probe: O6 code page and O7 workspace profiles

Author: xlings-ci

.github/probe/codepage.sh

@@ -0,0 +1,47 @@
+#!/usr/bin/env bash
+# TEMPORARY PROBE (do not merge): why Glob.EscapedSpellingIsUtf8WhateverTheName
+# fails on some Windows machines. Prints; asserts nothing.
+set -u
+M="${MCPP_BUILT:?}"
+section() { echo; echo "### $1"; }
+REPO="$PWD"
+
+section "the runner's code pages"
+powershell -NoProfile -Command "[System.Text.Encoding]::Default.CodePage; (Get-ItemProperty HKLM:\SYSTEM\CurrentControlSet\Control\Nls\CodePage).ACP"
+
+section "the unit test as CI runs it"
+"$M" test test_modgraph 2>&1 | grep -E "EscapedSpelling|PASSED|FAILED|passed|failed|error" | head -20
+echo "-- does the test binary embed the UTF-8 activeCodePage manifest? (mcpp.exe for comparison)"
+for exe in $(find target -name 'test_modgraph*.exe' | head -1) $(find target -name 'mcpp.exe' -path '*/bin/*' | head -1); do
+  printf '%s: ' "$exe"; grep -a -c "activeCodePage" "$exe" || true
+done
+
+section "the same narrow-to-path conversion under each code page"
+W="$RUNNER_TEMP/codepage"; rm -rf "$W"; mkdir -p "$W"
+for v in legacy utf8; do
+  mkdir -p "$W/$v/src"; cd "$W/$v"
+  { printf '[package]\nname = "cp_%s"\nversion = "0.1.0"\n\n[toolchain]\nwindows = "msvc@system"\n\n[targets.cp_%s]\nkind = "bin"\nmain = "src/main.cpp"\n' "$v" "$v"
+    [ "$v" = utf8 ] && printf 'windows_code_page = "utf-8"\n'; } > mcpp.toml
+  cat > src/main.cpp <<'EOF'
+#include <windows.h>
+#include <cstdio>
+#include <exception>
+#include <filesystem>
+static void conv(unsigned cp) {
+    const int n = MultiByteToWideChar(cp, MB_ERR_INVALID_CHARS, "caf\xE9", -1, nullptr, 0);
+    std::printf("  MultiByteToWideChar(cp=%u, \"caf\\xE9\") -> %s\n", cp, n ? "ok" : "fails (no mapping)");
+}
+int main() {
+    std::printf("GetACP() = %u\n", GetACP());
+    try {
+        std::filesystem::path p("caf\xE9");
+        std::printf("  std::filesystem::path(\"caf\\xE9\") ok, %zu UTF-16 units\n", p.native().size());
+    } catch (const std::exception& e) {
+        std::printf("  std::filesystem::path(\"caf\\xE9\") throws: %s\n", e.what());
+    }
+    for (unsigned cp : {1252u, 936u, 65001u}) conv(cp);
+}
+EOF
+  echo "-- $v"; "$M" run 2>&1 | grep -v "^ *\(Resolved\|Target\|Inferred\|Compiling\|Finished\|Running\)" | head -12
+done
+cd "$REPO"

.github/probe/ws_profile.sh

@@ -0,0 +1,90 @@
+#!/usr/bin/env bash
+# TEMPORARY PROBE (do not merge): where [profile.*], [workspace.build] and a
+# rooted workspace's own [build] reach, per member and per selection.
+# Prints the markers each ninja edge carries; asserts nothing.
+set -u
+M="${MCPP_BUILT:?}"
+W="$RUNNER_TEMP/ws-profile"; rm -rf "$W"; mkdir -p "$W"
+section() { echo; echo "### $1"; }
+markers() {   # every compile/link edge and the markers it (or the file-level default) carries
+  local found=0
+  for n in $(find target -name build.ninja | sort); do found=1
+  echo "  [graph $(basename "$(dirname "$n")")]"
+  python3 - "$n" <<'PY'
+import re, sys
+glob, edges, cur = {}, {}, None
+for line in open(sys.argv[1], encoding="utf-8", errors="replace"):
+    line = line.rstrip("\n")
+    if line.startswith("build "):
+        cur = line.split(" : ")[0][6:].strip()
+        rule = line.split(" : ", 1)[1].split()[0] if " : " in line else "?"
+        edges[cur] = {"_rule": rule}; continue
+    if line.startswith("  ") and cur:
+        k, _, v = line.strip().partition(" = "); edges[cur][k] = v; continue
+    cur = None
+    m = re.match(r"^(\w+)\s*=\s*(.*)$", line)
+    if m: glob[m.group(1)] = m.group(2)
+pat = re.compile(r"MARK_[A-Z_]+|-O[0-3s]\b")
+for e, v in sorted(edges.items()):
+    r = v["_rule"]
+    if r in ("cxx_object", "cxx_module", "c_object"):
+        text = " ".join([glob.get("cxxflags", ""), v.get("unit_cxxflags", ""), v.get("cxxflags", ""), v.get("package_cxxflags", "")] + [x for k, x in v.items() if "flags" in k])
+    elif r in ("cxx_link", "cxx_shared", "c_link"):
+        text = v.get("ldflags", glob.get("ldflags", "")) + " " + v.get("unit_ldflags", "")
+    else:
+        continue
+    found = sorted(set(pat.findall(text)))
+    print(f"    {r:11s} {e}: {' '.join(found) or '-'}")
+PY
+  done
+  [ $found = 1 ] || echo "(no build.ninja)"
+}
+build() { echo "\$ mcpp $*"; rm -rf target; "$M" "$@" 2>&1 | grep -E "warning|error|note|Finished" | head -8; markers; }
+
+prof() {  # a [profile.release] table whose markers name its file
+  cat <<EOF
+[profile.release]
+opt = 3
+cxxflags = ["-DMARK_PROFILE_$1"]
+ldflags = ["-Wl,-rpath,/MARK_PROFILE_$1"]
+EOF
+}
+member() {  # member <dir> <name> <kind> [extra toml]
+  mkdir -p "$1/src"
+  { printf '[package]\nname = "%s"\nversion = "0.1.0"\n\n' "$2"
+    printf '[build]\ncxxflags = ["-DMARK_OWN_%s"]\nldflags = ["-Wl,-rpath,/MARK_OWN_%s"]\n\n' "${2^^}" "${2^^}"
+    [ "$3" = shared ] && printf '[targets.%s]\nkind = "shared"\n' "$2"
+    [ "$3" = bin ] && printf '[dependencies]\nlib = { path = "%s" }\n\n[targets.%s]\nkind = "bin"\nmain = "src/main.cpp"\n' "${DEP_PATH:-../lib}" "$2"
+    printf '%s\n' "${4:-}"; } > "$1/mcpp.toml"
+  if [ "$3" = shared ]; then printf 'export module ws_lib;\nexport int lv() { return 1; }\n' > "$1/src/lib.cppm"
+  else printf 'import ws_lib;\nint main() { return lv() == 1 ? 0 : 1; }\n' > "$1/src/main.cpp"; fi
+}
+
+section "W1 virtual root: [profile.release] + [workspace.build] at the root; members declare no profile"
+mkdir -p "$W/w1" && cd "$W/w1"
+{ printf '[workspace]\nmembers = ["app", "lib"]\n\n[workspace.build]\ncxxflags = ["-DMARK_WS_BUILD"]\nldflags = ["-Wl,-rpath,/MARK_WS_BUILD"]\n\n'; prof WSROOT; } > mcpp.toml
+member lib lib shared; member app app bin
+build build --release
+build build --release -p app
+
+section "W2 virtual root with a profile, and member app declares its own [profile.release]"
+mkdir -p "$W/w2" && cd "$W/w2"
+{ printf '[workspace]\nmembers = ["app", "lib"]\n\n'; prof WSROOT; } > mcpp.toml
+member lib lib shared; member app app bin "$(prof APP)"
+build build --release
+build build --release -p app
+build build --release -p lib
+
+section "W3 rooted workspace: the root file is [package] app + [workspace] + [profile.release] + its own [build]"
+mkdir -p "$W/w3" && cd "$W/w3"
+DEP_PATH=lib member . app bin "$(printf '[workspace]\nmembers = ["lib"]\n\n[workspace.build]\ncxxflags = ["-DMARK_WS_BUILD"]\n\n'; prof ROOTFILE)"
+member lib lib shared
+build build --release
+build build --release --workspace
+build build --release -p lib
+
+section "W4 a plain project (no workspace), for comparison"
+mkdir -p "$W/w4" && cd "$W/w4"
+member lib lib shared
+mkdir -p app && member app app bin "$(prof APP)"
+cd app && build build --release

.github/workflows/probe-770-771.yml

@@ -32,6 +32,7 @@ jobs:
           path: facts.log
 
   mcpp-probes:
+    if: false   # measured in run 37291652508; not repeated
     name: mcpp (built from this commit) on #771, #770, #769
     runs-on: windows-latest
     timeout-minutes: 60
@@ -76,3 +77,63 @@ jobs:
         with:
           name: mcpp-probes
           path: probes*.log
+
+  codepage:
+    name: O6 - Glob.EscapedSpellingIsUtf8WhateverTheName and the code page (windows)
+    runs-on: windows-latest
+    timeout-minutes: 60
+    env:
+      MCPP_HOME: C:\Users\runneradmin\.mcpp
+    steps:
+      - uses: actions/checkout@v4
+      - uses: ./.github/actions/bootstrap-mcpp
+      - name: Build mcpp from source
+        shell: bash
+        run: |
+          set -euo pipefail
+          export MCPP_VENDORED_XLINGS="$XLINGS_BIN"
+          "$XLINGS_BIN" config --mirror GLOBAL 2>/dev/null || true
+          "$MCPP" self config --mirror GLOBAL 2>/dev/null || true
+          "$MCPP" build
+          built=$(find target -type f -name mcpp.exe -path '*/bin/*' | head -1)
+          mkdir -p dist && cp "$built" dist/mcpp.exe
+          echo "MCPP_BUILT=$PWD/dist/mcpp.exe" >> "$GITHUB_ENV"
+      - name: probe
+        shell: bash
+        run: |
+          export MCPP_VENDORED_XLINGS="$XLINGS_BIN"
+          bash .github/probe/codepage.sh 2>&1 | tee codepage.log
+          { echo '## code page'; echo '```'; cat codepage.log; echo '```'; } >> "$GITHUB_STEP_SUMMARY"
+      - uses: actions/upload-artifact@v4
+        if: always()
+        with:
+          name: codepage
+          path: codepage.log
+
+  ws-profile:
+    name: O7 - profiles, [workspace.build] and the rooted root (linux)
+    runs-on: ubuntu-latest
+    timeout-minutes: 45
+    steps:
+      - uses: actions/checkout@v4
+      - uses: ./.github/actions/bootstrap-mcpp
+      - name: Build mcpp from source
+        shell: bash
+        run: |
+          set -euo pipefail
+          export MCPP_VENDORED_XLINGS="$XLINGS_BIN"
+          "$MCPP" build
+          built=$(find target -type f -name mcpp -path '*/bin/*' | grep -v '/dist/' | head -1)
+          mkdir -p dist && cp "$built" dist/mcpp
+          echo "MCPP_BUILT=$PWD/dist/mcpp" >> "$GITHUB_ENV"
+      - name: probe
+        shell: bash
+        run: |
+          export MCPP_VENDORED_XLINGS="$XLINGS_BIN"
+          bash .github/probe/ws_profile.sh 2>&1 | tee ws-profile.log
+          { echo '## workspace profiles'; echo '```'; cat ws-profile.log; echo '```'; } >> "$GITHUB_STEP_SUMMARY"
+      - uses: actions/upload-artifact@v4
+        if: always()
+        with:
+          name: ws-profile
+          path: ws-profile.log