Proposed resolution:

In [class.directory_entry] Class directory_entry add the following observer declarations:

      uintmax_t file_size();
      uintmax_t file_size(error_code& ec) noexcept;
    

In directory_entry observers [directory_entry.obs] add the following:

      uintmax_t file_size();
      uintmax_t file_size(error_code& ec) noexcept;
    

Returns: if *this contains a cached file size, return it. Otherwise return file_size(path()) or file_size(path(), ec) respectively.

Throws: As specified in Error reporting (7).

[ 2016-08, Beman comments ]

This will be resolved by P0317R1, Directory Entry Caching for Filesystem.

Fri AM: Moved to Tentatively Resolved

[ Apr 2016 Issue updated to address the C++ Working Paper. Previously addressed File System TS ]

Previous resolution [SUPERSEDED]

In [fs.class.directory_entry] Class directory_entry add the following observer declarations:

      uintmax_t file_size();
      uintmax_t file_size(error_code& ec) noexcept;
    

In directory_entry observers [fs.dir.entry.obs] add the following:

      uintmax_t file_size();
      uintmax_t file_size(error_code& ec) noexcept;
    

Returns: if *this contains a cached file size, return it. Otherwise return file_size(path()) or file_size(path(), ec) respectively.

Throws: As specified in Error reporting (7).

[ Mar 2016 Jacksonville Beman to provide paper about this ]

[ 23 Nov 2015 Editorally correct name of data structure mentioned in discussion. ]

[ 17 Jun 2014 Rapperswil LWG will investigate issue at a subsequent meeting. ]

On Windows, the FindFileData WIN32_FIND_DATA structure, which is the underlying data type for directory_entry, contains the file size as one of the fields. Thus efficient enumeration of files and getting their sizes is possible without doing a separate query for the file size.