Simple file system is not read only

[WeidiDeng]

In the documentation there is a section about Static in-memory file systems.

The file system this creates is not technically read-only, as files and directories can be removed without any error.

There is even todo in the code:

go-fuse/fs/bridge.go

Lines 395 to 408 in f1d40d0

	func (b *rawBridge) Rmdir(cancel <-chan struct{}, header *fuse.InHeader, name string) fuse.Status {
	parent, _ := b.inode(header.NodeId, 0)
	var errno syscall.Errno
	if mops, ok := parent.ops.(NodeRmdirer); ok {
	errno = mops.Rmdir(&fuse.Context{Caller: header.Caller, Cancel: cancel}, name)
	}
	// TODO - this should not succeed silently.
	if errno == 0 {
	parent.RmChild(name)
	}
	return errnoToStatus(errno)
	}

go-fuse/fs/bridge.go

Lines 410 to 423 in f1d40d0

	func (b *rawBridge) Unlink(cancel <-chan struct{}, header *fuse.InHeader, name string) fuse.Status {
	parent, _ := b.inode(header.NodeId, 0)
	var errno syscall.Errno
	if mops, ok := parent.ops.(NodeUnlinker); ok {
	errno = mops.Unlink(&fuse.Context{Caller: header.Caller, Cancel: cancel}, name)
	}
	// TODO - this should not succeed silently.
	if errno == 0 {
	parent.RmChild(name)
	}
	return errnoToStatus(errno)
	}

The docs for these methods also indicate success by default

NodeRmdirer.Rmdir

NodeRmdirer.Rmdir

[hanwen]

yes, you are correct. Unfortunately, I fear we would break existing callers if we changed this. I think it's best if the documentation is updated.