[k2] add file, is_file builtins

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

@@ -40,17 +40,18 @@ 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 ($name ::: string) ::: string[] | false;
-/** @kphp-extern-func-info stub generation-required */
 function file_put_contents ($name ::: string, $content ::: mixed, $flags ::: int = 0) ::: int | false;
 /** @kphp-extern-func-info stub generation-required */
 function filectime ($name ::: string) ::: int | false;
@@ -59,8 +60,6 @@ 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

@@ -418,6 +418,13 @@ inline int32_t stat(std::string_view path, struct stat* stat) noexcept {
   return k2::errno_ok;
 }
 
+inline 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 error_code;
+  }
+  return k2::errno_ok;
+}
+
 using CommandArg = CommandArg;
 
 enum class CommandStdoutPolicy : uint8_t { NoCapture, Capture };

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

@@ -735,6 +735,29 @@ 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/stdlib/file/file-system-functions.h

@@ -197,3 +197,55 @@ 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;
+  uint64_t file_fd;
+
+  if (k2::fopen(std::addressof(file_fd), name.c_str(), "r") != k2::errno_ok) {
+    return false;
+  }
+  if (k2::stat(name.c_str(), std::addressof(stat_buf)) != k2::errno_ok) {
+    k2::free_descriptor(file_fd);
+    return false;
+  }
+  if (!S_ISREG(stat_buf.st_mode)) {
+    k2::free_descriptor(file_fd);
+    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()) {
+    k2::free_descriptor(file_fd);
+    kphp::log::warning("file \"{}\" is too large", name.c_str());
+    return false;
+  }
+
+  string res(static_cast<string::size_type>(size), false);
+  char* s{res.buffer()};
+  if (k2::read(file_fd, size, s) < size) {
+    k2::free_descriptor(file_fd);
+    return false;
+  }
+  k2::free_descriptor(file_fd);
+
+  array<string> result;
+  int prev{-1};
+  for (size_t i = 0; i < size; i++) {
+    if (s[i] == '\n' || i + 1 == size) {
+      result.push_back(string(s + prev + 1, i - prev));
+      prev = i;
+    }
+  }
+
+  return result;
+}
+
+inline bool f$is_file(const string& name) noexcept {
+  struct stat stat_buf;
+  if (k2::lstat(name.c_str(), std::addressof(stat_buf)) != k2::errno_ok) {
+    return false;
+  }
+  return S_ISREG(stat_buf.st_mode);
+}