experimental playground mode) — toolchain-dependent and non-reproducible for consumers; published packages should pin a concrete standard (c++23/c++26)"
            fi
          done
          [ $fail -eq 0 ] && echo "All package files valid."
          exit $fail
      # ── Whole-repository check (needs every descriptor at once) ──────
      # An install() hook addressing a sibling package does so by
      # `<namespace>:<literal package.name>`, and a miss returns nil rather
      # than raising — so a stale spelling surfaces far downstream (a broken
      # libxcb showed up as a libX11 link error). Verified across the repo
      # because it needs the full set of declared identities.
      - name: Lint cross-package references
        run: lua5.4 tests/check_cross_package_refs.lua pkgs/*/*.lua
      # ── Partial version bumps ────────────────────────────────────────
      # A bump is a one-line-looking edit that has to land in N platform
      # sections. Editing `xpm.linux` and reading the file back gives a file
      # that CONTAINS the new version, so the author -- and any whole-file
      # grep -- sees success, while the other platforms are left behind. The
      # failure then surfaces as `<pkg>@<ver> not found` on a platform,
      # against a file that literally contains that version string.
      #
      # Measured 2026-08-06: xpkg 0.0.52 and 0.0.53 were both added to
      # `xpm.linux` alone; linux CI went green twice while macOS and Windows
      # failed, and eight checks made from the outside all came back correct
      # because each asked "is it in the index?" instead of "is it in THIS
      # platform's section?". Whole-repo, because a partial bump is only
      # visible by comparing sections against each other.
      - name: Lint platform version parity
        run: lua5.4 tests/check_platform_version_parity.lua pkgs/*/*.lua

      # A published version names one set of bytes. Nothing upstream checks
      # that a package changing its content also bumps its number, and a
      # repeated key here is accepted by Lua and silently collapsed --- so a
      # release that republished an old tag would leave no other trace.
      - name: Lint duplicate versions
        run: lua5.4 tests/check_duplicate_versions.lua pkgs/*/*.lua
      - name: Check Quill install hook boundaries
        run: lua5.4 tests/check_quill_install.lua
      # The compatibility measurement's classifier decides a PUBLISHED figure
      # -- which members "run", which "build", and which were correctly
      # "refused" -- and the only thing that exercised it was the four-hour
      # openkal-compat matrix, on real members, where a misclassification is
      # a number nobody can trace back to a rule. `classify_failure` is split
      # out so it can be stated in a second, here, with no toolchain and no
      # network. It runs in `lint` rather than in that workflow because its
      # cases are about the rule and not about any member.
      - name: The compatibility classifier agrees with its own rules
        run: python3 tests/openkal/compat.py selftest
      - name: The red-member rule agrees with its own cases
        run: python3 tests/red_members.py selftest
      # ── Single-source-of-truth grammar check ─────────────────────────
      # `mcpp xpkg parse` uses EXACTLY the resolver's parser, so what
      # passes here is what builds for users of the pinned MCPP_VERSION.
      # Strict by default: unknown mcpp-segment keys fail (they would be
      # silently ignored at build time). This also mechanically enforces
      # the rollout rule "floor first, new grammar after": descriptors
      # needing a newer grammar cannot pass a lint pinned to an older mcpp.
      - name: Download pinned mcpp
        run: |
          curl -L -fsS -o mcpp.tar.gz \
            "https://github.com/mcpp-community/mcpp/releases/download/v${MCPP_VERSION}/mcpp-${MCPP_VERSION}-linux-x86_64.tar.gz"
          tar -xzf mcpp.tar.gz
          echo "MCPP=$PWD/mcpp-${MCPP_VERSION}-linux-x86_64/bin/mcpp" >> "$GITHUB_ENV"
      - name: Parse descriptors with the resolver grammar (mcpp xpkg parse)
        run: |
          fail=0
          for f in pkgs/*/*.lua; do
            if ! "$MCPP" xpkg parse "$f" > /dev/null; then
              echo "::error file=$f::mcpp xpkg parse failed (resolver grammar)"
              fail=1
            fi
          done
          [ $fail -eq 0 ] && echo "All descriptors parse with mcpp ${MCPP_VERSION}."
          exit $fail

  mirror-cn-reachable:
    # Closed-loop guard for the CN mirror: every CN url referenced by a
    # descriptor must be a live, downloadable gitcode release asset.
    #
    # ── Scoped to the CHANGED descriptors ─────────────────────────────────
    # This used to walk all 87 descriptors on every run — 105 unique urls,
    # each fetched in full (`curl -o /dev/null` downloads the body; that is
    # the point, a gitcode release page can 200 while the asset behind it is
    # gone) — while a PR mirrors one package's assets at a time. Everything
    # else it re-downloaded was mirrored by an earlier PR that already
    # checked it, and a url that was live then can only have gone dark
    # SINCE, which is decay over time rather than something this PR did.
    #
    # Decay is real, and it is what the weekly cron below is for: `select`
    # answers `__ALL__` for schedule/dispatch, so the full 105-url sweep
    # still happens — once a week, and on demand — instead of on every push.
    # It also still happens on any PR that touches this workflow file or a
    # shared test script (including tests/list_cn_urls.lua, the extractor
    # this job reads the urls with), because those widen `select` to a full
    # run for the same reason they widen the workspace matrix.
    #
    # `needs: select` rather than a second diff here: the "can I trust this
    # diff" question is already answered there, in one place.
    needs: select
    # Empty when the diff names no descriptor at all — a docs-only or
    # tools-only PR mirrors nothing.
    if: needs.select.outputs.pkgs != ''
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v4
      - name: Install lua
        run: sudo apt-get install -y --no-install-recommends lua5.4
      # REACHABLE IS NOT ENOUGH, and the gap is not hypothetical: a mirror
      # asset whose CONTENT is stale answers 200 all day. It happens whenever a
      # fork tag is re-cut at the same upstream version — gitcode cannot
      # overwrite a same-named asset (and DELETE on a release returns 405), so
      # the old tarball stays up while GLOBAL moves on, and CN users silently
      # get the previous build. Measured twice on mcpplibs/libglvnd v1.7.0.
      #
      # tests/list_cn_urls.lua already emits `url<TAB>sha256`; this now uses
      # the second column instead of throwing it away.
      - name: Check CN mirror assets are reachable AND match the descriptor
        env:
          PKGS: ${{ needs.select.outputs.pkgs }}
        run: |
          fail=0
          if [ "$PKGS" = "__ALL__" ]; then
            files=$(printf '%s\n' pkgs/*/*.lua)
            echo "full sweep: every descriptor"
          else
            files=$PKGS
            echo "changed descriptor(s): $files"
          fi
          # collect unique CN urls across the selected descriptors
          : > /tmp/cn.tsv
          for f in $files; do
            lua5.4 tests/list_cn_urls.lua "$f" >> /tmp/cn.tsv || true
          done
          sort -u /tmp/cn.tsv -o /tmp/cn.tsv
          total=$(grep -c . /tmp/cn.tsv || true)
          echo "checking $total CN mirror url(s)"
          while IFS=$'\t' read -r url sha; do
            [ -z "$url" ] && continue
            # follow redirects; gitcode release assets resolve to object storage.
            # The body is downloaded either way — a gitcode release page can 200
            # while the asset behind it is gone — so hashing it costs nothing
            # extra and answers the second question too.
            got=$(curl -fsSL --retry 2 --max-time 120 "$url" 2>/dev/null | sha256sum | cut -d' ' -f1)
            if [ -z "$got" ] || [ "$got" = "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855" ]; then
              # empty stream hashes to that constant; treat it as unreachable
              echo "::error::CN mirror unreachable or empty: $url"
              fail=1
            elif [ -n "$sha" ] && [ "$got" != "$sha" ]; then
              echo "::error::CN mirror CONTENT differs from the descriptor: $url"
              echo "           descriptor: $sha"
              echo "           mirror:     $got"
              echo "           gitcode cannot overwrite a same-named asset; delete the"
              echo "           release asset and re-publish it with gtc."
              fail=1
            else
              echo "ok: $url"
            fi
          done < /tmp/cn.tsv
          [ $fail -eq 0 ] && echo "All CN mirror urls reachable and matching."
          exit $fail

  # ── The whole test surface, as a mcpp workspace ───────────────────────
  # mcpp-index is a mcpp [workspace]; every per-library test project under
  # tests/examples/ is a member. `mcpp test --workspace` builds + runs each
  # member's tests/ (behavioral assertions) on each OS — members self-gate by
  # `[target.'cfg(...)']` (e.g. the X11/glfw stack is linux-only, openblas is
  # windows-only), so one command covers the matrix with no shell driver.
  # The ~/.mcpp/registry cache carries the built compat packages (xpkgs) across
  # runs, so repeat builds are fast.
  #
  # timeout-minutes is sized for the COLD build, not the cached path. The opencv
  # module package carries a from-source OpenCV 5 build, and each feature variant
  # re-keys the store into a full recompile, so a full run (forced whenever this
  # workflow file changes — e.g. a version bump) serially builds three OpenCV
  # variants on one runner: the opencv-module base member plus the `unifont` and
  # `dnn` feature members. The registry cache (restore-keys prefix below)
  # amortizes those across subsequent runs. 150 covers the one-time cold full
  # build with headroom; it is a ceiling, not a target.
  # ── The plan, computed ONCE ───────────────────────────────────────────
  # Was inlined in every workspace job — three runners each re-deriving the
  # same answer. It now also has to be decided BEFORE the matrix exists,
  # because the matrix's shard dimension depends on it: a full run fans out,
  # a selective one does not.
  select:
    runs-on: ubuntu-latest
    permissions:
      contents: read
      pull-requests: read
    outputs:
      members: ${{ steps.fanout.outputs.members }}
      matrix:  ${{ steps.fanout.outputs.matrix }}
      plan:    ${{ steps.plan_shards.outputs.plan }}
      # The descriptors this change actually touches, for `mirror-cn-reachable`.
      # `__ALL__` on a full run, empty when the diff names no descriptor at all.
      pkgs:    ${{ steps.plan.outputs.pkgs }}
    steps:
      - uses: actions/checkout@v4
        with:
          fetch-depth: 0
      - name: Install lua
        run: sudo apt-get install -y --no-install-recommends lua5.4
      # ── Selective member testing ──────────────────────────────────────
      # `mcpp test --workspace` builds every member (opencv, ffmpeg, …) and
      # dominates CI wall-clock, while a PR almost always touches one
      # package. Map changed files → affected members and test only those:
      #   pkgs/<x>/<lib>.lua    → members whose mcpp.toml references <lib>
      #   tests/examples/<m>/** → member <m>
      # A push to main is mapped the same way — the merge's own diff — so a
      # merge costs what its PR cost. Run the FULL workspace when the change
      # can affect everything: the weekly cron and manual dispatch (which mean
      # "check everything" by definition), this workflow file (it carries the
      # mcpp version pins, so a version bump always re-validates every
      # package), a non-member edit to the workspace manifest, or shared test
      # scripts. Docs-only and tools/-only changes select nothing.
      # Note: bash 3.2 on macOS runners — no associative arrays here.
      #
      # The same walk also collects the CHANGED DESCRIPTORS (`pkgs` output) for
      # `mirror-cn-reachable`. Two selections, one classification: every reason
      # to distrust the diff — this workflow file, a shared test script, an
      # unclassified path, a push with no predecessor, cron/dispatch — has to
      # widen both, and writing that decision twice is how the two drift.
      - name: Select affected workspace members
        id: plan
        shell: bash
        env:
          UPDATE_PR: ${{ inputs.update_pr }}
          GH_TOKEN: ${{ github.token }}
          DEFAULT_BRANCH: ${{ github.event.repository.default_branch }}
        run: |
          # A FULL RUN IS NO LONGER SOMETHING A PR CAN TRIGGER BY ACCIDENT.
          #
          # Every reason to distrust the diff used to escalate to the whole
          # workspace. On a PR that is the wrong trade twice over: it buries
          # the change under ~22 shards of unrelated work, and it puts the
          # PR's own legs behind an hour of queue, so the thing being reviewed
          # is the last thing to report. Touching this file was enough to
          # trigger it -- which meant a one-line CI edit could not be reviewed
          # against a fast signal, measured repeatedly on #390, #391 and #392.
          #
          # So `full` now answers only where "check everything" is the actual
          # request: the weekly cron and a manual `workflow_dispatch`. On a PR
          # or a push, the same reasons instead WIDEN NOTHING and say so --
          # `widen` records them, the summary prints them, and a human who
          # wants the sweep runs the workflow by hand.
          #
          # What is deliberately NOT weakened: descriptor and member changes
          # still select their members exactly as before. The only thing that
          # changed is what happens when the diff cannot be classified.
          full() {
            echo "MEMBERS=__ALL__" >> "$GITHUB_ENV"
            echo "pkgs=__ALL__" >> "$GITHUB_OUTPUT"
            echo "full run: $1"
            exit 0
          }

          UNTRUSTED=""
          widen() {
            case "${{ github.event_name }}" in
              schedule|workflow_dispatch) full "$1" ;;
            esac
            UNTRUSTED="${UNTRUSTED}${UNTRUSTED:+; }$1"
            echo "note: '$1' would once have forced a full workspace run;" \
                 "selecting only what the diff names. Run this workflow" \
                 "manually (workflow_dispatch) for the full sweep."
          }

          # A push to main has a diff too — it was just never asked for.
          #
          # This used to be `event != pull_request -> full`, so every merge
          # re-tested all 67 members on all three platforms: ~11 hours of
          # runner time to re-confirm what the PR had already gone green on
          # minutes earlier. The premise was that a push has no base to diff
          # against, and that is not true: merges here are squashes, so
          # `github.event.before` is the previous main and
          # `before..HEAD` reproduces exactly the file list the PR saw
          # (verified on 698b95ee — same ten paths).
          #
          # schedule and workflow_dispatch stay full. They are not "a change
          # landed", they are "check everything", which is the whole point of
          # the weekly net above.
          #
          # Two-dot for push, three-dot for pull_request, deliberately: a PR
          # wants its own commits against the merge base, while a push wants
          # what actually landed on this branch.
          case "${{ github.event_name }}" in
            pull_request)
              base="origin/${{ github.base_ref }}"; range="$base...HEAD" ;;
            push)
              base="${{ github.event.before }}"
              # All-zero on branch creation; absent object after a
              # force-push that dropped it. Either way there is nothing to
              # diff against.
              #
              # `widen` no longer exits on a push, so this has to leave a
              # USABLE range behind. `HEAD` alone (the root-to-HEAD diff)
              # would name every file in the repo and select every member --
              # a full run by another name, which is what this change exists
              # to stop. An empty range selects nothing, and `widen` has
              # already said why; the summary carries it to a human, who can
              # dispatch the sweep. Rare either way: merges here are squashes,
              # so `event.before` is present on every normal push.
              usable=1
              case "$base" in
                ""|0000000000000000000000000000000000000000)
                  widen "push with no predecessor"; usable=0 ;;
              esac
              if [ "$usable" = 1 ]; then
                git cat-file -e "$base^{commit}" 2>/dev/null \
                  || { widen "push predecessor $base not in history"; usable=0; }
              fi
              if [ "$usable" = 1 ]; then range="$base..HEAD"; else range=""; fi ;;
            workflow_dispatch)
              if [ -n "$UPDATE_PR" ]; then
                base=$(python3 tests/autoupdate/automation.py verify --pr "$UPDATE_PR")
                range="$base...HEAD"
              else
                full "event=workflow_dispatch"
              fi ;;
            *)
              full "event=${{ github.event_name }}" ;;
          esac
          if [ -n "$range" ]; then
            changed=$(git diff --name-only $range)
          else
            changed=""
          fi
          printf 'changed files vs %s:\n%s\n' "${base:-<none>}" "$changed"
          sel=""; pkgsel=""
          add() { case " $sel " in *" $1 "*) ;; *) sel="$sel $1" ;; esac; }
          while IFS= read -r f; do
            [ -n "$f" ] || continue
            case "$f" in
              .github/workflows/validate.yml|tests/*.sh) widen "$f" ;;
              mcpp.toml)
                # Workspace manifest. Every new-package PR appends to the
                # members list, so that alone must NOT force a full run:
                # select the added members; anything else in this file
                # (indices, settings) affects everyone → full.
                if ! diff -q <(git show "$base:mcpp.toml" | grep -v 'tests/examples/') \
                             <(grep -v 'tests/examples/' mcpp.toml) >/dev/null; then
                  widen "mcpp.toml non-member change"
                fi
                for p in $(comm -13 <(git show "$base:mcpp.toml" | grep -o 'tests/examples/[A-Za-z0-9._-]*' | sort -u) \
                                    <(grep -o 'tests/examples/[A-Za-z0-9._-]*' mcpp.toml | sort -u)); do
                  add "${p#tests/examples/}"
                done ;;
              tests/examples/*)
                m=${f#tests/examples/}; m=${m%%/*}
                # A deleted/renamed member dir implies a mcpp.toml edit,
                # which already forces a full run above.
                [ -d "tests/examples/$m" ] && add "$m" ;;
              pkgs/*.lua|pkgs/*/*.lua)
                # A DELETED descriptor is in the diff but not on disk, and its
                # CN urls are gone with it — nothing left to reach.
                if [ -f "$f" ]; then pkgsel="$pkgsel $f"; fi
                lib=$(basename "$f" .lua); lib=${lib#compat.}
                hit=0
                for mt in tests/examples/*/mcpp.toml; do
                  if grep -q "$lib" "$mt"; then add "$(basename "$(dirname "$mt")")"; hit=1; fi
                done
                [ "$hit" = 1 ] || echo "note: no workspace member exercises $f" ;;
              # tools/ holds OFFLINE descriptor-generation and publishing
              # helpers (tools/compat-*/, tools/gtc/, publish_mcpp_index.sh).
              # Nothing under it is consumed by a package build: when one of
              # them actually changes a package, the generated pkgs/*.lua
              # changes with it and the rule above selects the right members.
              # So a tools/ edit alone selects nothing rather than forcing a
              # full workspace rebuild.
              # The timing table decides how work is DISTRIBUTED across
              # shards, never what gets built — no member's result can change
              # because a measured number moved. It used to fall through to
              # the `*)` catch-all below and force a full run, which is the
              # most expensive way in this workflow to test nothing: the next
              # full run reads the new numbers anyway.
              #
              # Note this makes the file invisible to CI. A malformed row is
              # then silent, and `plan_shards` prices anything it cannot parse
              # at the median — the exact failure mode that overflowed a shard
              # here. Guard it in `lint` if that ever bites.
              tests/member-timings.tsv) : ;;
              *.md|docs/*|.agents/*|.github/*|tools/*) : ;;
              *) widen "unclassified change: $f" ;;
            esac
          done <<EOF
          $changed
          EOF
          sel=${sel# }
          echo "MEMBERS=$sel" >> "$GITHUB_ENV"
          echo "selected members: ${sel:-<none>}"
          pkgsel=${pkgsel# }
          echo "pkgs=$pkgsel" >> "$GITHUB_OUTPUT"
          echo "changed descriptors: ${pkgsel:-<none>}"

          # Carried to the run summary, not just the log. A widened reason is
          # the one thing here a reviewer has to see WITHOUT opening a job:
          # it says this run deliberately tested less than the old behaviour
          # would have, and what to do about it.
          if [ -n "$UNTRUSTED" ]; then
            {
              echo "### Selective run (full sweep not triggered)"
              echo
              echo "These would once have forced a full workspace run:"
              echo
              echo "- $UNTRUSTED"
              echo
              echo "Selected members: \`${sel:-<none>}\`"
              echo
              echo "Run this workflow manually (**Run workflow** /"
              echo "\`workflow_dispatch\`) for the full sweep."
            } >> "$GITHUB_STEP_SUMMARY"
          fi
      # Sharding is for the FULL run only, and the shard count per platform is
      # that platform's RUNNER CONCURRENCY — not a round number.
      #
      # Measured on this repo (24 jobs queued, 6 running):
      #     macos 1 · linux 3 · windows 2
      #
      # That measurement is what makes over-sharding a real cost rather than a
      # theoretical one: at concurrency 1, eight macOS shards run BACK TO BACK
      # and each pays its own checkout + mcpp download + cache restore, so the
      # split is strictly slower than not splitting. Wall-clock is
      # ceil(shards / concurrency) x slowest-shard; shards beyond the
      # concurrency only add fixed cost.
      #
      # Re-measure with:
      #   gh api repos/<owner>/<repo>/actions/runs/<id>/jobs --paginate \
      #     --jq '[.jobs[]|select(.status=="in_progress")]|length'
      - name: Decide the fan-out
        id: fanout
        shell: bash
        run: |
          full=0; [ "$MEMBERS" = "__ALL__" ] && full=1
          emit() {   # platform os suffix ext mcpp xlings toolchain shards
            for i in $(seq 0 $(( $8 - 1 ))); do
              printf '{"platform":"%s","os":"%s","suffix":"%s","ext":"%s","mcpp":"%s","xlings":"%s","toolchain":"%s","shard":%d,"shards":%d},' \
                     "$1" "$2" "$3" "$4" "$5" "$6" "$7" "$i" "$8"
            done
          }
          # Shard count follows the WORK, per platform, measured.
          #
          # It used to be binary — full run: linux 3 / macos 1 / windows 2,
          # anything else: 1 each — which asserts two things that are not
          # true. "A partial run is small": touching a widely consumed
          # descriptor selects every member that consumes it, and this PR's
          # `pkgs/c/compat.openssl.lua` selected four; one linux shard took
          # grpc-codegen (3563s) and grpc-module (1701s) back to back and was
          # cancelled at exactly 1h30m, the job cap. "A full macOS run fits in
          # one shard": at 6922s measured it does not — it fit only while the
          # total sat just under the cap, and the run that added grpc-codegen
          # (1715s on macOS) pushed it to 1h30m20s.
          #
          # Sharding is not only a wall-clock lever. Each shard is its own job
          # with its own `timeout-minutes`, so it is also how the work is made
          # to FIT. That is the half the "macOS concurrency is 1, so extra
          # shards run back to back" note above left out: back to back is fine
          # when the alternative is not finishing.
          #
          # tests/member-timings.tsv already holds every member's measured
          # cost and plan_shards.lua already packs by it (LPT), so the fan-out
          # comes from the same table: sum this run's members for that
          # platform and take one shard per ~70 minutes, which leaves ~20
          # minutes of headroom under the 90-minute cap for a cold cache.
          #
          # The caps are where runner concurrency comes back in: linux 4,
          # macOS 2, windows 2. Past those, more shards buy fit that is
          # already there and pay another checkout + mcpp download + cache
          # restore.
          #
          # linux was 3 — "the measured concurrency, a 4th shard would queue".
          # That cap outlived its premise twice over: the workspace grew, and
          # the second toolchain leg means linux emits 2 x N jobs, so it
          # queues at any N. It is now 4, which is what the formula asks for.
          #
          # Full run, from the table: linux 15891s over 4 -> ~74min slowest
          # shard, macOS 6922s over 2 -> ~57, windows 8043s over 2 -> ~67.
          #
          # linux is still the tight one, and the number to watch is the
          # SLOWEST shard against the 90-minute job cap, not the average.
          # Simulated on the current table: 3-way 98min, 4-way 74, 5-way 60.
          # The levers, in order: raise the linux cap to 5, then the job
          # timeout.
          #
          # Keep tests/member-timings.tsv fresh — that is not housekeeping. An
          # untimed member is priced at the MEDIAN, so a heavy newcomer packs
          # like a trivial one. `mysql-connector-cpp` (881s) and
          # `libmysqlclient` (329s) landed on shard 0 next to grpc-codegen
          # exactly that way and pushed it past the cap, and the stale table
          # then hid the overflow by under-counting the total.
          # `plan_shards.lua <platform> 0 0` reports the platform's total cost.
          #
          # This used to sum the timing table's rows inline, which was the same
          # answer only while the table listed every member. It now lists just
          # the heavy ones, so summing rows measures a fraction of the work:
          # 10965s instead of 16095s on linux, i.e. 3 shards where 4 are
          # needed, i.e. a slowest shard of 107 minutes against a 90-minute
          # cap. Asking plan_shards keeps ONE definition of both "which members
          # are there" and "what does an untimed one cost".
          #
          # The estimate is tuned on linux (16095s vs 15891s measured, +1.3%).
          # macOS and windows members are cheaper on average than the default,
          # so their totals come out high — harmless here because both are
          # already at their cap, and erring toward more shards is the safe
          # direction. Revisit if either cap moves.
          shards_for() {   # platform cap -> shard count
            local secs sel=""
            [ "$MEMBERS" = "__ALL__" ] || sel="$MEMBERS"
            secs=$(lua5.4 tests/plan_shards.lua "$1" 0 0 $sel)
            local n=$(( secs / 4200 + 1 ))
            [ "$n" -gt "$2" ] && n="$2"
            echo "$1 work: ${secs}s (measured) -> $n shard(s)" >&2
            echo "$n"
          }
          # linux 3 -> 4. The cap was the binding constraint, not the ~70min
          # target: `secs / 4200 + 1` asks for 4 on the refreshed table
          # (15891s) and has been getting clamped back to 3. Simulated on that
          # table, slowest shard: 3-way 98min (OVER the 90min job cap — and
          # `linux default 0/3` was in fact cancelled at 90min on run
          # 31266814148), 4-way 74min, 5-way 60min. 4 is the formula's own
          # answer and leaves 16 minutes for a cold cache; 5 is there if that
          # stops being enough.
          #
          # SUPERSEDED 2026-09-06: it stopped being enough. The table is now
          # 23925s rather than 15891s, so the paragraph above describes a
          # measurement that no longer holds; see the dated block below the
          # concurrency note for what replaced it.
          #
          # This does cross the concurrency line the old comment drew: with two
          # toolchain legs linux now emits 8 jobs against a measured runner
          # concurrency of 3, so shards queue. Queuing is the right trade —
          # back-to-back shards still finish, a shard over the cap does not.
          # ── 2026-09-06: linux 4 -> 6, windows 2 -> 5 ────────────────────
          #
          # The caps were the binding constraint again, and the table has grown
          # about 50% since they were set. Measured with `plan_shards.lua` on
          # this tree: linux 23925s, macos 16676s, windows 18435s, so the
          # formula asks for 6 / 4 / 5 and was being clamped to 4 / 2 / 2.
          # Windows at 18435s over two shards is roughly 150 minutes a shard,
          # over the 90-minute job cap before a cold cache is even considered.
          #
          # Observed, not predicted. On the 2026.9.6.3 pin move (#361) three
          # shards were cancelled at exactly 90 minutes: `linux default 0/4`,
          # `linux default 3/4` and `windows default 0/2`. The two linux ones
          # passed on a re-run against the warm registry cache; the Windows one
          # timed out AGAIN at 1h30m13s with the cache restored, because the
          # pin change invalidates every MEMBER build and that shard carries
          # godot-cpp-m twice (9.5 and 10 minutes) plus opencv.
          #
          # macos stays at 2. It has never been cancelled, so raising it would
          # be a prediction rather than a measurement; the formula's 4 is there
          # the day it is.
          ln=$(shards_for linux 6)
          mn=$(shards_for macos 2)
          wn=$(shards_for windows 5)
          # THE TWO LINUX LEGS ARE NOT THE SAME COST, AND THE PLANNER CANNOT
          # SEE THAT. `plan_shards.lua` takes a PLATFORM and reads one timing
          # table; it has no toolchain dimension, so `linux default` and
          # `linux llvm` receive an identical split of an identical estimate.
          # They are not identical work: the llvm leg resolves a different
          # compiler with no sysroot, and it is the slower of the two.
          #
          # Measured on this PR: with both legs at 6, every `linux default`
          # shard finished and `linux llvm 5/6` was cancelled at 1h30m16s. The
          # split was right for the leg the table was measured on and wrong for
          # the other one.
          #
          # Giving llvm its own count is the smallest correct fix. Teaching the
          # planner about toolchains is the larger one, and it needs a timing
          # table per leg that CI does not produce yet.
          lln=$(( ln + 3 ))
          {
            printf '{"include":['
            # linux runs TWICE: once on mcpp's default toolchain and once on
            # llvm. Not redundancy — the two differ in a way that decides
            # whether a descriptor is correct. mcpp's linux default (gcc)
            # reaches the compiler through --sysroot into a clean xlings
            # subos, so the HOST /usr/include is not on the search path;
            # llvm has no sysroot and the host's headers are. compat.ffmpeg
            # put its vendored source root on -idirafter (below the system
            # dirs) and was green here for months while being broken on any
            # developer box with libavutil-dev installed — the gcc leg
            # structurally cannot see that class of bug.
            #
            # Cost, measured: this doubles linux to 6 jobs (2 x 3 shards) at
            # ~75min each, against a measured linux runner concurrency of 3 —
            # so the second leg QUEUES behind the first rather than running
            # beside it, and a full run's linux wall-clock roughly doubles.
            # That is the price of the coverage; the levers if it hurts are
            # in the shards_for comment above.
            emit linux   ubuntu-latest  linux-x86_64   tar.gz bin/mcpp     registry/bin/xlings     default "$ln"
            emit linux   ubuntu-latest  linux-x86_64   tar.gz bin/mcpp     registry/bin/xlings     llvm    "$lln"
            emit macos   macos-15       macosx-arm64   tar.gz bin/mcpp     registry/bin/xlings     default "$mn"
            # WINDOWS IS PINNED, like macOS above, and for a reason that cost
            # a day to find. `windows-latest` rolled to the Visual Studio 18
            # image (MSVC STL 14.51), and mcpp builds windows with clang
            # targeting x86_64-pc-windows-msvc -- so every package here
            # compiles against an STL written for a different front end.
            #
            # Most survive that. `huxerui.huxerui` did not:
            #
            #     xutility:320: error: static assertion failed: unexpected size
            #     in instantiation of 'std::_Find_vectorized<
            #         const huxerui::detail::NodeExtensionHandle, ...>'
            #
            # MSVC STL's vectorized `std::find` is guarded by a trait that
            # decides whether the element type can be compared bitwise.
            # `NodeExtensionHandle` is 24 bytes with no padding, trivially
            # copyable, `operator==` defaulted -- the guard admits it under
            # clang, and the helper it dispatches to implements 1/2/4/8-byte
            # elements and static_asserts on the rest. Guard and implementation
            # disagree about what "vectorizable" means, and only clang is there
            # to notice.
            #
            # NOT this index's bug, and not the package's: the same source, the
            # same clang, compiles on the 2022 image's STL. Upstream HuxerUI's
            # own mcpp CI is green for exactly that reason -- it pins
            # `windows-2022`. Reported as mcpp-community/mcpp#609 so the pin can
            # be lifted when the toolchain combination works.
            #
            # 13 of the 14 members on the shard that failed were unaffected, so
            # this is not a blanket breakage -- which is precisely why a rolling
            # label is the wrong thing to stand on: the next image moves the set
            # of packages that happen to trip it, and the failure arrives
            # attributed to whatever descriptor changed that week.
            emit windows windows-2022   windows-x86_64 zip    bin/mcpp.exe registry/bin/xlings.exe default "$wn"
            printf ']}'
          } | sed 's/,]}/]}/' > /tmp/matrix.json
          echo "matrix=$(cat /tmp/matrix.json)" >> "$GITHUB_OUTPUT"
          echo "members=$MEMBERS" >> "$GITHUB_OUTPUT"
          cat /tmp/matrix.json

      # The split is computed ONCE, here, and shipped to the runners as data.
      # It used to run on each runner, which needed lua5.4 on all three
      # platforms — windows has no apt or brew, and macOS's brew installs
      # `lua`, not `lua5.4`, so every non-linux shard died with
      # `lua5.4: command not found` after 16 seconds. Deciding once is also
      # simply correct: one plan, not three runners each re-deriving it.
      - name: Plan the shards
        id: plan_shards
        shell: bash
        run: |
          plan='${{ steps.fanout.outputs.members }}'
          [ "$plan" = "__ALL__" ] && plan=""
          {
            printf '{'
            first=1
            # `.shards`, NOT the number of matrix entries.
            #
            # These stopped being the same number the moment linux grew a
            # second toolchain leg: the platform emits 2 x 3 = 6 entries while
            # the split is still 3 ways, and every entry carries shard 0..2.
            # Planning 6 ways and consuming three of them dropped 36 of 65
            # members on the floor — silently, because a member that is never
            # assigned is indistinguishable from one that passed. `ffmpeg`,
            # `opencv-module*`, `catch2-v2` and `openssl` were among them, on
            # the very run that added the leg meant to test them.
            #
            # `.shards` is the value emit() already wrote per entry and the
            # same one the job reads for its own `matrix.shards`, so plan and
            # consumer now read one number instead of two that agreed by
            # accident.
            for spec in linux:$(jq -r 'first(.include[]|select(.platform=="linux")).shards' /tmp/matrix.json) \
                        macos:$(jq -r 'first(.include[]|select(.platform=="macos")).shards' /tmp/matrix.json) \
                        windows:$(jq -r 'first(.include[]|select(.platform=="windows")).shards' /tmp/matrix.json); do
              p=${spec%%:*}; n=${spec##*:}
              [ "$first" = 1 ] || printf ','
              first=0
              printf '"%s":{' "$p"
              for i in $(seq 0 $((n - 1))); do
                [ "$i" = 0 ] || printf ','
                m=$(lua5.4 tests/plan_shards.lua "$p" "$i" "$n" $plan)
                printf '"%s":"%s"' "$i" "$m"
              done
              printf '}'
            done
            printf '}'
          } > /tmp/plan.json
          echo "plan=$(cat /tmp/plan.json)" >> "$GITHUB_OUTPUT"
          jq . /tmp/plan.json

  # Installing the graphics stack must not change what UNRELATED members link
  # against. This is the check #179 did not have: every test in this repo asks
  # about its own package, so nothing asked whether installing one changes
  # things for everyone else -- and `asio-module` / `core` went red on a
  # change that never touched them.
  #
  # Its own job, on Linux, because it needs a real toolchain and a home it can
  # watch payloads appear in. It reports INCONCLUSIVE (and fails) rather than
  # green whenever its preconditions do not hold; see the script.
  graphics-side-effects:
    name: "graphics install: no side effects on unrelated members"
    runs-on: ubuntu-latest
    timeout-minutes: 60
    steps:
      - uses: actions/checkout@v4
      - name: Download mcpp
        shell: bash
        env:
          MCPP_ARCHIVE: mcpp-${{ env.MCPP_VERSION }}-linux-x86_64.tar.gz
          MCPP_ROOT: mcpp-${{ env.MCPP_VERSION }}-linux-x86_64
        run: |
          curl -L -fsS -o "$MCPP_ARCHIVE" \
            "https://github.com/mcpp-community/mcpp/releases/download/v${MCPP_VERSION}/${MCPP_ARCHIVE}"
          tar -xzf "$MCPP_ARCHIVE"
          root="$PWD/$MCPP_ROOT"
          # MCPP_HOME is the tarball root, not ~/.mcpp.
          #
          # A released mcpp is self-contained: it resolves its registry from
          # beside its own executable. Copying the payload tree into ~/.mcpp
          # and pointing MCPP_HOME there therefore watches one home while the
          # build uses another -- which this check detects and refuses to call
          # green (it did, on the first run). Aligning them is the fix; the
          # check was right.
          echo "MCPP=$root/bin/mcpp" >> "$GITHUB_ENV"
          echo "MCPP_HOME=$root" >> "$GITHUB_ENV"
          echo "MCPP_VENDORED_XLINGS=$root/registry/bin/xlings" >> "$GITHUB_ENV"
      - name: Side-effect check
        shell: bash
        run: |
          "$MCPP" self config --mirror GLOBAL 2>/dev/null || true
          bash tests/check_graphics_install_side_effects.sh

  workspace:
    # The shard suffix appears only when the platform is actually split, and
    # the toolchain only when it is not mcpp's default — so the existing job
    # names are unchanged and only the new llvm leg is labelled.
    name: workspace (${{ matrix.platform }}${{ matrix.toolchain == 'default' && '' || format(' {0}', matrix.toolchain) }}${{ matrix.shards == 1 && '' || format(' {0}/{1}', matrix.shard, matrix.shards) }})
    needs: select
    if: needs.select.outputs.members != ''
    runs-on: ${{ matrix.os }}
    # One shard is a fraction of the work, so this is a real ceiling rather
    # than the thing that decides whether the job finishes (a full linux run
    # used to hit 150 exactly and get cancelled).
    timeout-minutes: 90
    strategy:
      fail-fast: false
      # Whole matrix from `select`: the shard count is per-platform, because it
      # tracks that platform's runner concurrency.
      matrix: ${{ fromJSON(needs.select.outputs.matrix) }}
    steps:
      # Full history: the member-selection step below diffs against the PR
      # base to decide which workspace members to test.
      - uses: actions/checkout@v4
        with:
          fetch-depth: 0
      # The cache key is computed ONCE, here, instead of inline in the cache
      # step. `hashFiles()` globs the WORKING TREE, and actions/cache
      # re-evaluates its `key` in the post (save) step — i.e. AFTER the build,
      # when `tests/**` no longer matches 80-odd tracked sources but tens of
      # thousands of build-output files under tests/examples/*/target and
      # .mcpp (multi-GB; .gitignore does not apply to hashFiles). Hashing that
      # tree blew past the runner's 120s template-evaluation cap on windows
      # and failed an otherwise all-green job:
      #   "hashFiles('pkgs/**/*.lua, tests/**, .github/workflows/validate.yml')
      #    couldn't finish within 120 seconds"
      # `git ls-files -s` reads the INDEX, so it sees exactly the tracked
      # inputs, never build output, and reports blob SHAs git already has —
      # no file content is read at all. Freezing the result in the job env
      # also guarantees the save step keys on the same string the restore
      # step used, no matter what the build left behind.
      - name: Compute registry cache key
        shell: bash
        run: |
          # git hash-object rather than sha256sum/cut: git is already a hard
          # requirement here (checkout ran), coreutils on the windows leg is
          # only a Git-Bash convenience.
          h=$(git ls-files -s -- 'pkgs/**/*.lua' 'tests/**' '.github/workflows/validate.yml' \
              | git hash-object --stdin)
          # matrix.toolchain is in the key: both linux legs are runner.os
          # Linux, and the cache holds TOOLCHAINS and built compat packages.
          # Sharing one entry would let the gcc leg's objects answer for the
          # llvm leg — exactly the confusion this second leg exists to catch.
          echo "REGISTRY_CACHE_KEY=mcpp-registry-${{ runner.os }}-${{ matrix.toolchain }}-${{ env.MCPP_VERSION }}-$h" >> "$GITHUB_ENV"
      - name: Restore mcpp registry cache
        uses: actions/cache@v4
        with:
          # Holds toolchains AND the built compat packages (data/xpkgs), so a
          # repeat `mcpp test` rebuilds little.
          #
          # THE INDEX TREES ARE EXCLUDED, and that exclusion is the whole point
          # of this entry's shape. `data/xim-pkgindex` and `data/mcpplibs` are
          # the RESOLVED indices -- the descriptors mcpp reads. Caching them
          # under a key whose `restore-keys` prefix ends at MCPP_VERSION means
          # a run restores whatever index the last run happened to leave, no
          # matter what landed upstream since.
          #
          # That is not hypothetical. Two successive fixes to `xim:wix`
          # (openxlings/xim-pkgindex#808, then #809) were each merged, each
          # re-run here, and each appeared to change nothing -- because neither
          # was ever loaded. The tell was that `7zip`, which #809 declares,
          # appeared ZERO times in a log that was supposedly testing #809. The
          # only thing that ever refreshed the index was moving MCPP_VERSION,
          # which is a coincidence of cache-key composition rather than a
          # mechanism anyone chose.
          #
          # Excluding rather than keying on the index revision is deliberate.
          # Keying would evict everything on every upstream commit; measured on
          # a developer machine, the index trees are ~8 MB against ~28 GB of
          # `data/xpkgs`. Re-fetching 8 MB per job is the cheap half, and it is
          # the half that has to be current.
          path: |
            ~/.mcpp/registry
            !~/.mcpp/registry/data/xim-pkgindex
            !~/.mcpp/registry/data/mcpplibs
          key: ${{ env.REGISTRY_CACHE_KEY }}
          restore-keys: |
            mcpp-registry-${{ runner.os }}-${{ matrix.toolchain }}-${{ env.MCPP_VERSION }}-
      # Host tools built from source (protoc, grpc_cpp_plugin, …).
      #
      # Measured on the run that added grpc-codegen: 636s to build protoc and
      # 660s to build grpc_cpp_plugin — 1296s of a 3363s member, repeated in
      # every job of every run, because the registry cache above covers
      # ~/.mcpp/registry and the tool store lives in ~/.mcpp/build-cache.
      # `protobuf-protoc` and `grpc-module` build the SAME protoc, so a full
      # run pays for it several times over.
      #
      # A coarse rolling key is safe here, which is why this is cheap. The
      # store is content-addressed by `<pkg>@<version>/<hash>` and mcpp
      # validates an entry FIELD BY FIELD against the recorded entry.json
      # (epoch, target, host triple, compiler identity, profile, features,
      # transitive dependency closure). A stale entry is not used; it is
      # rebuilt. The worst case of a cache miss-match is the status quo.
      #
      # Deliberately NOT build-cache/v1/pkg: 6.0 GB measured locally against
      # Actions' 10 GB per-repo budget, which would evict the registry cache
      # this job needs more. The tool store is 116 MB.
      - name: Cache host tool store
        uses: actions/cache@v4
        with:
          path: ~/.mcpp/build-cache/v1/tool
          key: mcpp-toolstore-${{ runner.os }}-${{ matrix.toolchain }}-${{ env.MCPP_VERSION }}-${{ github.run_id }}-${{ matrix.platform }}-${{ matrix.shard }}
          restore-keys: |
            mcpp-toolstore-${{ runner.os }}-${{ matrix.toolchain }}-${{ env.MCPP_VERSION }}-
      - name: Download mcpp
        shell: bash
        env:
          MCPP_ARCHIVE: mcpp-${{ env.MCPP_VERSION }}-${{ matrix.suffix }}.${{ matrix.ext }}
          MCPP_ROOT: mcpp-${{ env.MCPP_VERSION }}-${{ matrix.suffix }}
        run: |
          curl -L -fsS -o "$MCPP_ARCHIVE" \
            "https://github.com/mcpp-community/mcpp/releases/download/v${MCPP_VERSION}/${MCPP_ARCHIVE}"
          case "$MCPP_ARCHIVE" in
            *.zip) powershell -NoProfile -Command "Expand-Archive -Force -Path '${MCPP_ARCHIVE}' -DestinationPath '.'" ;;
            *)     tar -xzf "$MCPP_ARCHIVE" ;;
          esac
          root="$PWD/$MCPP_ROOT"
          mkdir -p "$HOME/.mcpp/registry"
          cp -a "$root/registry/." "$HOME/.mcpp/registry/"
          # Point mcpp AT the home the caches above cover.
          #
          # Without this the copy above is one-way scenery: a released mcpp is
          # self-contained and resolves its registry from beside its own
          # executable, so every build used <tarball>/registry while the cache
          # restored, and saved, ~/.mcpp/registry. Proof from the run that
          # added grpc-codegen: abseil compiled from
          # `<tarball>/registry/data/xpkgs/compat-x-abseil/...`, and
          # xim:glibc@2.44 / xim:python@3.13.12 were downloaded again on a job
          # that reported a registry cache HIT.
          #
          # With MCPP_HOME set, the restored registry is the one in play — so
          # payloads stop being re-downloaded per job, and the tool store cache
          # below lands where the build looks for it.
          echo "MCPP_HOME=$HOME/.mcpp" >> "$GITHUB_ENV"
          if [[ "$RUNNER_OS" == "Windows" ]]; then
            echo "MCPP=$(cygpath -m "$root/${{ matrix.mcpp }}")" >> "$GITHUB_ENV"
            echo "MCPP_VENDORED_XLINGS=$(cygpath -m "$root/${{ matrix.xlings }}")" >> "$GITHUB_ENV"
            echo "$(cygpath -m "$root/bin")" >> "$GITHUB_PATH"
          else
            echo "MCPP=$root/${{ matrix.mcpp }}" >> "$GITHUB_ENV"
            echo "MCPP_VENDORED_XLINGS=$root/${{ matrix.xlings }}" >> "$GITHUB_ENV"
            echo "$root/bin" >> "$GITHUB_PATH"
          fi
      # compat.ffmpeg / compat.opencv5 carry NASM .asm sources. No host
      # install and no index-refresh pre-step needed: mcpp >= 0.0.97
      # resolves nasm itself through the same synchronous gate as the
      # toolchain (index refresh + install + payload check BEFORE the build
      # plans, mcpp#232). The sandbox copy lands in ~/.mcpp/registry, so
      # the cache carries it across runs.

      # ── Make the llvm leg look like a developer's box ─────────────────
      # A bare GitHub runner has no libav* under /usr/include, so a
      # descriptor that loses its vendored headers to the host's would still
      # be green here. That is precisely how compat.ffmpeg's -idirafter hole
      # survived: nothing in CI had a system ffmpeg to lose to. Installing
      # the dev headers is what turns this leg into a real test of include
      # ORDER rather than of include EXISTENCE.
      #
      # Deliberately only on the llvm leg. The default leg reaches its
      # compiler through --sysroot, so it is immune by construction and
      # installing there would change an existing job's environment for no
      # coverage. Keeping the blast radius to the new leg means a red here
      # is always about the thing this leg was added to watch.
      - name: Install host ffmpeg dev headers (llvm leg only)
        if: matrix.platform == 'linux' && matrix.toolchain == 'llvm'
        shell: bash
        run: |
          sudo apt-get update -qq
          sudo apt-get install -y -qq \
            libavutil-dev libavcodec-dev libavformat-dev \
            libavfilter-dev libavdevice-dev libswscale-dev libswresample-dev
          dpkg -l | grep -E 'libav(util|codec)-dev|libswscale-dev'

      # ── Select the toolchain for this leg ─────────────────────────────
      # No version pin: `llvm` resolves to whatever the toolchain family's
      # current release is for the PINNED MCPP_VERSION, which is the version
      # that mcpp itself vouches for. Pinning a concrete llvm here would fail
      # hard the day that mcpp release stops shipping it, and the point of
      # this leg is the ABSENCE of a sysroot, not a particular llvm.
      - name: Select toolchain
        if: matrix.toolchain != 'default'
        shell: bash
        run: |
          "$MCPP" toolchain install ${{ matrix.toolchain }}
          "$MCPP" toolchain default ${{ matrix.toolchain }}
          "$MCPP" toolchain list

      # ── This shard's slice of the plan ────────────────────────────────
      # `select` decided WHAT runs; this decides which part of it runs HERE.
      # Round-robin by position, which is what spreads the expensive members:
      # opencv-module / -dnn / -unifont are adjacent in the list, so `% N`
      # necessarily puts them on three different runners. A single job that
      # builds all three spends 45+ minutes on opencv alone.
      # ── This shard's slice ────────────────────────────────────────────
      # Already decided by `select` (measured-time bin packing, see
      # tests/plan_shards.lua). Arrives as data, so a runner needs no lua.
      - name: Take this shard's members
        shell: bash
        run: |
          mine='${{ fromJSON(needs.select.outputs.plan)[matrix.platform][format('{0}', matrix.shard)] }}'

          # ── A member its toolchain cannot build, and why that lives HERE ──
          #
          # Members self-gate with `[target.'cfg(...)']` (see the workspace
          # comment above), and that has a platform axis and no STANDARD
          # LIBRARY axis -- a cfg selector is not a platform. So a member that
          # cannot be built under libc++ has nowhere of its own to say so, and
          # the leg is the only layer that knows which standard library is in
          # play. That is the whole reason this list is in CI and not in a
          # descriptor; it is not a convenience.
          #
          # `mysql-connector-cpp`: the package builds STATIC libraries with
          # CMake and the system compiler, so `std::__cxx11::` crosses the
          # boundary and a libc++ consumer fails at link:
          #
          #     ld.lld: error: undefined symbol:
          #         std::__cxx11::basic_string<...>::_M_create(...)
          #
          # Its install hook cannot adapt: mcpp resolves the toolchain after the
          # dependency graph, so the consumer's standard library does not exist
          # yet when the hook runs, and MCPP_CXX_STDLIB is nil (older mcpp) or
          # empty (2026.9.12.2 and later) there (#392; mcpp-community/mcpp#613,
          # measured by its tests/e2e/648). The descriptor states the
          # requirement instead, `requires = { "mcpp:c++-abi=libstdc++" }` in its
          # `linux` table, and a libc++ consumer is refused at resolution naming
          # both implementations. (On macOS the libraries are built with Apple
          # clang, and the `macosx` table requires libc++, which the default
          # toolchain there resolves.)
          # That refusal follows the hook's source build, so the member stays off
          # this leg: the leg would spend the build to reach a known refusal.
          #
          # SKIPPED, NOT MARKED GREEN: the incompatibility is real for users on
          # a libc++ toolchain. This stops CI from re-measuring a known answer
          # every run; it does not claim the combination works.
          if [ "${{ matrix.toolchain }}" = "llvm" ]; then
            kept=""
            for m in $mine; do
              case "$m" in
                mysql-connector-cpp)
                  echo "skipping '$m' on the llvm leg: static libs are built" \
                       "against the system stdlib; see the comment above" ;;
                *) kept="$kept $m" ;;
              esac
            done
            mine="${kept# }"
          fi

          # `mp-units-module`: the reverse case. `mpusz.mp-units` requires
          # libc++ on linux because GCC 16.1 cannot import its modules
          # (`recursive lazy load`, mpusz/mp-units#717), so the gcc leg would
          # only reach that refusal. Lift once xim ships a GCC that works.
          if [ "${{ matrix.platform }}" = "linux" ] && [ "${{ matrix.toolchain }}" = "default" ]; then
            kept=""
            for m in $mine; do
              case "$m" in
                mp-units-module)
                  echo "skipping '$m' on the gcc leg: the package requires libc++" ;;
                *) kept="$kept $m" ;;
              esac
            done
            mine="${kept# }"
          fi

          echo "MEMBERS=$mine" >> "$GITHUB_ENV"
          echo "shard ${{ matrix.shard }}/${{ matrix.shards }}: ${mine:-<none>}"

      # ── Refresh the PUBLISHED index before testing ────────────────────
      # Most members resolve everything from this checkout, but a member that
      # redirects a namespace other than `compat` gets the REST from the
      # published index — and nothing here ever refreshed it. The snapshot in
      # play is whatever the pinned mcpp release vendored (the Download step
      # `cp -a`s the release's registry/ over ~/.mcpp/registry, on top of the
      # restored cache), so it is by construction older than main, and it
      # never moves: the cache is saved with that same stale copy inside it.
      #
      # mcpp does refresh on a miss for a DIRECT dependency, which is why this
      # went unnoticed — the gap is a Form-A package's TRANSITIVE dependency.
      # tests/examples/godot-cpp-module hit it head-on: the module package's
      # own compat.godot-cpp dep resolved against a snapshot predating the
      # commit that added it, and failed with `index: local index <sha> (never
      # refreshed)` even though the artifact had already been republished.
      # (Older members never noticed: their compat packages have been in the
      # index far longer than any snapshot.)
      - name: Refresh the published package index
        shell: bash
        env:
          MCPP_INDEX_MIRROR: GLOBAL
        run: |
          "$MCPP" index update

      - name: Report temporarily untestable consumers of the unavailable MySQL source
        shell: bash
        run: |
          if [[ "$RUNNER_OS" == Windows ]]; then
            python tests/autoupdate/source_outage.py members
          else
            python3 tests/autoupdate/source_outage.py members
          fi

      - name: Install nanodbc macOS development headers
        if: runner.os == 'macOS' && contains(format(' {0} ', env.MEMBERS), ' nanodbc ')
        run: brew install libiodbc

      - name: mcpp test (workspace or affected members)
        shell: bash
        env:
          MCPP_INDEX_MIRROR: GLOBAL
          # The GLOBAL package build cache is on (mcpp >= 2026.7.30.2), which
          # is the default — this step used to set `MCPP_BUILD_CACHE: local`
          # and no longer does.
          #
          # That bypass existed for mcpp#344: object-path disambiguation fires
          # on basename collisions across the WHOLE build dir — i.e. on what
          # the CONSUMER pulls in — while the cache key covered only the
          # dependency, so one entry could hold two layouts and ninja died at
          # graph time with "missing and no known rule to make it". #344
          # landed in 2026.8.3.4 with per-package Merkle keys that cover the
          # consumer-dependent layout, so the reason is gone.
          #
          # Keeping it cost real time, and the full run is where it showed:
          # with `local`, EVERY member recompiles EVERY dependency from
          # scratch. 59 members that mostly share abseil / protobuf / opencv
          # meant the same sources were built over and over —
          #
          #   linux    2h30m -> cancelled at the 150-minute timeout
          #   windows  2h20m
          #   macos    1h26m
          #
          # — and a workspace cannot be validated by a job that cannot finish.
          # With the cache on, a given (package, version, features, toolchain)
          # is built once per run and every later member hits it.
        run: |
          "$MCPP" --version
          # No `timeout` wrapper: absent on macOS runners; job-level timeout-minutes bounds it.
          # One code path: the shard step above already expanded `__ALL__`
          # into this runner's actual member names, so `mcpp test --workspace`
          # — which would ignore the sharding and rebuild everything here — is
          # gone.
          #
          # tests/run_members.sh is the SAME script you run locally. A timing
          # table that only exists in CI cannot be used while deciding what to
          # optimise, and a local harness that differs from CI measures
          # something else.
          if [ -z "$MEMBERS" ]; then
            echo "No workspace member to test here: none affected by this" \
                 "change, or every member of this shard was skipped on this" \
                 "leg (the shard step above says which, and why)."
          else
            MCPP_TIMINGS="$PWD/timings.tsv" bash tests/run_members.sh $MEMBERS
          fi

      # Per-shard timings, merged by the `timings` job below. `always()`: a
      # run that failed is exactly when knowing where the time went matters.
      - name: Upload this shard's timings
        if: always() && hashFiles('timings.tsv') != ''
        uses: actions/upload-artifact@v4
        with:
          # The toolchain is part of the name because upload-artifact@v4
          # refuses duplicates: without it both linux legs would race to
          # upload `timings-linux-0` and the second one would fail the job.
          name: timings-${{ matrix.platform }}-${{ matrix.toolchain }}-${{ matrix.shard }}
          path: timings.tsv
          retention-days: 14

      # install()-driven packages (openssl, openblas) build through their own
      # Make/Configure system, whose output xim's interface mode swallows; a
      # failed hook surfaces only as `E_INTERNAL: [<pkg>] failed:`. Each writes
      # a log into its install prefix, so on failure surface those — otherwise
      # diagnosing a platform-specific build break costs a full CI round-trip
      # per guess.
      - name: Dump install() build logs on failure
        if: failure()
        shell: bash
        run: |
          found=0
          while IFS= read -r log; do
            found=1
            echo "::group::$log"
            tail -80 "$log"
            echo "::endgroup::"
          done < <(find tests/examples "$HOME/.mcpp/registry" -name 'mcpp_*_build.log' 2>/dev/null)
          [ "$found" = 1 ] || echo "no install() build logs found"

  # ── Where the time went ───────────────────────────────────────────────
  # Sharding hides the cost: eight runners each report their own slice, and
  # nobody can see which members actually dominate. This merges them into one
  # ranking per platform, in the run summary, so the next optimisation starts
  # from measurement instead of a guess.
  #
  # `always()` — a failed run is exactly when this is worth reading.
  timings:
    needs: [select, workspace]
    if: always() && needs.select.outputs.members != ''
    runs-on: ubuntu-latest
    steps:
      - uses: actions/download-artifact@v4
        with:
          pattern: timings-*
          path: timings
        continue-on-error: true
      - name: Rank members by wall-clock
        shell: bash
        run: |
          shopt -s nullglob
          files=(timings/*/timings.tsv)
          if [ ${#files[@]} -eq 0 ]; then
            echo "no timing data (every shard skipped or failed before testing)" \
              >> "$GITHUB_STEP_SUMMARY"
            exit 0
          fi

          # Artifact name carries platform AND toolchain leg:
          # timings-<platform>-<toolchain>-<shard>. Ranking per LEG, not per
          # platform: the two linux legs are different builds and averaging
          # them would describe neither.
          for leg in linux-default linux-llvm macos-default windows-default; do
            rows=$(mktemp)
            for f in timings/timings-$leg-*/timings.tsv; do
              [ -f "$f" ] && cat "$f" >> "$rows"
            done
            [ -s "$rows" ] || { rm -f "$rows"; continue; }

            total=$(awk -F'\t' '{s += $1} END {print s+0}' "$rows")
            count=$(wc -l < "$rows")
            {
              echo "### $leg — ${count} member(s), ${total}s of member wall-clock"
              echo
              echo "| rank | seconds | share | member | result |"
              echo "|---:|---:|---:|---|---|"
              sort -rn "$rows" | awk -F'\t' -v tot="$total" '
                { pct = tot > 0 ? ($1 * 100 / tot) : 0
                  printf "| %d | %s | %.1f%% | `%s` | %s |\n", NR, $1, pct, $2, $3 }'
              echo
            } >> "$GITHUB_STEP_SUMMARY"
            rm -f "$rows"
          done

          echo "_Total is the SUM across shards; wall-clock is the slowest shard._" \
            >> "$GITHUB_STEP_SUMMARY"

          # The table that feeds the NEXT run's sharding. Emitted as an
          # artifact rather than committed automatically: a number that
          # rewrites itself on every run would make every diff noisy and would
          # silently absorb a one-off slow runner. Refresh it deliberately —
          # download this artifact and replace tests/member-timings.tsv when
          # the numbers have actually moved.
          # Top 10 per platform, not every member.
          #
          # plan_shards prices anything absent at a fixed default, and the
          # packing decisions are made by the heavy members anyway — on linux
          # the top 10 of 67 are 69% of the total, and dropping the other 57
          # moves the slowest shard by 0 minutes. A table that lists everyone
          # goes stale the moment someone adds a test; one that lists ten only
          # goes stale when the heavy set actually changes. That is the whole
          # point: adding a member must not require touching this file.
          {
            echo "# <platform>\t<member>\t<seconds> — the HEAVY members only, top 10"
            echo "# per platform, from run ${{ github.run_id }}. Everything else is"
            echo "# priced at plan_shards.lua's fixed default; adding a test does NOT"
            echo "# require touching this file. Refresh when a member becomes heavy"
            echo "# enough to enter the top 10, or one of these numbers moves."
            # ONLY the default-toolchain leg feeds this table. The llvm leg
            # runs the SAME members again, so globbing every leg would put two
            # rows per (platform, member) into the file — `sort -u` keeps both,
            # since the seconds differ. Both legs are planned from these
            # numbers, so the default leg is the right single baseline.
            #
            # Then: heaviest first, keep 10. `sort -u` before that so a member
            # reported by two shards cannot occupy two of the ten slots.
            for plat in linux macos windows; do
              for f in timings/timings-$plat-default-*/timings.tsv; do
                [ -f "$f" ] || continue
                awk -F'\t' -v p="$plat" '{ printf "%s\t%s\t%s\n", p, $2, $1 }' "$f"
              done | sort -u | sort -t"$(printf '\t')" -k3,3nr | head -10
            done
          } > member-timings.tsv
          echo "wrote member-timings.tsv ($(grep -vc '^#' member-timings.tsv) rows, top 10/platform)"

      - name: Upload the timing table for the next run's sharding
        if: always() && hashFiles('member-timings.tsv') != ''
        uses: actions/upload-artifact@v4
        with:
          name: member-timings
          path: member-timings.tsv
          retention-days: 90

  # ── Every red member names its issue ────────────────────────────────────
  # The full sweeps of 2026-09-26 failed on a member and on the CN-mirror job,
  # and no issue recorded either. A failing member is either listed in
  # tests/known-red.tsv with the open issue that owns it, or it fails this job;
  # the run summary lists every failing member with its issue (task N2 of the
  # 2026-09-28 ecosystem design). `always()`: it reads the rows the members'
  # shards uploaded, whatever the members' job concluded.
  red-members:
    name: red members name their issues
    needs: [select, workspace]
    if: always() && needs.select.outputs.members != ''
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v4
      - uses: actions/download-artifact@v4
        with:
          pattern: timings-*
          path: timings
        continue-on-error: true
      - name: Every failing member names its issue
        run: |
          set -o pipefail
          mkdir -p timings
          python3 tests/red_members.py report timings tests/known-red.tsv \
            --members-result "${{ needs.workspace.result }}" | tee -a "$GITHUB_STEP_SUMMARY"

  # ─────────────────────────────────────────────────────────────────────────
  # A full sweep that fails must be SEEN.
  #
  # Selective CI is structurally unable to notice a published package that
  # broke because something UPSTREAM changed: no PR's diff names it, so no
  # PR selects it. The weekly cron exists for exactly that, and it already
  # ran — but a red cron blocks nothing and pages nobody, so it may as well
  # not have run. Measured: eight graphics members failed the same way for
  # weeks while `main` showed green, because the job that builds them was
  # `skipped` on every push.
  #
  # `main` is green and the packages install are two different claims. This
  # job turns the second one into an issue somebody sees.
  #
  # Only for full sweeps (schedule / manual dispatch). A PR's failure is
  # already visible on the PR.
  sweep-alert:
    name: full sweep failed — open an issue
    needs: [workspace, mirror-cn-reachable, red-members]
    if: >-
      always()
      && (needs.workspace.result == 'failure' || needs.mirror-cn-reachable.result == 'failure')
      && (github.event_name == 'schedule' || (github.event_name == 'workflow_dispatch' && inputs.update_pr == ''))
    runs-on: ubuntu-latest
    permissions:
      issues: write
    steps:
      - name: Open or update the sweep-failure issue
        env:
          GH_TOKEN: ${{ github.token }}
          # This job checks nothing out, so `gh` cannot infer the repository
          # from a work tree: on 2026-09-26 every `gh issue` call here failed
          # with "not a git repository", and the failing sweep opened no issue.
          GH_REPO: ${{ github.repository }}
          RUN_URL: ${{ github.server_url }}/${{ github.repository }}/actions/runs/${{ github.run_id }}
          FAILED_JOBS: >-
            ${{ needs.workspace.result == 'failure' && 'workspace' || '' }}
            ${{ needs.mirror-cn-reachable.result == 'failure' && 'mirror-cn-reachable' || '' }}
        run: |
          set -euo pipefail
          TITLE="full workspace sweep is failing"
          # One issue, reopened and appended to — not one per run. A tracker
          # that produces a new issue every week is a tracker people mute.
          NUM=$(gh issue list --state all --search "$TITLE in:title" \
                  --json number,title,state \
                  --jq "[.[] | select(.title == \"$TITLE\")] | first | .number // empty")
          BODY=$(printf '%s\n' \
            "The weekly full sweep failed: $RUN_URL" \
            "" \
            "This is the only signal that catches a published package broken by an" \
            "upstream change — selective CI cannot select a member no diff names." \
            "" \
            "Failed jobs: $FAILED_JOBS." \
            "" \
            "The \`red members name their issues\` job lists each failing member with" \
            "the issue that owns it (tests/known-red.tsv).")
          if [ -n "$NUM" ]; then
            gh issue reopen "$NUM" 2>/dev/null || true
            gh issue comment "$NUM" --body "$BODY"
            echo "commented on #$NUM"
          else
            gh issue create --title "$TITLE" --body "$BODY"
          fi