method
prevent_directory_traversal
ruby latest stable - Class:
WEBrick::HTTPServlet::FileHandler
prevent_directory_traversal(req, res)private
No documentation available.
# File lib/webrick/httpservlet/filehandler.rb, line 242
def prevent_directory_traversal(req, res)
# Preventing directory traversal on Windows platforms;
# Backslashes (0x5c) in path_info are not interpreted as special
# character in URI notation. So the value of path_info should be
# normalize before accessing to the filesystem.
# dirty hack for filesystem encoding; in nature, File.expand_path
# should not be used for path normalization. [Bug #3345]
path = req.path_info.dup.force_encoding(Encoding.find("filesystem"))
if trailing_pathsep?(req.path_info)
# File.expand_path removes the trailing path separator.
# Adding a character is a workaround to save it.
# File.expand_path("/aaa/") #=> "/aaa"
# File.expand_path("/aaa/" + "x") #=> "/aaa/x"
expanded = File.expand_path(path + "x")
expanded.chop! # remove trailing "x"
else
expanded = File.expand_path(path)
end
expanded.force_encoding(req.path_info.encoding)
req.path_info = expanded
end