status: preserve clean-status proofs across mixed Git writers

.github/workflows/codex-pgo-training.sh

@@ -0,0 +1,60 @@
+#!/bin/sh
+
+# Keep this workload short and biased toward the local Git operations Codex
+# invokes frequently. The full Git test suite is too slow for each release
+# target and would weight test-harness paths more heavily than status, diff,
+# clone, fetch, and repository maintenance.
+
+set -eu
+
+git_bin="$PWD/bin-wrappers/git"
+training_dir=$(mktemp -d "${TMPDIR:-/tmp}/codex-git-pgo.XXXXXX")
+repo="$training_dir/repo"
+clone="$training_dir/clone"
+
+cleanup () {
+	rm -rf "$training_dir"
+}
+trap cleanup EXIT HUP INT TERM
+
+mkdir -p "$training_dir/home"
+export HOME="$training_dir/home"
+export GIT_CONFIG_NOSYSTEM=1
+export GIT_TERMINAL_PROMPT=0
+
+"$git_bin" clone --quiet --no-local "$PWD" "$repo"
+"$git_bin" -C "$repo" config user.name "Codex Git PGO"
+"$git_bin" -C "$repo" config user.email "codex-git-pgo@openai.com"
+
+i=0
+while test "$i" -lt 256
+do
+	dir="$repo/training/$((i % 16))"
+	mkdir -p "$dir"
+	printf '%s\n' "$i" >"$dir/file-$i"
+	i=$((i + 1))
+done
+
+"$git_bin" -C "$repo" status --porcelain=v2 --branch >/dev/null
+"$git_bin" -C "$repo" status --porcelain=v2 --branch --untracked-files=all >/dev/null
+"$git_bin" -C "$repo" ls-files --others --exclude-standard >/dev/null
+"$git_bin" -C "$repo" add training
+"$git_bin" -C "$repo" diff --cached --stat >/dev/null
+"$git_bin" -C "$repo" commit --quiet -m "add training files"
+
+printf 'changed\n' >>"$repo/training/0/file-0"
+rm "$repo/training/1/file-1"
+mkdir -p "$repo/untracked"
+printf 'new\n' >"$repo/untracked/file"
+
+"$git_bin" -C "$repo" status --porcelain=v2 --branch >/dev/null
+"$git_bin" -C "$repo" diff --stat >/dev/null
+"$git_bin" -C "$repo" diff --name-status >/dev/null
+"$git_bin" -C "$repo" ls-files --stage >/dev/null
+"$git_bin" -C "$repo" log --oneline --decorate -20 >/dev/null
+"$git_bin" -C "$repo" rev-list --objects --all >/dev/null
+"$git_bin" -C "$repo" for-each-ref --format='%(refname) %(objectname)' >/dev/null
+"$git_bin" -C "$repo" repack -ad
+"$git_bin" clone --quiet --no-local "$repo" "$clone"
+"$git_bin" -C "$clone" status --porcelain=v2 --branch >/dev/null
+"$git_bin" -C "$clone" fetch --quiet "$repo"

.github/workflows/codex-release.yml

@@ -104,7 +104,7 @@ jobs:
     name: ${{ matrix.name }}
     needs: version
     runs-on: ${{ matrix.os }}
-    timeout-minutes: 30
+    timeout-minutes: 60
     strategy:
       fail-fast: false
       matrix:
@@ -117,6 +117,9 @@ jobs:
             binary: /tmp/build/git/bin/git
             file_pattern: Mach-O 64-bit executable arm64
             has_gcm: false
+            lto: thin
+            profile_format: LLVM
+            llvm_profdata: xcrun llvm-profdata
             max_tar_bytes: 67108864
           - name: macOS x64
             os: macos-15-intel
@@ -126,15 +129,21 @@ jobs:
             binary: /tmp/build/git/bin/git
             file_pattern: Mach-O 64-bit executable x86_64
             has_gcm: false
+            lto: thin
+            profile_format: LLVM
+            llvm_profdata: xcrun llvm-profdata
             max_tar_bytes: 67108864
+          # Keep arm64 builds native so release smoke tests can execute them.
           - name: Linux arm64
-            os: ubuntu-22.04
+            os: ubuntu-22.04-arm
             target_platform: ubuntu
             asset_platform: ubuntu
             arch: arm64
             binary: /tmp/build/git/bin/git
             file_pattern: ELF 64-bit.*ARM aarch64
             has_gcm: false
+            lto: auto
+            profile_format: GCC
             max_tar_bytes: 67108864
           - name: Linux x64
             os: ubuntu-22.04
@@ -144,6 +153,8 @@ jobs:
             binary: /tmp/build/git/bin/git
             file_pattern: ELF 64-bit.*x86-64
             has_gcm: false
+            lto: auto
+            profile_format: GCC
             max_tar_bytes: 67108864
           - name: Windows arm64
             os: windows-11-arm
@@ -153,6 +164,9 @@ jobs:
             binary: /tmp/build/git/clangarm64/bin/git.exe
             file_pattern: PE32\+.*ARM64
             has_gcm: true
+            lto: thin
+            profile_format: LLVM
+            llvm_profdata: llvm-profdata
             max_tar_bytes: 134217728
             sdk_arch: aarch64
             sdk_flavor: full
@@ -169,6 +183,8 @@ jobs:
             binary: /tmp/build/git/mingw64/bin/git.exe
             file_pattern: PE32\+.*x86-64
             has_gcm: true
+            lto: auto
+            profile_format: GCC
             max_tar_bytes: 134217728
             sdk_arch: x86_64
             sdk_flavor: full
@@ -184,8 +200,8 @@ jobs:
       - name: Check out Dugite Native
         uses: actions/checkout@v6
         with:
-          repository: desktop/dugite-native
-          ref: f97e50add48cdcff053a69d95aa343a4a4a258c2
+          repository: dreynaud-oai/dugite-native
+          ref: b6f4473557acb85433fdf9deffe0854a34fd9cc5
           path: dugite-native
           fetch-depth: 0
           persist-credentials: false
@@ -209,6 +225,11 @@ jobs:
             -c 'user.email=41898282+github-actions[bot]@users.noreply.github.com' \
             tag -a "$VERSION" -m "$VERSION"
 
+      - name: Install OpenAI build configuration
+        shell: bash
+        working-directory: dugite-native/git
+        run: cp config.mak.openai config.mak
+
       # Match Dugite Native's compatibility choice for its macOS x64 build.
       - name: Select Xcode 16.4
         if: matrix.target_platform == 'macOS' && matrix.arch == 'x64'
@@ -243,19 +264,13 @@ jobs:
       - name: Install Linux arm64 build dependencies
         if: matrix.target_platform == 'ubuntu' && matrix.arch == 'arm64'
         run: |
-          sudo sed -i "s/^deb/deb [arch=amd64,i386]/g" /etc/apt/sources.list
-          release="$(lsb_release -s -c)"
-          echo "deb [arch=arm64,armhf] http://azure.ports.ubuntu.com/ ${release} main universe multiverse restricted" | sudo tee -a /etc/apt/sources.list
-          echo "deb [arch=arm64,armhf] http://azure.ports.ubuntu.com/ ${release}-updates main universe multiverse restricted" | sudo tee -a /etc/apt/sources.list
-          sudo dpkg --add-architecture arm64
-          sudo apt-get update
           sudo apt-get install -y \
             binutils-aarch64-linux-gnu \
             gcc-aarch64-linux-gnu \
-            libcurl4-gnutls-dev:arm64 \
-            libexpat1-dev:arm64 \
-            libssl-dev:arm64 \
-            zlib1g-dev:arm64
+            libcurl4-gnutls-dev \
+            libexpat1-dev \
+            libssl-dev \
+            zlib1g-dev
 
       # Dugite Native currently pins MinGit 2.53. Keep its build script and
       # dependency schema, but match the runtime to the Git series we compile.
@@ -303,13 +318,18 @@ jobs:
             dependencies.json >"$updated"
           mv "$updated" dependencies.json
 
-      # Dugite cross-compiles several targets. Keep Git's optional Rust
-      # library disabled until its Makefile can direct Cargo at those targets.
+      # Build an instrumented Git, run a representative local workload, then
+      # rebuild with its profile. Keep Git's optional Rust library disabled
+      # until its Makefile can direct Cargo at these targets.
       - name: Build the Dugite Native distribution
         shell: bash
         working-directory: dugite-native
         env:
           NO_RUST: 1
+          OPENAI_LLVM_PROFDATA: ${{ matrix.llvm_profdata }}
+          OPENAI_LTO: ${{ matrix.lto }}
+          OPENAI_PROFILE: BUILD
+          OPENAI_PROFILE_FORMAT: ${{ matrix.profile_format }}
           TARGET_PLATFORM: ${{ matrix.target_platform }}
           TARGET_ARCH: ${{ matrix.arch }}
         run: |
@@ -335,6 +355,9 @@ jobs:
         working-directory: dugite-native/git
         env:
           MINGW_DIR: ${{ matrix.mingw_dir }}
+          OPENAI_LLVM_PROFDATA: ${{ matrix.llvm_profdata }}
+          OPENAI_LTO: ${{ matrix.lto }}
+          OPENAI_PROFILE_FORMAT: ${{ matrix.profile_format }}
         run: |
           set -euo pipefail
 
@@ -349,8 +372,8 @@ jobs:
             SKIP_DASHED_BUILT_INS=YesPlease
           )
           jobs="${NUMBER_OF_PROCESSORS:-2}"
-          make -j"$jobs" "${make_args[@]}" all
-          make "${make_args[@]}" DESTDIR=/tmp/build/git strip install
+          make -j"$jobs" "${make_args[@]}" OPENAI_PROFILE=BUILD all
+          make "${make_args[@]}" OPENAI_PROFILE=USE DESTDIR=/tmp/build/git strip install
 
       - name: Verify distribution layout and provenance
         shell: bash
@@ -360,6 +383,8 @@ jobs:
           GIT_BINARY: ${{ matrix.binary }}
           FILE_PATTERN: ${{ matrix.file_pattern }}
           HAS_GCM: ${{ matrix.has_gcm }}
+          LTO: ${{ matrix.lto }}
+          PROFILE_FORMAT: ${{ matrix.profile_format }}
         run: |
           set -euo pipefail
           if test "$TARGET_PLATFORM" = win32
@@ -387,9 +412,15 @@ jobs:
           file "$GIT_BINARY" | tee /tmp/git-file-type
           grep -E "$FILE_PATTERN" /tmp/git-file-type
           strings "$GIT_BINARY" | grep -F "$GITHUB_SHA"
+          grep -F -- "-flto=$LTO" dugite-native/git/GIT-CFLAGS
+          if test "$PROFILE_FORMAT" = LLVM
+          then
+            grep -F -- "-fprofile-instr-use=" dugite-native/git/GIT-CFLAGS
+          else
+            grep -F -- "-fprofile-use=" dugite-native/git/GIT-CFLAGS
+          fi
 
       - name: Smoke-test the native distribution
-        if: matrix.arch == 'x64' || matrix.target_platform == 'win32'
         shell: bash
         env:
           TARGET_PLATFORM: ${{ matrix.target_platform }}

.github/workflows/codex.yml

@@ -0,0 +1,48 @@
+name: Refresh codex
+
+on:
+  workflow_dispatch:
+  pull_request:
+    branches:
+      - codex
+      - codex-unstable
+    types:
+      - opened
+      - reopened
+      - synchronize
+      - ready_for_review
+  merge_group:
+    types:
+      - checks_requested
+
+permissions:
+  actions: read
+  contents: read
+  pull-requests: read
+
+jobs:
+  refresh:
+    if: github.event_name == 'workflow_dispatch'
+    uses: openai/git/.github/workflows/codex.yml@meta
+  admission:
+    name: Codex admission
+    if: >-
+      (github.event_name == 'pull_request' &&
+       github.event.pull_request.base.ref == 'codex') ||
+      (github.event_name == 'merge_group' &&
+       github.event.merge_group.base_ref == 'refs/heads/codex')
+    permissions:
+      contents: read
+      pull-requests: read
+    uses: openai/git/.github/workflows/codex-admission.yml@meta
+  unstable_admission:
+    name: Codex unstable admission
+    if: >-
+      (github.event_name == 'pull_request' &&
+       github.event.pull_request.base.ref == 'codex-unstable') ||
+      (github.event_name == 'merge_group' &&
+       github.event.merge_group.base_ref == 'refs/heads/codex-unstable')
+    permissions:
+      contents: read
+      pull-requests: read
+    uses: openai/git/.github/workflows/codex-admission.yml@meta

Documentation/Makefile

@@ -142,6 +142,7 @@ TECH_DOCS += technical/send-pack-pipeline
 TECH_DOCS += technical/shallow
 TECH_DOCS += technical/sparse-checkout
 TECH_DOCS += technical/sparse-index
+TECH_DOCS += technical/status-clean-proof
 TECH_DOCS += technical/trivial-merge
 TECH_DOCS += technical/unambiguous-types
 TECH_DOCS += technical/unit-tests

Documentation/config/core.adoc

@@ -729,6 +729,17 @@ relatively high IO latencies.  When enabled, Git will do the
 index comparison to the filesystem data in parallel, allowing
 overlapping IO's.  Defaults to true.
 
+core.preloadIndexBulk::
+	On supported filesystems, scan working tree directories in bulk before
+	the parallel index preload.
++
+This replaces per-entry filesystem lookups with a physical directory scan,
+but may cost more than normal preload depending on filesystem and cache
+state. Inconclusive scans are discarded before continuing with the normal
+preload. Currently this is supported on APFS, ext-family filesystems, and
+XFS, and only has an effect when `core.preloadIndex` is enabled. Defaults
+to false.
+
 core.unsetenvvars::
 	Windows-only: comma-separated list of environment variables'
 	names that need to be unset before spawning any other process.

Documentation/gitformat-index.adoc

@@ -366,6 +366,49 @@ The remaining data of each directory block is grouped by type:
   - An ewah bitmap, the n-th bit indicates whether the n-th index entry
     is not CE_FSMONITOR_VALID.
 
+== File System Monitor untracked-cache token
+
+  The file system monitor untracked-cache token records the provider
+  token associated with an untracked-cache snapshot. The signature for
+  this extension is { 'F', 'S', 'U', 'C' }.
+
+  The extension consists of:
+
+  - 32-bit version number: the current version is 1.
+
+  - A NUL-terminated string containing the opaque file system monitor
+    token associated with the untracked-cache data.
+
+== File System Monitor semantic proof
+
+  The file system monitor semantic proof records the configuration and
+  attribute inputs for a completed worktree-content verification. Its
+  signature is { 'F', 'S', 'C', 'F' }.
+
+  The extension consists of:
+
+  - 32-bit version number (currently 1).
+
+  - 32-bit magic number identifying version 1 records (`FSC1`).
+
+  - 32-bit flags. The low four bits respectively indicate a complete
+    attribute manifest, a provider-token binding, a stat-data binding, and
+    coverage of the full index. All other bits must be zero.
+
+  - 32-bit length of the provider token.
+
+  - 32-bit length of the attribute manifest.
+
+  - The provider token, without a terminating NUL.
+
+  - Three hashes, using the index hash algorithm, over the relevant Git
+    configuration, semantic-conversion configuration, and attribute state.
+
+  - The attribute manifest described by its length above.
+
+  - A hash over all preceding bytes in this extension, using the index hash
+    algorithm.
+
 == End of Index Entry
 
   The End of Index Entry (EOIE) is used to locate the end of the variable

Documentation/technical/meson.build

@@ -31,6 +31,7 @@ articles = [
   'shallow.adoc',
   'sparse-checkout.adoc',
   'sparse-index.adoc',
+  'status-clean-proof.adoc',
   'trivial-merge.adoc',
   'unambiguous-types.adoc',
   'unit-tests.adoc',