agent/codex
2
0
Fork 0
mirror of https://github.com/pchuan98/codex.git synced 2026-07-01 00:31:56 +08:00
Code Issues Packages Projects Releases Wiki Activity

CI: Customize v8 building (#22086)

Browse Source 
## Summary

Move the rusty_v8 artifact production into hermetic Bazel path and bump
the `v8` crate to `147.4.0`

The new flow builds V8 release artifacts from source for Darwin and
Linux targets, publishes both the current release-compatible artifacts
and sandbox-enabled variants, and keeps Cargo consumers on prebuilt
binaries by continuing to feed the `v8` crate the archive and generated
binding files it already expects.

## Why

We need control over V8 build-time features without giving up prebuilt
artifacts for downstream Cargo builds.

Upstream `rusty_v8` already supports source-only features such as
`v8_enable_sandbox`, but its normal prebuilt release assets do not cover
every feature combination we need. Building the artifacts ourselves lets
us enable settings such as the V8 sandbox and pointer compression at
artifact build time, then publish those outputs so ordinary Cargo builds
can still consume prebuilts instead of compiling V8 locally.

This keeps the fast consumer experience of prebuilt `rusty_v8` archives
while giving us a reproducible path to ship featureful variants that
upstream does not currently publish for us.

## Implementation Notes

The Bazel graph in this PR is not copied wholesale from `rusty_v8`;
`rusty_v8`'s normal source build is still GN/Ninja-based.

Instead, this change starts from upstream V8's Bazel rules and adapts
them to Codex's hermetic toolchains and dependency layout. Where we
intentionally follow `rusty_v8`, we mirror its existing artifact
contract:

- the same `v8` crate version and generated binding expectations
- the same sandbox feature relationship, where sandboxing requires
pointer compression
- the same custom libc++ model expected by Cargo's default
`use_custom_libcxx` feature
- the same release-style archive plus `src_binding` outputs consumed by
the `v8` crate

To preserve that contract, the Bazel release path pins the libc++,
libc++abi, and llvm-libc revisions used by `rusty_v8 v147.4.0`, builds
release artifacts with `--config=rusty-v8-upstream-libcxx`, and folds
the matching runtime objects into the final static archive.

## Windows

Windows is annoyingly handled differently.

Codex's current hermetic Bazel Windows C++ platform is `windows-gnullvm`
/ `x86_64-w64-windows-gnu`, while upstream `rusty_v8` publishes Windows
prebuilts for `*-pc-windows-msvc`. Those are different ABIs, so the
Bazel graph cannot truthfully reproduce the upstream MSVC artifacts
until we add a real MSVC-targeting C++ toolchain.

For now:

- Windows MSVC consumers continue to use upstream `rusty_v8` release
archives.
- Windows GNU targets are built in-tree so they link against a matching
GNU ABI.
- The canary workflow separately exercises upstream `rusty_v8` source
builds for MSVC sandbox artifacts, but MSVC is not yet part of the
Bazel-produced release matrix.

## Validation
This PR is technically self validating through CI. I have already
published it as a release tag so the artifacts from this branch are
published to
https://github.com/openai/codex/releases/tag/rusty-v8-v147.4.0 CI for
this PR should therefore consume our own release targets. I have also
locally tested for linux and darwin.

---------

Co-authored-by: Codex <noreply@openai.com>
This commit is contained in:
Channing Conger
2026-05-18 21:33:05 -07:00
committed by GitHub
co-authored by Codex
parent a668379abf
commit 7cdeab33d1
29 changed files with 2173 additions and 454 deletions
Download Patch File Download Diff File Expand all files Collapse all files
.github/scripts/rusty_v8_bazel.py
+131 -100
View File
@@ -5,17 +5,18 @@ from __future__ import annotations
import argparse
import gzip
import hashlib
import os
import re
import shutil
import subprocess
import sys
import tempfile
import tomllib
from pathlib import Path
from rusty_v8_module_bazel import (
RustyV8ChecksumError,
check_module_bazel,
rusty_v8_http_file_versions,
update_module_bazel,
)
@@ -23,12 +24,16 @@ from rusty_v8_module_bazel import (
ROOT = Path(__file__).resolve().parents[2]
MODULE_BAZEL = ROOT / "MODULE.bazel"
RUSTY_V8_CHECKSUMS_DIR = ROOT / "third_party" / "v8"
MUSL_RUNTIME_ARCHIVE_LABELS = [
"@llvm//runtimes/libcxx:libcxx.static",
"@llvm//runtimes/libcxx:libcxxabi.static",
]
LLVM_AR_LABEL = "@llvm//tools:llvm-ar"
LLVM_RANLIB_LABEL = "@llvm//tools:llvm-ranlib"
RELEASE_ARTIFACT_PROFILE = "release"
SANDBOX_ARTIFACT_PROFILE = "ptrcomp_sandbox_release"
ARTIFACT_BAZEL_CONFIGS = ["rusty-v8-upstream-libcxx"]
def bazel_remote_args() -> list[str]:
buildbuddy_api_key = os.environ.get("BUILDBUDDY_API_KEY")
if not buildbuddy_api_key:
return []
return [f"--remote_header=x-buildbuddy-api-key={buildbuddy_api_key}"]
def bazel_execroot() -> Path:
@@ -75,6 +80,7 @@ def bazel_output_files(
compilation_mode,
f"--platforms=@llvm//platforms:{platform}",
*[f"--config={config}" for config in bazel_configs],
*bazel_remote_args(),
"--output=files",
expression,
],
@@ -91,8 +97,10 @@ def bazel_build(
labels: list[str],
compilation_mode: str = "fastbuild",
bazel_configs: list[str] | None = None,
download_toplevel: bool = False,
) -> None:
bazel_configs = bazel_configs or []
download_args = ["--remote_download_toplevel"] if download_toplevel else []
subprocess.run(
[
"bazel",
@@ -101,6 +109,8 @@ def bazel_build(
compilation_mode,
f"--platforms=@llvm//platforms:{platform}",
*[f"--config={config}" for config in bazel_configs],
*bazel_remote_args(),
*download_args,
*labels,
],
cwd=ROOT,
@@ -114,11 +124,15 @@ def ensure_bazel_output_files(
compilation_mode: str = "fastbuild",
bazel_configs: list[str] | None = None,
) -> list[Path]:
outputs = bazel_output_files(platform, labels, compilation_mode, bazel_configs)
if all(path.exists() for path in outputs):
return outputs
bazel_build(platform, labels, compilation_mode, bazel_configs)
# Bazel output paths can be reused across config flips, so existence alone
# does not prove the files match the requested flags.
bazel_build(
platform,
labels,
compilation_mode,
bazel_configs,
download_toplevel=True,
)
outputs = bazel_output_files(platform, labels, compilation_mode, bazel_configs)
missing = [str(path) for path in outputs if not path.exists()]
if missing:
@@ -126,9 +140,18 @@ def ensure_bazel_output_files(
return outputs
def release_pair_label(target: str) -> str:
def artifact_bazel_configs(bazel_configs: list[str] | None = None) -> list[str]:
configured = list(ARTIFACT_BAZEL_CONFIGS)
for config in bazel_configs or []:
if config not in configured:
configured.append(config)
return configured
def release_pair_label(target: str, sandbox: bool = False) -> str:
target_suffix = target.replace("-", "_")
return f"//third_party/v8:rusty_v8_release_pair_{target_suffix}"
pair_kind = "sandbox_release_pair" if sandbox else "release_pair"
return f"//third_party/v8:rusty_v8_{pair_kind}_{target_suffix}"
def resolved_v8_crate_version() -> str:
@@ -169,6 +192,16 @@ def rusty_v8_checksum_manifest_path(version: str) -> Path:
def command_version(version: str | None) -> str:
if version is not None:
return version
manifest_versions = rusty_v8_http_file_versions(MODULE_BAZEL.read_text())
if len(manifest_versions) == 1:
return manifest_versions[0]
if len(manifest_versions) > 1:
raise SystemExit(
"expected at most one rusty_v8 http_file version in MODULE.bazel, "
f"found: {manifest_versions}; pass --version explicitly"
)
return resolved_v8_crate_version()
@@ -180,66 +213,76 @@ def command_manifest_path(manifest: Path | None, version: str) -> Path:
return ROOT / manifest
def staged_archive_name(target: str, source_path: Path) -> str:
if source_path.suffix == ".lib":
return f"rusty_v8_release_{target}.lib.gz"
return f"librusty_v8_release_{target}.a.gz"
def staged_archive_name(target: str, source_path: Path, artifact_profile: str) -> str:
if target.endswith("-pc-windows-msvc"):
return f"rusty_v8_{artifact_profile}_{target}.lib.gz"
return f"librusty_v8_{artifact_profile}_{target}.a.gz"
def is_musl_archive_target(target: str, source_path: Path) -> bool:
return target.endswith("-unknown-linux-musl") and source_path.suffix == ".a"
def staged_binding_name(target: str, artifact_profile: str) -> str:
return f"src_binding_{artifact_profile}_{target}.rs"
def single_bazel_output_file(
platform: str,
label: str,
compilation_mode: str = "fastbuild",
bazel_configs: list[str] | None = None,
) -> Path:
outputs = ensure_bazel_output_files(platform, [label], compilation_mode, bazel_configs)
if len(outputs) != 1:
raise SystemExit(f"expected exactly one output for {label}, found {outputs}")
return outputs[0]
def staged_checksums_name(target: str, artifact_profile: str) -> str:
return f"rusty_v8_{artifact_profile}_{target}.sha256"
def merged_musl_archive(
platform: str,
def stage_artifacts(
target: str,
lib_path: Path,
compilation_mode: str = "fastbuild",
bazel_configs: list[str] | None = None,
) -> Path:
llvm_ar = single_bazel_output_file(platform, LLVM_AR_LABEL, compilation_mode, bazel_configs)
llvm_ranlib = single_bazel_output_file(
platform,
LLVM_RANLIB_LABEL,
compilation_mode,
bazel_configs,
)
runtime_archives = [
single_bazel_output_file(platform, label, compilation_mode, bazel_configs)
for label in MUSL_RUNTIME_ARCHIVE_LABELS
]
binding_path: Path,
output_dir: Path,
sandbox: bool,
) -> None:
missing_paths = [str(path) for path in [lib_path, binding_path] if not path.exists()]
if missing_paths:
raise SystemExit(f"missing release outputs for {target}: {missing_paths}")
temp_dir = Path(tempfile.mkdtemp(prefix="rusty-v8-musl-stage-"))
merged_archive = temp_dir / lib_path.name
merge_commands = "\n".join(
[
f"create {merged_archive}",
f"addlib {lib_path}",
*[f"addlib {archive}" for archive in runtime_archives],
"save",
"end",
]
)
subprocess.run(
[str(llvm_ar), "-M"],
cwd=ROOT,
check=True,
input=merge_commands,
text=True,
)
subprocess.run([str(llvm_ranlib), str(merged_archive)], cwd=ROOT, check=True)
return merged_archive
output_dir.mkdir(parents=True, exist_ok=True)
artifact_profile = SANDBOX_ARTIFACT_PROFILE if sandbox else RELEASE_ARTIFACT_PROFILE
staged_library = output_dir / staged_archive_name(target, lib_path, artifact_profile)
staged_binding = output_dir / staged_binding_name(target, artifact_profile)
with lib_path.open("rb") as src, staged_library.open("wb") as dst:
with gzip.GzipFile(
filename="",
mode="wb",
fileobj=dst,
compresslevel=6,
mtime=0,
) as gz:
shutil.copyfileobj(src, gz)
shutil.copyfile(binding_path, staged_binding)
staged_checksums = output_dir / staged_checksums_name(target, artifact_profile)
with staged_checksums.open("w", encoding="utf-8") as checksums:
for path in [staged_library, staged_binding]:
digest = hashlib.sha256()
with path.open("rb") as artifact:
for chunk in iter(lambda: artifact.read(1024 * 1024), b""):
digest.update(chunk)
checksums.write(f"{digest.hexdigest()} {path.name}\n")
print(staged_library)
print(staged_binding)
print(staged_checksums)
def upstream_release_pair_paths(source_root: Path, target: str) -> tuple[Path, Path]:
lib_name = "rusty_v8.lib" if target.endswith("-pc-windows-msvc") else "librusty_v8.a"
gn_out = source_root / "target" / target / "release" / "gn_out"
return gn_out / "obj" / lib_name, gn_out / "src_binding.rs"
def stage_upstream_release_pair(
source_root: Path,
target: str,
output_dir: Path,
sandbox: bool = False,
) -> None:
lib_path, binding_path = upstream_release_pair_paths(source_root, target)
stage_artifacts(target, lib_path, binding_path, output_dir, sandbox)
def stage_release_pair(
@@ -248,10 +291,12 @@ def stage_release_pair(
output_dir: Path,
compilation_mode: str = "fastbuild",
bazel_configs: list[str] | None = None,
sandbox: bool = False,
) -> None:
bazel_configs = artifact_bazel_configs(bazel_configs)
outputs = ensure_bazel_output_files(
platform,
[release_pair_label(target)],
[release_pair_label(target, sandbox)],
compilation_mode,
bazel_configs,
)
@@ -266,39 +311,7 @@ def stage_release_pair(
except StopIteration as exc:
raise SystemExit(f"missing Rust binding output for {target}") from exc
output_dir.mkdir(parents=True, exist_ok=True)
staged_library = output_dir / staged_archive_name(target, lib_path)
staged_binding = output_dir / f"src_binding_release_{target}.rs"
source_archive = (
merged_musl_archive(platform, lib_path, compilation_mode, bazel_configs)
if is_musl_archive_target(target, lib_path)
else lib_path
)
with source_archive.open("rb") as src, staged_library.open("wb") as dst:
with gzip.GzipFile(
filename="",
mode="wb",
fileobj=dst,
compresslevel=6,
mtime=0,
) as gz:
shutil.copyfileobj(src, gz)
shutil.copyfile(binding_path, staged_binding)
staged_checksums = output_dir / f"rusty_v8_release_{target}.sha256"
with staged_checksums.open("w", encoding="utf-8") as checksums:
for path in [staged_library, staged_binding]:
digest = hashlib.sha256()
with path.open("rb") as artifact:
for chunk in iter(lambda: artifact.read(1024 * 1024), b""):
digest.update(chunk)
checksums.write(f"{digest.hexdigest()} {path.name}\n")
print(staged_library)
print(staged_binding)
print(staged_checksums)
stage_artifacts(target, lib_path, binding_path, output_dir, sandbox)
def parse_args() -> argparse.Namespace:
@@ -309,6 +322,7 @@ def parse_args() -> argparse.Namespace:
stage_release_pair_parser.add_argument("--platform", required=True)
stage_release_pair_parser.add_argument("--target", required=True)
stage_release_pair_parser.add_argument("--output-dir", required=True)
stage_release_pair_parser.add_argument("--sandbox", action="store_true")
stage_release_pair_parser.add_argument(
"--bazel-config",
action="append",
@@ -321,6 +335,14 @@ def parse_args() -> argparse.Namespace:
choices=["fastbuild", "opt", "dbg"],
)
stage_upstream_release_pair_parser = subparsers.add_parser(
"stage-upstream-release-pair"
)
stage_upstream_release_pair_parser.add_argument("--source-root", type=Path, required=True)
stage_upstream_release_pair_parser.add_argument("--target", required=True)
stage_upstream_release_pair_parser.add_argument("--output-dir", required=True)
stage_upstream_release_pair_parser.add_argument("--sandbox", action="store_true")
subparsers.add_parser("resolved-v8-crate-version")
check_module_bazel_parser = subparsers.add_parser("check-module-bazel")
@@ -353,6 +375,15 @@ def main() -> int:
output_dir=Path(args.output_dir),
compilation_mode=args.compilation_mode,
bazel_configs=args.bazel_configs,
sandbox=args.sandbox,
)
return 0
if args.command == "stage-upstream-release-pair":
stage_upstream_release_pair(
source_root=args.source_root,
target=args.target,
output_dir=Path(args.output_dir),
sandbox=args.sandbox,
)
return 0
if args.command == "resolved-v8-crate-version":