guanlun / fstream

Advanced FS Streaming for Node

Like FS streams, but with stat on them, and supporting directories and symbolic links, as well as normal files. Also, you can use this to set the stats on a file, even if you don't change its contents, or to create a symlink, etc.

So, for example, you can "write" a directory, and it'll call mkdir. You can specify a uid and gid, and it'll call chown. You can specify a mtime and atime, and it'll call utimes. You can call it a symlink and provide a linkpath and it'll call symlink.

Note that it won't automatically resolve symbolic links. So, if you call fstream.Reader('/some/symlink') then you'll get an object that stats and then ends immediately (since it has no data). To follow symbolic links, do this: fstream.Reader({path:'/some/symlink', follow: true }).

There are various checks to make sure that the bytes emitted are the same as the intended size, if the size is set.

Examples

fstream
  .Writer({ path: "path/to/file"
          , mode: 0755
          , size: 6
          })
  .write("hello\n")
  .end()

This will create the directories if they're missing, and then write hello\n into the file, chmod it to 0755, and assert that 6 bytes have been written when it's done.

fstream
  .Writer({ path: "path/to/file"
          , mode: 0755
          , size: 6
          , flags: "a"
          })
  .write("hello\n")
  .end()

You can pass flags in, if you want to append to a file.

fstream
  .Writer({ path: "path/to/symlink"
          , linkpath: "./file"
          , SymbolicLink: true
          , mode: "0755" // octal strings supported
          })
  .end()

If isSymbolicLink is a function, it'll be called, and if it returns true, then it'll treat it as a symlink. If it's not a function, then any truish value will make a symlink, or you can set type: 'SymbolicLink', which does the same thing.

Note that the linkpath is relative to the symbolic link location, not the parent dir or cwd.

fstream
  .Reader("path/to/dir")
  .pipe(fstream.Writer("path/to/other/dir"))

This will do like cp -Rp path/to/dir path/to/other/dir. If the other dir exists and isn't a directory, then it'll emit an error. It'll also set the uid, gid, mode, etc. to be identical. In this way, it's more like rsync -a than simply a copy.