[k2] add file, is_file builtins (#1483)

Author: denisichh

builtin-functions/kphp-light/stdlib/file-functions.txt

@@ -45,25 +45,24 @@ function vsprintf ($format ::: string, $args ::: array) ::: string;
 
 function file_exists ($name ::: string) ::: bool;
 
-// === UNSUPPORTED ===
+function file ($name ::: string) ::: string[] | false;
 
+function is_file ($name ::: string) ::: bool;
+
+// === UNSUPPORTED ===
 /** @kphp-extern-func-info stub generation-required */
 function chmod ($name ::: string, $mode ::: int) ::: bool;
 /** @kphp-extern-func-info stub generation-required */
 function copy ($from ::: string, $to ::: string) ::: bool;
 /** @kphp-extern-func-info stub generation-required */
 function dirname ($name ::: string) ::: string;
 /** @kphp-extern-func-info stub generation-required */
-function file ($name ::: string) ::: string[] | false;
-/** @kphp-extern-func-info stub generation-required */
 function filectime ($name ::: string) ::: int | false;
 /** @kphp-extern-func-info stub generation-required */
 function filemtime ($name ::: string) ::: int | false;
 /** @kphp-extern-func-info stub generation-required */
 function is_dir ($name ::: string) ::: bool;
 /** @kphp-extern-func-info stub generation-required */
-function is_file ($name ::: string) ::: bool;
-/** @kphp-extern-func-info stub generation-required */
 function is_readable ($name ::: string) ::: bool;
 /** @kphp-extern-func-info stub generation-required */
 function mkdir ($name ::: string, $mode ::: int = 0777, $recursive ::: bool = false) ::: bool;

runtime-light/k2-platform/k2-api.h

@@ -415,11 +415,18 @@ inline auto canonicalize(std::string_view path) noexcept {
   return return_type{{unique_ptr_type{resolved_path, std::invoke(deleter_creator, resolved_path_len, resolved_path_align)}, resolved_path_len}};
 }
 
-inline int32_t stat(std::string_view path, struct stat* stat) noexcept {
+inline std::expected<void, int32_t> stat(std::string_view path, struct stat* stat) noexcept {
   if (auto error_code{k2_stat(path.data(), path.size(), stat)}; error_code != k2::errno_ok) [[unlikely]] {
-    return error_code;
+    return std::unexpected{error_code};
   }
-  return k2::errno_ok;
+  return {};
+}
+
+inline std::expected<void, int32_t> lstat(std::string_view path, struct stat* stat) noexcept {
+  if (auto error_code{k2_lstat(path.data(), path.size(), stat)}; error_code != k2::errno_ok) [[unlikely]] {
+    return std::unexpected{error_code};
+  }
+  return {};
 }
 
 using CommandArg = CommandArg;

runtime-light/k2-platform/k2-header.h

@@ -742,6 +742,29 @@ int32_t k2_canonicalize(const char* path, size_t pathlen, char* const* resolved_
  */
 int32_t k2_stat(const char* pathname, size_t pathname_len, struct stat* statbuf);
 
+/**
+ * Semantically equivalent to libc's `lstat`.
+ *
+ * Possible `errno`:
+ * `EACCES` => Search permission is denied for one of the directories in the path prefix of `pathname`.
+ * `EINVAL` => `pathname` or `statbuf` is `NULL`.
+ * `EFAULT` => Bad address.
+ * `ELOOP` => Too many symbolic links encountered while traversing the path.
+ * `ENAMETOOLONG` => `pathname` is too long.
+ * `ENOENT` => A component of `pathname` does not exist or is a dangling symbolic link.
+ * `ENOMEM` => Out of memory (i.e., kernel memory).
+ * `ENOTDIR` => A component of the path prefix of `pathname` is not a directory.
+ * `EOVERFLOW` => `pathname` refers to a file whose size, inode number,
+ *                or number of blocks cannot be represented in, respectively,
+ *                the types `off_t`, `ino_t`, or `blkcnt_t`.  This error can occur
+ *                when, for example, an application compiled on a 32-bit
+ *                platform without `-D_FILE_OFFSET_BITS=64` calls `lstat()` on a
+ *                file whose size exceeds `(1<<31)-1` bytes.
+ * `ERANGE` => Failed to convert `st_size`, `st_blksize`, or `st_blocks` to `int64_t`.
+ * `ENOSYS` => Internal error.
+ */
+int32_t k2_lstat(const char* pathname, size_t pathname_len, struct stat* statbuf);
+
 struct CommandArg {
   const char* arg;
   size_t arg_len;

runtime-light/state/component-state.cpp

@@ -54,8 +54,9 @@ void ComponentState::parse_runtime_config_arg(std::string_view value_view) noexc
   }
 
   struct stat stat {};
-  if (auto error_code{k2::stat({runtime_config_path.get(), runtime_config_path_size}, std::addressof(stat))}; error_code != k2::errno_ok) [[unlikely]] {
-    return kphp::log::warning("error getting runtime-config stat: error code -> {}", error_code);
+  if (auto expected_stat_result{k2::stat({runtime_config_path.get(), runtime_config_path_size}, std::addressof(stat))}; !expected_stat_result.has_value())
+      [[unlikely]] {
+    return kphp::log::warning("error getting runtime-config stat: error code -> {}", expected_stat_result.error());
   }
 
   const auto runtime_config_mem{std::unique_ptr<char, decltype(std::addressof(kphp::memory::script::free))>{

runtime-light/stdlib/file/file-system-functions.h

@@ -16,7 +16,9 @@
 #include <unistd.h>
 #include <utility>
 
+#include "runtime-common/core/allocator/script-allocator.h"
 #include "runtime-common/core/runtime-core.h"
+#include "runtime-common/core/std/containers.h"
 #include "runtime-common/stdlib/array/array-functions.h"
 #include "runtime-common/stdlib/string/string-functions.h"
 #include "runtime-light/coroutine/task.h"
@@ -61,7 +63,7 @@ inline string f$basename(const string& path, const string& suffix = {}) noexcept
 
 inline Optional<int64_t> f$filesize(const string& filename) noexcept {
   struct stat stat {};
-  if (auto errc{k2::stat({filename.c_str(), filename.size()}, std::addressof(stat))}; errc != k2::errno_ok) [[unlikely]] {
+  if (auto stat_result{k2::stat({filename.c_str(), filename.size()}, std::addressof(stat))}; !stat_result.has_value()) [[unlikely]] {
     return false;
   }
   return static_cast<int64_t>(stat.st_size);
@@ -201,6 +203,57 @@ inline Optional<string> f$file_get_contents(const string& stream) noexcept {
   return false;
 }
 
+inline Optional<array<string>> f$file(const string& name) noexcept {
+  struct stat stat_buf {};
+
+  auto expected_file{kphp::fs::file::open(name.c_str(), "r")};
+  if (!expected_file.has_value()) {
+    return false;
+  }
+  if (!k2::stat(name.c_str(), std::addressof(stat_buf)).has_value()) {
+    return false;
+  }
+  if (!S_ISREG(stat_buf.st_mode)) {
+    kphp::log::warning("regular file expected as first argument in function file, \"{}\" is given", name.c_str());
+    return false;
+  }
+
+  const size_t size{static_cast<size_t>(stat_buf.st_size)};
+  if (size > string::max_size()) {
+    kphp::log::warning("file \"{}\" is too large", name.c_str());
+    return false;
+  }
+
+  kphp::stl::vector<std::byte, kphp::memory::script_allocator> file_content;
+  file_content.resize(size);
+  {
+    auto file{std::move(*expected_file)};
+    if (auto expected_read_result{file.read(file_content)}; !expected_read_result.has_value() || *expected_read_result < size) {
+      return false;
+    }
+  }
+
+  array<string> result;
+  int32_t prev{-1};
+  for (size_t i{0}; i < size; i++) {
+    if (static_cast<char>(file_content[i]) == '\n' || i + 1 == size) {
+      result.push_back(string{reinterpret_cast<char*>(file_content.data()) + prev + 1, static_cast<string::size_type>(i - prev)});
+      prev = i;
+    }
+  }
+
+  return result;
+}
+
+inline bool f$is_file(const string& name) noexcept {
+  struct stat stat_buf {};
+  // TODO: the semantics in PHP are different: PHP expects stat
+  if (!k2::lstat(name.c_str(), std::addressof(stat_buf)).has_value()) {
+    return false;
+  }
+  return S_ISREG(stat_buf.st_mode);
+}
+
 inline Optional<int64_t> f$file_put_contents(const string& stream, const mixed& content_var, int64_t flags = 0) noexcept {
   string content{content_var.is_array() ? f$implode(string{}, content_var.to_array()) : content_var.to_string()};
 

runtime-light/stdlib/kml/kml-file-api.h

@@ -120,8 +120,8 @@ class dir_traverser final : public kphp::kml::dir_traverser_interface<dir_traver
     }
 
     struct stat stat {};
-    if (auto error_code{k2::stat(direntry_path, std::addressof(stat))}; error_code != k2::errno_ok) [[unlikely]] {
-      kphp::log::warning("[kml] failed to get stat: error code -> {}, path -> {}", error_code, direntry_path);
+    if (auto expected_stat_result{k2::stat(direntry_path, std::addressof(stat))}; !expected_stat_result.has_value()) [[unlikely]] {
+      kphp::log::warning("[kml] failed to get stat: error code -> {}, path -> {}", expected_stat_result.error(), direntry_path);
       return std::nullopt;
     }