Skip to content

Commit fcef56c

Browse files
dario-piotrowiczaduh95
dario-piotrowicz
authored and
aduh95
committed
fs: improve cpSync no-filter copyDir performance
move the logic in `cpSync` that copies a directory from src to dest from JavaScript to C++ increasing its performance Note: this improvement is not applied if the filter option is provided, such optimization will be looked into separately PR-URL: #58461 Reviewed-By: Yagiz Nizipli <yagiz@nizipli.com>
1 parent d6ea36a commit fcef56c

File tree

5 files changed

+249
-8
lines changed

5 files changed

+249
-8
lines changed

‎lib/internal/fs/cp/cp-sync.js

Lines changed: 21 additions & 8 deletions
Original file line numberDiff line numberDiff line change
@@ -133,18 +133,27 @@ function setDestTimestamps(src, dest) {
133133

134134
// TODO(@anonrig): Move this function to C++.
135135
function onDir(srcStat, destStat, src, dest, opts) {
136-
if (!destStat) return mkDirAndCopy(srcStat.mode, src, dest, opts);
136+
if (!destStat) return copyDir(src, dest, opts, true, srcStat.mode);
137137
return copyDir(src, dest, opts);
138138
}
139139

140-
function mkDirAndCopy(srcMode, src, dest, opts) {
141-
mkdirSync(dest);
142-
copyDir(src, dest, opts);
143-
return setDestMode(dest, srcMode);
144-
}
140+
function copyDir(src, dest, opts, mkDir, srcMode) {
141+
if (!opts.filter) {
142+
// The caller didn't provide a js filter function, in this case
143+
// we can run the whole function faster in C++
144+
// TODO(dario-piotrowicz): look into making cpSyncCopyDir also accept the potential filter function
145+
return fsBinding.cpSyncCopyDir(src, dest,
146+
opts.force,
147+
opts.dereference,
148+
opts.errorOnExist,
149+
opts.verbatimSymlinks,
150+
opts.preserveTimestamps);
151+
}
152+
153+
if (mkDir) {
154+
mkdirSync(dest);
155+
}
145156

146-
// TODO(@anonrig): Move this function to C++.
147-
function copyDir(src, dest, opts) {
148157
const dir = opendirSync(src);
149158

150159
try {
@@ -169,6 +178,10 @@ function copyDir(src, dest, opts) {
169178
}
170179
} finally {
171180
dir.closeSync();
181+
182+
if (srcMode !== undefined) {
183+
setDestMode(dest, srcMode);
184+
}
172185
}
173186
}
174187

‎src/node_errors.h

Lines changed: 2 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -78,7 +78,9 @@ void OOMErrorHandler(const char* location, const v8::OOMDetails& details);
7878
V(ERR_FS_CP_EINVAL, Error) \
7979
V(ERR_FS_CP_DIR_TO_NON_DIR, Error) \
8080
V(ERR_FS_CP_NON_DIR_TO_DIR, Error) \
81+
V(ERR_FS_CP_SYMLINK_TO_SUBDIRECTORY, Error) \
8182
V(ERR_FS_EISDIR, Error) \
83+
V(ERR_FS_CP_EEXIST, Error) \
8284
V(ERR_FS_CP_SOCKET, Error) \
8385
V(ERR_FS_CP_FIFO_PIPE, Error) \
8486
V(ERR_FS_CP_UNKNOWN, Error) \

‎src/node_file.cc

Lines changed: 221 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -43,6 +43,8 @@
4343
#include "uv.h"
4444
#include "v8-fast-api-calls.h"
4545

46+
#include <errno.h>
47+
#include <cerrno>
4648
#include <cstdio>
4749
#include <filesystem>
4850

@@ -3390,6 +3392,223 @@ static void CpSyncOverrideFile(const FunctionCallbackInfo<Value>& args) {
33903392
}
33913393
}
33923394

3395+
std::vector<std::string> normalizePathToArray(
3396+
const std::filesystem::path& path) {
3397+
std::vector<std::string> parts;
3398+
std::filesystem::path absPath = std::filesystem::absolute(path);
3399+
for (const auto& part : absPath) {
3400+
if (!part.empty()) parts.push_back(part.string());
3401+
}
3402+
return parts;
3403+
}
3404+
3405+
bool isInsideDir(const std::filesystem::path& src,
3406+
const std::filesystem::path& dest) {
3407+
auto srcArr = normalizePathToArray(src);
3408+
auto destArr = normalizePathToArray(dest);
3409+
if (srcArr.size() > destArr.size()) return false;
3410+
return std::equal(srcArr.begin(), srcArr.end(), destArr.begin());
3411+
}
3412+
3413+
static void CpSyncCopyDir(const FunctionCallbackInfo<Value>& args) {
3414+
CHECK_EQ(args.Length(), 7); // src, dest, force, dereference, errorOnExist,
3415+
// verbatimSymlinks, preserveTimestamps
3416+
3417+
Environment* env = Environment::GetCurrent(args);
3418+
Isolate* isolate = env->isolate();
3419+
3420+
BufferValue src(isolate, args[0]);
3421+
CHECK_NOT_NULL(*src);
3422+
ToNamespacedPath(env, &src);
3423+
3424+
BufferValue dest(isolate, args[1]);
3425+
CHECK_NOT_NULL(*dest);
3426+
ToNamespacedPath(env, &dest);
3427+
3428+
bool force = args[2]->IsTrue();
3429+
bool dereference = args[3]->IsTrue();
3430+
bool error_on_exist = args[4]->IsTrue();
3431+
bool verbatim_symlinks = args[5]->IsTrue();
3432+
bool preserve_timestamps = args[6]->IsTrue();
3433+
3434+
std::error_code error;
3435+
std::filesystem::create_directories(*dest, error);
3436+
if (error) {
3437+
return env->ThrowStdErrException(error, "cp", *dest);
3438+
}
3439+
3440+
auto file_copy_opts = std::filesystem::copy_options::recursive;
3441+
if (force) {
3442+
file_copy_opts |= std::filesystem::copy_options::overwrite_existing;
3443+
} else if (error_on_exist) {
3444+
file_copy_opts |= std::filesystem::copy_options::none;
3445+
} else {
3446+
file_copy_opts |= std::filesystem::copy_options::skip_existing;
3447+
}
3448+
3449+
std::function<bool(std::filesystem::path, std::filesystem::path)>
3450+
copy_dir_contents;
3451+
copy_dir_contents = [verbatim_symlinks,
3452+
&copy_dir_contents,
3453+
&env,
3454+
file_copy_opts,
3455+
preserve_timestamps,
3456+
force,
3457+
error_on_exist,
3458+
dereference,
3459+
&isolate](std::filesystem::path src,
3460+
std::filesystem::path dest) {
3461+
std::error_code error;
3462+
for (auto dir_entry : std::filesystem::directory_iterator(src)) {
3463+
auto dest_file_path = dest / dir_entry.path().filename();
3464+
auto dest_str = PathToString(dest);
3465+
3466+
if (dir_entry.is_symlink()) {
3467+
if (verbatim_symlinks) {
3468+
std::filesystem::copy_symlink(
3469+
dir_entry.path(), dest_file_path, error);
3470+
if (error) {
3471+
env->ThrowStdErrException(error, "cp", dest_str.c_str());
3472+
return false;
3473+
}
3474+
} else {
3475+
auto symlink_target =
3476+
std::filesystem::read_symlink(dir_entry.path().c_str(), error);
3477+
if (error) {
3478+
env->ThrowStdErrException(error, "cp", dest_str.c_str());
3479+
return false;
3480+
}
3481+
3482+
if (std::filesystem::exists(dest_file_path)) {
3483+
if (std::filesystem::is_symlink((dest_file_path.c_str()))) {
3484+
auto current_dest_symlink_target =
3485+
std::filesystem::read_symlink(dest_file_path.c_str(), error);
3486+
if (error) {
3487+
env->ThrowStdErrException(error, "cp", dest_str.c_str());
3488+
return false;
3489+
}
3490+
3491+
if (!dereference &&
3492+
std::filesystem::is_directory(symlink_target) &&
3493+
isInsideDir(symlink_target, current_dest_symlink_target)) {
3494+
std::string message =
3495+
"Cannot copy %s to a subdirectory of self %s";
3496+
THROW_ERR_FS_CP_EINVAL(env,
3497+
message.c_str(),
3498+
symlink_target.c_str(),
3499+
current_dest_symlink_target.c_str());
3500+
return false;
3501+
}
3502+
3503+
// Prevent copy if src is a subdir of dest since unlinking
3504+
// dest in this case would result in removing src contents
3505+
// and therefore a broken symlink would be created.
3506+
if (std::filesystem::is_directory(dest_file_path) &&
3507+
isInsideDir(current_dest_symlink_target, symlink_target)) {
3508+
std::string message = "cannot overwrite %s with %s";
3509+
THROW_ERR_FS_CP_SYMLINK_TO_SUBDIRECTORY(
3510+
env,
3511+
message.c_str(),
3512+
current_dest_symlink_target.c_str(),
3513+
symlink_target.c_str());
3514+
return false;
3515+
}
3516+
3517+
// symlinks get overridden by cp even if force: false, this is
3518+
// being applied here for backward compatibility, but is it
3519+
// correct? or is it a bug?
3520+
std::filesystem::remove(dest_file_path, error);
3521+
if (error) {
3522+
env->ThrowStdErrException(error, "cp", dest_str.c_str());
3523+
return false;
3524+
}
3525+
} else if (std::filesystem::is_regular_file(dest_file_path)) {
3526+
if (!dereference || (!force && error_on_exist)) {
3527+
auto dest_file_path_str = PathToString(dest_file_path);
3528+
env->ThrowStdErrException(
3529+
std::make_error_code(std::errc::file_exists),
3530+
"cp",
3531+
dest_file_path_str.c_str());
3532+
return false;
3533+
}
3534+
}
3535+
}
3536+
auto symlink_target_absolute = std::filesystem::weakly_canonical(
3537+
std::filesystem::absolute(src / symlink_target));
3538+
if (dir_entry.is_directory()) {
3539+
std::filesystem::create_directory_symlink(
3540+
symlink_target_absolute, dest_file_path, error);
3541+
} else {
3542+
std::filesystem::create_symlink(
3543+
symlink_target_absolute, dest_file_path, error);
3544+
}
3545+
if (error) {
3546+
env->ThrowStdErrException(error, "cp", dest_str.c_str());
3547+
return false;
3548+
}
3549+
}
3550+
} else if (dir_entry.is_directory()) {
3551+
auto entry_dir_path = src / dir_entry.path().filename();
3552+
std::filesystem::create_directory(dest_file_path);
3553+
auto success = copy_dir_contents(entry_dir_path, dest_file_path);
3554+
if (!success) {
3555+
return false;
3556+
}
3557+
} else if (dir_entry.is_regular_file()) {
3558+
std::filesystem::copy_file(
3559+
dir_entry.path(), dest_file_path, file_copy_opts, error);
3560+
if (error) {
3561+
if (error.value() == EEXIST) {
3562+
THROW_ERR_FS_CP_EEXIST(isolate,
3563+
"[ERR_FS_CP_EEXIST]: Target already exists: "
3564+
"cp returned EEXIST (%s already exists)",
3565+
dest_file_path.c_str());
3566+
return false;
3567+
}
3568+
env->ThrowStdErrException(error, "cp", dest_str.c_str());
3569+
return false;
3570+
}
3571+
3572+
if (preserve_timestamps) {
3573+
uv_fs_t req;
3574+
auto cleanup = OnScopeLeave([&req]() { uv_fs_req_cleanup(&req); });
3575+
3576+
auto dir_entry_path_str = PathToString(dir_entry.path());
3577+
int result =
3578+
uv_fs_stat(nullptr, &req, dir_entry_path_str.c_str(), nullptr);
3579+
if (is_uv_error(result)) {
3580+
env->ThrowUVException(
3581+
result, "stat", nullptr, dir_entry_path_str.c_str());
3582+
return false;
3583+
}
3584+
3585+
const uv_stat_t* const s = static_cast<const uv_stat_t*>(req.ptr);
3586+
const double source_atime =
3587+
s->st_atim.tv_sec + s->st_atim.tv_nsec / 1e9;
3588+
const double source_mtime =
3589+
s->st_mtim.tv_sec + s->st_mtim.tv_nsec / 1e9;
3590+
3591+
auto dest_file_path_str = PathToString(dest_file_path);
3592+
int utime_result = uv_fs_utime(nullptr,
3593+
&req,
3594+
dest_file_path_str.c_str(),
3595+
source_atime,
3596+
source_mtime,
3597+
nullptr);
3598+
if (is_uv_error(utime_result)) {
3599+
env->ThrowUVException(
3600+
utime_result, "utime", nullptr, dest_file_path_str.c_str());
3601+
return false;
3602+
}
3603+
}
3604+
}
3605+
}
3606+
return true;
3607+
};
3608+
3609+
copy_dir_contents(std::filesystem::path(*src), std::filesystem::path(*dest));
3610+
}
3611+
33933612
BindingData::FilePathIsFileReturnType BindingData::FilePathIsFile(
33943613
Environment* env, const std::string& file_path) {
33953614
THROW_IF_INSUFFICIENT_PERMISSIONS(
@@ -3726,6 +3945,7 @@ static void CreatePerIsolateProperties(IsolateData* isolate_data,
37263945

37273946
SetMethod(isolate, target, "cpSyncCheckPaths", CpSyncCheckPaths);
37283947
SetMethod(isolate, target, "cpSyncOverrideFile", CpSyncOverrideFile);
3948+
SetMethod(isolate, target, "cpSyncCopyDir", CpSyncCopyDir);
37293949

37303950
StatWatcher::CreatePerIsolateProperties(isolate_data, target);
37313951
BindingData::CreatePerIsolateProperties(isolate_data, target);
@@ -3837,6 +4057,7 @@ void RegisterExternalReferences(ExternalReferenceRegistry* registry) {
38374057

38384058
registry->Register(CpSyncCheckPaths);
38394059
registry->Register(CpSyncOverrideFile);
4060+
registry->Register(CpSyncCopyDir);
38404061

38414062
registry->Register(Chmod);
38424063
registry->Register(FChmod);

‎test/parallel/test-fs-cp.mjs

Lines changed: 3 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -521,6 +521,9 @@ if (!isWindows && !isInsideDirWithUnusualChars) {
521521
const dest = nextdir();
522522
mkdirSync(dest, mustNotMutateObjectDeep({ recursive: true }));
523523
writeFileSync(join(src, 'foo.txt'), 'foo', mustNotMutateObjectDeep({ mode: 0o444 }));
524+
// Small wait to make sure that destStat.mtime.getTime() would produce a time
525+
// different from srcStat.mtime.getTime() if preserveTimestamps wasn't set to true
526+
await setTimeout(5);
524527
cpSync(src, dest, mustNotMutateObjectDeep({ preserveTimestamps: true, recursive: true }));
525528
assertDirEquivalent(src, dest);
526529
const srcStat = lstatSync(join(src, 'foo.txt'));

‎typings/internalBinding/fs.d.ts

Lines changed: 2 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -78,6 +78,7 @@ declare namespace InternalFSBinding {
7878

7979
function cpSyncCheckPaths(src: StringOrBuffer, dest: StringOrBuffer, dereference: boolean, recursive: boolean): void;
8080
function cpSyncOverrideFile(src: StringOrBuffer, dest: StringOrBuffer, mode: number, preserveTimestamps: boolean): void;
81+
function cpSyncCopyDir(src: StringOrBuffer, dest: StringOrBuffer, force: boolean, errorOnExist: boolean, verbatimSymlinks: boolean, dereference: boolean): void;
8182

8283
function fchmod(fd: number, mode: number, req: FSReqCallback): void;
8384
function fchmod(fd: number, mode: number): void;
@@ -262,6 +263,7 @@ export interface FsBinding {
262263
copyFile: typeof InternalFSBinding.copyFile;
263264
cpSyncCheckPaths: typeof InternalFSBinding.cpSyncCheckPaths;
264265
cpSyncOverrideFile: typeof InternalFSBinding.cpSyncOverrideFile;
266+
cpSyncCopyDir: typeof InternalFSBinding.cpSyncCopyDir;
265267
fchmod: typeof InternalFSBinding.fchmod;
266268
fchown: typeof InternalFSBinding.fchown;
267269
fdatasync: typeof InternalFSBinding.fdatasync;

0 commit comments

Comments
 (0)