llvm-project/libcxx/test/support/filesystem_test_helper.h

Ignoring revisions in .git-blame-ignore-revs. Click here to bypass and see the normal blame view.

529 lines
16 KiB
C
Raw Normal View History

#ifndef FILESYSTEM_TEST_HELPER_H
#define FILESYSTEM_TEST_HELPER_H
#include "filesystem_include.h"
#include <unistd.h> // for ftruncate
#include <cassert>
#include <cstdio> // for printf
#include <string>
#include <fstream>
#include <random>
#include <chrono>
[libc++] Implement Directory Entry Caching -- Sort of. Summary: This patch implements directory_entry caching *almost* as specified in P0317r1. However, I explicitly chose to deviate from the standard as I'll explain below. The approach I decided to take is a fully caching one. When `refresh()` is called, the cache is populated by calls to `stat` and `lstat` as needed. During directory iteration the cache is only populated with the `file_type` as reported by `readdir`. The cache can be in the following states: * `_Empty`: There is nothing in the cache (likely due to an error) * `_IterSymlink`: Created by directory iteration when we walk onto a symlink only the symlink file type is known. * `_IterNonSymlink`: Created by directory iteration when we walk onto a non-symlink. Both the regular file type and symlink file type are known. * `_RefreshSymlink` and `_RefreshNonSymlink`: A full cache created by `refresh()`. This case includes dead symlinks. * `_RefreshSymlinkUnresolved`: A partial cache created by refresh when we fail to resolve the file pointed to by a symlink (likely due to permissions). Symlink attributes are cached, but attributes about the linked entity are not. As mentioned, this implementation purposefully deviates from the standard. According to some readings of the specification, and the Windows filesystem implementation, the constructors and modifiers which don't pass an `error_code` must throw when the `directory_entry` points to a entity which doesn't exist. or when attribute resolution fails for another reason. @BillyONeal has proposed a more reasonable set of requirements, where modifiers other than refresh ignore errors. This is the behavior libc++ currently implements, with the expectation some form of the new language will be accepted into the standard. Some additional semantics which differ from the Windows implementation: 1. `refresh` will not throw when the entry doesn't exist. In this case we can still meet the functions specification, so we don't treat it as an error. 2. We don't clear the path name when a constructor fails via refresh (this will hopefully be changed in the standard as well). It should be noted that libstdc++'s current implementation has the same behavior as libc++, except for point (2). If the changes to the specification don't get accepted, we'll be able to make the changes later. [1] http://www.open-std.org/jtc1/sc22/wg21/docs/papers/2016/p0317r1.html Reviewers: mclow.lists, gromer, ldionne, aaron.ballman Subscribers: BillyONeal, christof, cfe-commits Differential Revision: https://reviews.llvm.org/D49530 llvm-svn: 337516
2018-07-20 09:22:32 +08:00
#include <vector>
#include <regex>
[libc++] Implement Directory Entry Caching -- Sort of. Summary: This patch implements directory_entry caching *almost* as specified in P0317r1. However, I explicitly chose to deviate from the standard as I'll explain below. The approach I decided to take is a fully caching one. When `refresh()` is called, the cache is populated by calls to `stat` and `lstat` as needed. During directory iteration the cache is only populated with the `file_type` as reported by `readdir`. The cache can be in the following states: * `_Empty`: There is nothing in the cache (likely due to an error) * `_IterSymlink`: Created by directory iteration when we walk onto a symlink only the symlink file type is known. * `_IterNonSymlink`: Created by directory iteration when we walk onto a non-symlink. Both the regular file type and symlink file type are known. * `_RefreshSymlink` and `_RefreshNonSymlink`: A full cache created by `refresh()`. This case includes dead symlinks. * `_RefreshSymlinkUnresolved`: A partial cache created by refresh when we fail to resolve the file pointed to by a symlink (likely due to permissions). Symlink attributes are cached, but attributes about the linked entity are not. As mentioned, this implementation purposefully deviates from the standard. According to some readings of the specification, and the Windows filesystem implementation, the constructors and modifiers which don't pass an `error_code` must throw when the `directory_entry` points to a entity which doesn't exist. or when attribute resolution fails for another reason. @BillyONeal has proposed a more reasonable set of requirements, where modifiers other than refresh ignore errors. This is the behavior libc++ currently implements, with the expectation some form of the new language will be accepted into the standard. Some additional semantics which differ from the Windows implementation: 1. `refresh` will not throw when the entry doesn't exist. In this case we can still meet the functions specification, so we don't treat it as an error. 2. We don't clear the path name when a constructor fails via refresh (this will hopefully be changed in the standard as well). It should be noted that libstdc++'s current implementation has the same behavior as libc++, except for point (2). If the changes to the specification don't get accepted, we'll be able to make the changes later. [1] http://www.open-std.org/jtc1/sc22/wg21/docs/papers/2016/p0317r1.html Reviewers: mclow.lists, gromer, ldionne, aaron.ballman Subscribers: BillyONeal, christof, cfe-commits Differential Revision: https://reviews.llvm.org/D49530 llvm-svn: 337516
2018-07-20 09:22:32 +08:00
#include "test_macros.h"
#include "rapid-cxx-test.h"
#include "format_string.h"
// For creating socket files
#if !defined(__FreeBSD__) && !defined(__APPLE__)
# include <sys/socket.h>
# include <sys/un.h>
#endif
// static test helpers
namespace StaticEnv {
// Tests that use these utilities should add '<...>/Inputs/static_test_env'
// to their FILE_DEPENDENCIES, to make sure the directory is made available
// to the test. Assuming that, the 'static_test_env' will be available in the
// directory where the test is run.
static const fs::path Root = fs::current_path() / "static_test_env";
inline fs::path makePath(fs::path const& p) {
// env_path is expected not to contain symlinks.
fs::path const& env_path = Root;
return env_path / p;
}
static const fs::path TestFileList[] = {
makePath("empty_file"),
makePath("non_empty_file"),
makePath("dir1/file1"),
makePath("dir1/file2")
};
const std::size_t TestFileListSize = sizeof(TestFileList) / sizeof(fs::path);
static const fs::path TestDirList[] = {
makePath("dir1"),
makePath("dir1/dir2"),
makePath("dir1/dir2/dir3")
};
const std::size_t TestDirListSize = sizeof(TestDirList) / sizeof(fs::path);
static const fs::path File = TestFileList[0];
static const fs::path Dir = TestDirList[0];
static const fs::path Dir2 = TestDirList[1];
static const fs::path Dir3 = TestDirList[2];
static const fs::path SymlinkToFile = makePath("symlink_to_empty_file");
static const fs::path SymlinkToDir = makePath("symlink_to_dir");
static const fs::path BadSymlink = makePath("bad_symlink");
static const fs::path DNE = makePath("DNE");
static const fs::path EmptyFile = TestFileList[0];
static const fs::path NonEmptyFile = TestFileList[1];
static const fs::path CharFile = "/dev/null"; // Hopefully this exists
static const fs::path DirIterationList[] = {
makePath("dir1/dir2"),
makePath("dir1/file1"),
makePath("dir1/file2")
};
const std::size_t DirIterationListSize = sizeof(DirIterationList)
/ sizeof(fs::path);
static const fs::path DirIterationListDepth1[] = {
makePath("dir1/dir2/afile3"),
makePath("dir1/dir2/dir3"),
makePath("dir1/dir2/symlink_to_dir3"),
makePath("dir1/dir2/file4"),
};
static const fs::path RecDirIterationList[] = {
makePath("dir1/dir2"),
makePath("dir1/file1"),
makePath("dir1/file2"),
makePath("dir1/dir2/afile3"),
makePath("dir1/dir2/dir3"),
makePath("dir1/dir2/symlink_to_dir3"),
makePath("dir1/dir2/file4"),
makePath("dir1/dir2/dir3/file5")
};
static const fs::path RecDirFollowSymlinksIterationList[] = {
makePath("dir1/dir2"),
makePath("dir1/file1"),
makePath("dir1/file2"),
makePath("dir1/dir2/afile3"),
makePath("dir1/dir2/dir3"),
makePath("dir1/dir2/file4"),
makePath("dir1/dir2/dir3/file5"),
makePath("dir1/dir2/symlink_to_dir3"),
makePath("dir1/dir2/symlink_to_dir3/file5"),
};
} // namespace StaticEnv
namespace random_utils {
inline char to_hex(int ch) {
return ch < 10 ? static_cast<char>('0' + ch)
: static_cast<char>('a' + (ch - 10));
}
inline char random_hex_char() {
static std::mt19937 rd{std::random_device{}()};
static std::uniform_int_distribution<int> mrand{0, 15};
return to_hex(mrand(rd));
}
} // namespace random_utils
struct scoped_test_env
{
scoped_test_env() : test_root(random_path()) {
std::string cmd = "mkdir -p " + test_root.native();
int ret = std::system(cmd.c_str());
assert(ret == 0);
// Ensure that the root_path is fully resolved, i.e. it contains no
// symlinks. The filesystem tests depend on that. We do this after
// creating the root_path, because `fs::canonical` requires the
// path to exist.
test_root = fs::canonical(test_root);
}
~scoped_test_env() {
std::string cmd = "chmod -R 777 " + test_root.native();
int ret = std::system(cmd.c_str());
assert(ret == 0);
cmd = "rm -r " + test_root.native();
ret = std::system(cmd.c_str());
assert(ret == 0);
}
scoped_test_env(scoped_test_env const &) = delete;
scoped_test_env & operator=(scoped_test_env const &) = delete;
fs::path make_env_path(std::string p) { return sanitize_path(p); }
std::string sanitize_path(std::string raw) {
assert(raw.find("..") == std::string::npos);
std::string const& root = test_root.native();
if (root.compare(0, root.size(), raw, 0, root.size()) != 0) {
assert(raw.front() != '\\');
fs::path tmp(test_root);
tmp /= raw;
return std::move(const_cast<std::string&>(tmp.native()));
}
return raw;
}
// Purposefully using a size potentially larger than off_t here so we can
// test the behavior of libc++fs when it is built with _FILE_OFFSET_BITS=64
// but the caller is not (std::filesystem also uses uintmax_t rather than
// off_t). On a 32-bit system this allows us to create a file larger than
// 2GB.
std::string create_file(std::string filename, uintmax_t size = 0) {
#if defined(__LP64__)
auto large_file_fopen = fopen;
auto large_file_ftruncate = ftruncate;
using large_file_offset_t = off_t;
#else
auto large_file_fopen = fopen64;
auto large_file_ftruncate = ftruncate64;
using large_file_offset_t = off64_t;
#endif
filename = sanitize_path(std::move(filename));
if (size > std::numeric_limits<large_file_offset_t>::max()) {
fprintf(stderr, "create_file(%s, %ju) too large\n",
filename.c_str(), size);
abort();
}
FILE* file = large_file_fopen(filename.c_str(), "we");
if (file == nullptr) {
fprintf(stderr, "fopen %s failed: %s\n", filename.c_str(),
strerror(errno));
abort();
}
if (large_file_ftruncate(
fileno(file), static_cast<large_file_offset_t>(size)) == -1) {
fprintf(stderr, "ftruncate %s %ju failed: %s\n", filename.c_str(),
size, strerror(errno));
fclose(file);
abort();
}
fclose(file);
return filename;
}
std::string create_dir(std::string filename) {
filename = sanitize_path(std::move(filename));
std::string cmd = "mkdir " + filename;
int ret = std::system(cmd.c_str());
assert(ret == 0);
return filename;
}
std::string create_symlink(std::string source, std::string to) {
source = sanitize_path(std::move(source));
to = sanitize_path(std::move(to));
std::string cmd = "ln -s " + source + ' ' + to;
int ret = std::system(cmd.c_str());
assert(ret == 0);
return to;
}
std::string create_hardlink(std::string source, std::string to) {
source = sanitize_path(std::move(source));
to = sanitize_path(std::move(to));
std::string cmd = "ln " + source + ' ' + to;
int ret = std::system(cmd.c_str());
assert(ret == 0);
return to;
}
std::string create_fifo(std::string file) {
file = sanitize_path(std::move(file));
std::string cmd = "mkfifo " + file;
int ret = std::system(cmd.c_str());
assert(ret == 0);
return file;
}
// OS X and FreeBSD doesn't support socket files so we shouldn't even
// allow tests to call this unguarded.
#if !defined(__FreeBSD__) && !defined(__APPLE__)
std::string create_socket(std::string file) {
file = sanitize_path(std::move(file));
::sockaddr_un address;
address.sun_family = AF_UNIX;
assert(file.size() <= sizeof(address.sun_path));
::strncpy(address.sun_path, file.c_str(), sizeof(address.sun_path));
int fd = ::socket(AF_UNIX, SOCK_STREAM, 0);
::bind(fd, reinterpret_cast<::sockaddr*>(&address), sizeof(address));
return file;
}
#endif
fs::path test_root;
private:
static std::string unique_path_suffix() {
std::string model = "test.%%%%%%";
for (auto & ch : model) {
if (ch == '%')
ch = random_utils::random_hex_char();
}
return model;
}
// This could potentially introduce a filesystem race with other tests
// running at the same time, but oh well, it's just test code.
static inline fs::path random_path() {
fs::path tmp = fs::temp_directory_path();
fs::path p = fs::path(tmp) / unique_path_suffix();
return p;
}
};
// Misc test types
#define MKSTR(Str) {Str, TEST_CONCAT(L, Str), TEST_CONCAT(u, Str), TEST_CONCAT(U, Str)}
struct MultiStringType {
const char* s;
const wchar_t* w;
const char16_t* u16;
const char32_t* u32;
operator const char* () const { return s; }
operator const wchar_t* () const { return w; }
operator const char16_t* () const { return u16; }
operator const char32_t* () const { return u32; }
};
const MultiStringType PathList[] = {
MKSTR(""),
MKSTR(" "),
MKSTR("//"),
MKSTR("."),
MKSTR(".."),
MKSTR("foo"),
MKSTR("/"),
MKSTR("/foo"),
MKSTR("foo/"),
MKSTR("/foo/"),
MKSTR("foo/bar"),
MKSTR("/foo/bar"),
MKSTR("//net"),
MKSTR("//net/foo"),
MKSTR("///foo///"),
MKSTR("///foo///bar"),
MKSTR("/."),
MKSTR("./"),
MKSTR("/.."),
MKSTR("../"),
MKSTR("foo/."),
MKSTR("foo/.."),
MKSTR("foo/./"),
MKSTR("foo/./bar"),
MKSTR("foo/../"),
MKSTR("foo/../bar"),
MKSTR("c:"),
MKSTR("c:/"),
MKSTR("c:foo"),
MKSTR("c:/foo"),
MKSTR("c:foo/"),
MKSTR("c:/foo/"),
MKSTR("c:/foo/bar"),
MKSTR("prn:"),
MKSTR("c:\\"),
MKSTR("c:\\foo"),
MKSTR("c:foo\\"),
MKSTR("c:\\foo\\"),
MKSTR("c:\\foo/"),
MKSTR("c:/foo\\bar"),
MKSTR("//"),
MKSTR("/finally/we/need/one/really/really/really/really/really/really/really/long/string")
};
const unsigned PathListSize = sizeof(PathList) / sizeof(MultiStringType);
template <class Iter>
Iter IterEnd(Iter B) {
using VT = typename std::iterator_traits<Iter>::value_type;
for (; *B != VT{}; ++B)
;
return B;
}
template <class CharT>
const CharT* StrEnd(CharT const* P) {
return IterEnd(P);
}
template <class CharT>
std::size_t StrLen(CharT const* P) {
return StrEnd(P) - P;
}
// Testing the allocation behavior of the code_cvt functions requires
// *knowing* that the allocation was not done by "path::__str_".
// This hack forces path to allocate enough memory.
inline void PathReserve(fs::path& p, std::size_t N) {
auto const& native_ref = p.native();
const_cast<std::string&>(native_ref).reserve(N);
}
template <class Iter1, class Iter2>
bool checkCollectionsEqual(
Iter1 start1, Iter1 const end1
, Iter2 start2, Iter2 const end2
)
{
while (start1 != end1 && start2 != end2) {
if (*start1 != *start2) {
return false;
}
++start1; ++start2;
}
return (start1 == end1 && start2 == end2);
}
template <class Iter1, class Iter2>
bool checkCollectionsEqualBackwards(
Iter1 const start1, Iter1 end1
, Iter2 const start2, Iter2 end2
)
{
while (start1 != end1 && start2 != end2) {
--end1; --end2;
if (*end1 != *end2) {
return false;
}
}
return (start1 == end1 && start2 == end2);
}
// We often need to test that the error_code was cleared if no error occurs
// this function returns an error_code which is set to an error that will
// never be returned by the filesystem functions.
[libc++] Implement Directory Entry Caching -- Sort of. Summary: This patch implements directory_entry caching *almost* as specified in P0317r1. However, I explicitly chose to deviate from the standard as I'll explain below. The approach I decided to take is a fully caching one. When `refresh()` is called, the cache is populated by calls to `stat` and `lstat` as needed. During directory iteration the cache is only populated with the `file_type` as reported by `readdir`. The cache can be in the following states: * `_Empty`: There is nothing in the cache (likely due to an error) * `_IterSymlink`: Created by directory iteration when we walk onto a symlink only the symlink file type is known. * `_IterNonSymlink`: Created by directory iteration when we walk onto a non-symlink. Both the regular file type and symlink file type are known. * `_RefreshSymlink` and `_RefreshNonSymlink`: A full cache created by `refresh()`. This case includes dead symlinks. * `_RefreshSymlinkUnresolved`: A partial cache created by refresh when we fail to resolve the file pointed to by a symlink (likely due to permissions). Symlink attributes are cached, but attributes about the linked entity are not. As mentioned, this implementation purposefully deviates from the standard. According to some readings of the specification, and the Windows filesystem implementation, the constructors and modifiers which don't pass an `error_code` must throw when the `directory_entry` points to a entity which doesn't exist. or when attribute resolution fails for another reason. @BillyONeal has proposed a more reasonable set of requirements, where modifiers other than refresh ignore errors. This is the behavior libc++ currently implements, with the expectation some form of the new language will be accepted into the standard. Some additional semantics which differ from the Windows implementation: 1. `refresh` will not throw when the entry doesn't exist. In this case we can still meet the functions specification, so we don't treat it as an error. 2. We don't clear the path name when a constructor fails via refresh (this will hopefully be changed in the standard as well). It should be noted that libstdc++'s current implementation has the same behavior as libc++, except for point (2). If the changes to the specification don't get accepted, we'll be able to make the changes later. [1] http://www.open-std.org/jtc1/sc22/wg21/docs/papers/2016/p0317r1.html Reviewers: mclow.lists, gromer, ldionne, aaron.ballman Subscribers: BillyONeal, christof, cfe-commits Differential Revision: https://reviews.llvm.org/D49530 llvm-svn: 337516
2018-07-20 09:22:32 +08:00
inline std::error_code GetTestEC(unsigned Idx = 0) {
using std::errc;
auto GetErrc = [&]() {
switch (Idx) {
case 0:
return errc::address_family_not_supported;
case 1:
return errc::address_not_available;
case 2:
return errc::address_in_use;
case 3:
return errc::argument_list_too_long;
default:
assert(false && "Idx out of range");
std::abort();
}
};
return std::make_error_code(GetErrc());
}
inline bool ErrorIsImp(const std::error_code& ec,
std::vector<std::errc> const& errors) {
for (auto errc : errors) {
if (ec == std::make_error_code(errc))
return true;
}
return false;
}
template <class... ErrcT>
inline bool ErrorIs(const std::error_code& ec, std::errc First, ErrcT... Rest) {
std::vector<std::errc> errors = {First, Rest...};
return ErrorIsImp(ec, errors);
}
// Provide our own Sleep routine since std::this_thread::sleep_for is not
// available in single-threaded mode.
void SleepFor(std::chrono::seconds dur) {
using namespace std::chrono;
#if defined(_LIBCPP_HAS_NO_MONOTONIC_CLOCK)
using Clock = system_clock;
#else
using Clock = steady_clock;
#endif
const auto wake_time = Clock::now() + dur;
while (Clock::now() < wake_time)
;
}
inline bool PathEq(fs::path const& LHS, fs::path const& RHS) {
return LHS.native() == RHS.native();
}
[libc++] Implement Directory Entry Caching -- Sort of. Summary: This patch implements directory_entry caching *almost* as specified in P0317r1. However, I explicitly chose to deviate from the standard as I'll explain below. The approach I decided to take is a fully caching one. When `refresh()` is called, the cache is populated by calls to `stat` and `lstat` as needed. During directory iteration the cache is only populated with the `file_type` as reported by `readdir`. The cache can be in the following states: * `_Empty`: There is nothing in the cache (likely due to an error) * `_IterSymlink`: Created by directory iteration when we walk onto a symlink only the symlink file type is known. * `_IterNonSymlink`: Created by directory iteration when we walk onto a non-symlink. Both the regular file type and symlink file type are known. * `_RefreshSymlink` and `_RefreshNonSymlink`: A full cache created by `refresh()`. This case includes dead symlinks. * `_RefreshSymlinkUnresolved`: A partial cache created by refresh when we fail to resolve the file pointed to by a symlink (likely due to permissions). Symlink attributes are cached, but attributes about the linked entity are not. As mentioned, this implementation purposefully deviates from the standard. According to some readings of the specification, and the Windows filesystem implementation, the constructors and modifiers which don't pass an `error_code` must throw when the `directory_entry` points to a entity which doesn't exist. or when attribute resolution fails for another reason. @BillyONeal has proposed a more reasonable set of requirements, where modifiers other than refresh ignore errors. This is the behavior libc++ currently implements, with the expectation some form of the new language will be accepted into the standard. Some additional semantics which differ from the Windows implementation: 1. `refresh` will not throw when the entry doesn't exist. In this case we can still meet the functions specification, so we don't treat it as an error. 2. We don't clear the path name when a constructor fails via refresh (this will hopefully be changed in the standard as well). It should be noted that libstdc++'s current implementation has the same behavior as libc++, except for point (2). If the changes to the specification don't get accepted, we'll be able to make the changes later. [1] http://www.open-std.org/jtc1/sc22/wg21/docs/papers/2016/p0317r1.html Reviewers: mclow.lists, gromer, ldionne, aaron.ballman Subscribers: BillyONeal, christof, cfe-commits Differential Revision: https://reviews.llvm.org/D49530 llvm-svn: 337516
2018-07-20 09:22:32 +08:00
struct ExceptionChecker {
std::errc expected_err;
[libc++] Implement Directory Entry Caching -- Sort of. Summary: This patch implements directory_entry caching *almost* as specified in P0317r1. However, I explicitly chose to deviate from the standard as I'll explain below. The approach I decided to take is a fully caching one. When `refresh()` is called, the cache is populated by calls to `stat` and `lstat` as needed. During directory iteration the cache is only populated with the `file_type` as reported by `readdir`. The cache can be in the following states: * `_Empty`: There is nothing in the cache (likely due to an error) * `_IterSymlink`: Created by directory iteration when we walk onto a symlink only the symlink file type is known. * `_IterNonSymlink`: Created by directory iteration when we walk onto a non-symlink. Both the regular file type and symlink file type are known. * `_RefreshSymlink` and `_RefreshNonSymlink`: A full cache created by `refresh()`. This case includes dead symlinks. * `_RefreshSymlinkUnresolved`: A partial cache created by refresh when we fail to resolve the file pointed to by a symlink (likely due to permissions). Symlink attributes are cached, but attributes about the linked entity are not. As mentioned, this implementation purposefully deviates from the standard. According to some readings of the specification, and the Windows filesystem implementation, the constructors and modifiers which don't pass an `error_code` must throw when the `directory_entry` points to a entity which doesn't exist. or when attribute resolution fails for another reason. @BillyONeal has proposed a more reasonable set of requirements, where modifiers other than refresh ignore errors. This is the behavior libc++ currently implements, with the expectation some form of the new language will be accepted into the standard. Some additional semantics which differ from the Windows implementation: 1. `refresh` will not throw when the entry doesn't exist. In this case we can still meet the functions specification, so we don't treat it as an error. 2. We don't clear the path name when a constructor fails via refresh (this will hopefully be changed in the standard as well). It should be noted that libstdc++'s current implementation has the same behavior as libc++, except for point (2). If the changes to the specification don't get accepted, we'll be able to make the changes later. [1] http://www.open-std.org/jtc1/sc22/wg21/docs/papers/2016/p0317r1.html Reviewers: mclow.lists, gromer, ldionne, aaron.ballman Subscribers: BillyONeal, christof, cfe-commits Differential Revision: https://reviews.llvm.org/D49530 llvm-svn: 337516
2018-07-20 09:22:32 +08:00
fs::path expected_path1;
fs::path expected_path2;
unsigned num_paths;
const char* func_name;
std::string opt_message;
explicit ExceptionChecker(std::errc first_err, const char* func_name,
std::string opt_msg = {})
: expected_err{first_err}, num_paths(0), func_name(func_name),
opt_message(opt_msg) {}
explicit ExceptionChecker(fs::path p, std::errc first_err,
const char* func_name, std::string opt_msg = {})
: expected_err(first_err), expected_path1(p), num_paths(1),
func_name(func_name), opt_message(opt_msg) {}
[libc++] Implement Directory Entry Caching -- Sort of. Summary: This patch implements directory_entry caching *almost* as specified in P0317r1. However, I explicitly chose to deviate from the standard as I'll explain below. The approach I decided to take is a fully caching one. When `refresh()` is called, the cache is populated by calls to `stat` and `lstat` as needed. During directory iteration the cache is only populated with the `file_type` as reported by `readdir`. The cache can be in the following states: * `_Empty`: There is nothing in the cache (likely due to an error) * `_IterSymlink`: Created by directory iteration when we walk onto a symlink only the symlink file type is known. * `_IterNonSymlink`: Created by directory iteration when we walk onto a non-symlink. Both the regular file type and symlink file type are known. * `_RefreshSymlink` and `_RefreshNonSymlink`: A full cache created by `refresh()`. This case includes dead symlinks. * `_RefreshSymlinkUnresolved`: A partial cache created by refresh when we fail to resolve the file pointed to by a symlink (likely due to permissions). Symlink attributes are cached, but attributes about the linked entity are not. As mentioned, this implementation purposefully deviates from the standard. According to some readings of the specification, and the Windows filesystem implementation, the constructors and modifiers which don't pass an `error_code` must throw when the `directory_entry` points to a entity which doesn't exist. or when attribute resolution fails for another reason. @BillyONeal has proposed a more reasonable set of requirements, where modifiers other than refresh ignore errors. This is the behavior libc++ currently implements, with the expectation some form of the new language will be accepted into the standard. Some additional semantics which differ from the Windows implementation: 1. `refresh` will not throw when the entry doesn't exist. In this case we can still meet the functions specification, so we don't treat it as an error. 2. We don't clear the path name when a constructor fails via refresh (this will hopefully be changed in the standard as well). It should be noted that libstdc++'s current implementation has the same behavior as libc++, except for point (2). If the changes to the specification don't get accepted, we'll be able to make the changes later. [1] http://www.open-std.org/jtc1/sc22/wg21/docs/papers/2016/p0317r1.html Reviewers: mclow.lists, gromer, ldionne, aaron.ballman Subscribers: BillyONeal, christof, cfe-commits Differential Revision: https://reviews.llvm.org/D49530 llvm-svn: 337516
2018-07-20 09:22:32 +08:00
explicit ExceptionChecker(fs::path p1, fs::path p2, std::errc first_err,
const char* func_name, std::string opt_msg = {})
: expected_err(first_err), expected_path1(p1), expected_path2(p2),
num_paths(2), func_name(func_name), opt_message(opt_msg) {}
[libc++] Implement Directory Entry Caching -- Sort of. Summary: This patch implements directory_entry caching *almost* as specified in P0317r1. However, I explicitly chose to deviate from the standard as I'll explain below. The approach I decided to take is a fully caching one. When `refresh()` is called, the cache is populated by calls to `stat` and `lstat` as needed. During directory iteration the cache is only populated with the `file_type` as reported by `readdir`. The cache can be in the following states: * `_Empty`: There is nothing in the cache (likely due to an error) * `_IterSymlink`: Created by directory iteration when we walk onto a symlink only the symlink file type is known. * `_IterNonSymlink`: Created by directory iteration when we walk onto a non-symlink. Both the regular file type and symlink file type are known. * `_RefreshSymlink` and `_RefreshNonSymlink`: A full cache created by `refresh()`. This case includes dead symlinks. * `_RefreshSymlinkUnresolved`: A partial cache created by refresh when we fail to resolve the file pointed to by a symlink (likely due to permissions). Symlink attributes are cached, but attributes about the linked entity are not. As mentioned, this implementation purposefully deviates from the standard. According to some readings of the specification, and the Windows filesystem implementation, the constructors and modifiers which don't pass an `error_code` must throw when the `directory_entry` points to a entity which doesn't exist. or when attribute resolution fails for another reason. @BillyONeal has proposed a more reasonable set of requirements, where modifiers other than refresh ignore errors. This is the behavior libc++ currently implements, with the expectation some form of the new language will be accepted into the standard. Some additional semantics which differ from the Windows implementation: 1. `refresh` will not throw when the entry doesn't exist. In this case we can still meet the functions specification, so we don't treat it as an error. 2. We don't clear the path name when a constructor fails via refresh (this will hopefully be changed in the standard as well). It should be noted that libstdc++'s current implementation has the same behavior as libc++, except for point (2). If the changes to the specification don't get accepted, we'll be able to make the changes later. [1] http://www.open-std.org/jtc1/sc22/wg21/docs/papers/2016/p0317r1.html Reviewers: mclow.lists, gromer, ldionne, aaron.ballman Subscribers: BillyONeal, christof, cfe-commits Differential Revision: https://reviews.llvm.org/D49530 llvm-svn: 337516
2018-07-20 09:22:32 +08:00
void operator()(fs::filesystem_error const& Err) {
TEST_CHECK(ErrorIsImp(Err.code(), {expected_err}));
[libc++] Implement Directory Entry Caching -- Sort of. Summary: This patch implements directory_entry caching *almost* as specified in P0317r1. However, I explicitly chose to deviate from the standard as I'll explain below. The approach I decided to take is a fully caching one. When `refresh()` is called, the cache is populated by calls to `stat` and `lstat` as needed. During directory iteration the cache is only populated with the `file_type` as reported by `readdir`. The cache can be in the following states: * `_Empty`: There is nothing in the cache (likely due to an error) * `_IterSymlink`: Created by directory iteration when we walk onto a symlink only the symlink file type is known. * `_IterNonSymlink`: Created by directory iteration when we walk onto a non-symlink. Both the regular file type and symlink file type are known. * `_RefreshSymlink` and `_RefreshNonSymlink`: A full cache created by `refresh()`. This case includes dead symlinks. * `_RefreshSymlinkUnresolved`: A partial cache created by refresh when we fail to resolve the file pointed to by a symlink (likely due to permissions). Symlink attributes are cached, but attributes about the linked entity are not. As mentioned, this implementation purposefully deviates from the standard. According to some readings of the specification, and the Windows filesystem implementation, the constructors and modifiers which don't pass an `error_code` must throw when the `directory_entry` points to a entity which doesn't exist. or when attribute resolution fails for another reason. @BillyONeal has proposed a more reasonable set of requirements, where modifiers other than refresh ignore errors. This is the behavior libc++ currently implements, with the expectation some form of the new language will be accepted into the standard. Some additional semantics which differ from the Windows implementation: 1. `refresh` will not throw when the entry doesn't exist. In this case we can still meet the functions specification, so we don't treat it as an error. 2. We don't clear the path name when a constructor fails via refresh (this will hopefully be changed in the standard as well). It should be noted that libstdc++'s current implementation has the same behavior as libc++, except for point (2). If the changes to the specification don't get accepted, we'll be able to make the changes later. [1] http://www.open-std.org/jtc1/sc22/wg21/docs/papers/2016/p0317r1.html Reviewers: mclow.lists, gromer, ldionne, aaron.ballman Subscribers: BillyONeal, christof, cfe-commits Differential Revision: https://reviews.llvm.org/D49530 llvm-svn: 337516
2018-07-20 09:22:32 +08:00
TEST_CHECK(Err.path1() == expected_path1);
TEST_CHECK(Err.path2() == expected_path2);
LIBCPP_ONLY(check_libcxx_string(Err));
}
void check_libcxx_string(fs::filesystem_error const& Err) {
std::string message = std::make_error_code(expected_err).message();
std::string additional_msg = "";
if (!opt_message.empty()) {
additional_msg = opt_message + ": ";
}
auto transform_path = [](const fs::path& p) {
if (p.native().empty())
return "\"\"";
return p.c_str();
};
std::string format = [&]() -> std::string {
switch (num_paths) {
case 0:
return format_string("filesystem error: in %s: %s%s", func_name,
additional_msg, message);
case 1:
return format_string("filesystem error: in %s: %s%s [%s]", func_name,
additional_msg, message,
transform_path(expected_path1));
case 2:
return format_string("filesystem error: in %s: %s%s [%s] [%s]",
func_name, additional_msg, message,
transform_path(expected_path1),
transform_path(expected_path2));
default:
TEST_CHECK(false && "unexpected case");
return "";
}
}();
TEST_CHECK(format == Err.what());
if (format != Err.what()) {
fprintf(stderr,
"filesystem_error::what() does not match expected output:\n");
fprintf(stderr, " expected: \"%s\"\n", format.c_str());
fprintf(stderr, " actual: \"%s\"\n\n", Err.what());
}
[libc++] Implement Directory Entry Caching -- Sort of. Summary: This patch implements directory_entry caching *almost* as specified in P0317r1. However, I explicitly chose to deviate from the standard as I'll explain below. The approach I decided to take is a fully caching one. When `refresh()` is called, the cache is populated by calls to `stat` and `lstat` as needed. During directory iteration the cache is only populated with the `file_type` as reported by `readdir`. The cache can be in the following states: * `_Empty`: There is nothing in the cache (likely due to an error) * `_IterSymlink`: Created by directory iteration when we walk onto a symlink only the symlink file type is known. * `_IterNonSymlink`: Created by directory iteration when we walk onto a non-symlink. Both the regular file type and symlink file type are known. * `_RefreshSymlink` and `_RefreshNonSymlink`: A full cache created by `refresh()`. This case includes dead symlinks. * `_RefreshSymlinkUnresolved`: A partial cache created by refresh when we fail to resolve the file pointed to by a symlink (likely due to permissions). Symlink attributes are cached, but attributes about the linked entity are not. As mentioned, this implementation purposefully deviates from the standard. According to some readings of the specification, and the Windows filesystem implementation, the constructors and modifiers which don't pass an `error_code` must throw when the `directory_entry` points to a entity which doesn't exist. or when attribute resolution fails for another reason. @BillyONeal has proposed a more reasonable set of requirements, where modifiers other than refresh ignore errors. This is the behavior libc++ currently implements, with the expectation some form of the new language will be accepted into the standard. Some additional semantics which differ from the Windows implementation: 1. `refresh` will not throw when the entry doesn't exist. In this case we can still meet the functions specification, so we don't treat it as an error. 2. We don't clear the path name when a constructor fails via refresh (this will hopefully be changed in the standard as well). It should be noted that libstdc++'s current implementation has the same behavior as libc++, except for point (2). If the changes to the specification don't get accepted, we'll be able to make the changes later. [1] http://www.open-std.org/jtc1/sc22/wg21/docs/papers/2016/p0317r1.html Reviewers: mclow.lists, gromer, ldionne, aaron.ballman Subscribers: BillyONeal, christof, cfe-commits Differential Revision: https://reviews.llvm.org/D49530 llvm-svn: 337516
2018-07-20 09:22:32 +08:00
}
ExceptionChecker(ExceptionChecker const&) = delete;
ExceptionChecker& operator=(ExceptionChecker const&) = delete;
[libc++] Implement Directory Entry Caching -- Sort of. Summary: This patch implements directory_entry caching *almost* as specified in P0317r1. However, I explicitly chose to deviate from the standard as I'll explain below. The approach I decided to take is a fully caching one. When `refresh()` is called, the cache is populated by calls to `stat` and `lstat` as needed. During directory iteration the cache is only populated with the `file_type` as reported by `readdir`. The cache can be in the following states: * `_Empty`: There is nothing in the cache (likely due to an error) * `_IterSymlink`: Created by directory iteration when we walk onto a symlink only the symlink file type is known. * `_IterNonSymlink`: Created by directory iteration when we walk onto a non-symlink. Both the regular file type and symlink file type are known. * `_RefreshSymlink` and `_RefreshNonSymlink`: A full cache created by `refresh()`. This case includes dead symlinks. * `_RefreshSymlinkUnresolved`: A partial cache created by refresh when we fail to resolve the file pointed to by a symlink (likely due to permissions). Symlink attributes are cached, but attributes about the linked entity are not. As mentioned, this implementation purposefully deviates from the standard. According to some readings of the specification, and the Windows filesystem implementation, the constructors and modifiers which don't pass an `error_code` must throw when the `directory_entry` points to a entity which doesn't exist. or when attribute resolution fails for another reason. @BillyONeal has proposed a more reasonable set of requirements, where modifiers other than refresh ignore errors. This is the behavior libc++ currently implements, with the expectation some form of the new language will be accepted into the standard. Some additional semantics which differ from the Windows implementation: 1. `refresh` will not throw when the entry doesn't exist. In this case we can still meet the functions specification, so we don't treat it as an error. 2. We don't clear the path name when a constructor fails via refresh (this will hopefully be changed in the standard as well). It should be noted that libstdc++'s current implementation has the same behavior as libc++, except for point (2). If the changes to the specification don't get accepted, we'll be able to make the changes later. [1] http://www.open-std.org/jtc1/sc22/wg21/docs/papers/2016/p0317r1.html Reviewers: mclow.lists, gromer, ldionne, aaron.ballman Subscribers: BillyONeal, christof, cfe-commits Differential Revision: https://reviews.llvm.org/D49530 llvm-svn: 337516
2018-07-20 09:22:32 +08:00
};
#endif /* FILESYSTEM_TEST_HELPER_HPP */