Change lfs_dir_read to return 0 on success

[geky]

If #491 is merged, it means breaking changes to the API, and a major version bump of some form (but not disk-breaking). This means it's a good time to look at open issues we couldn't fix early without breaking the API.

---

Here's a requested change to lfs_dir_read discussed in #114

Before:

• If an entry is found, lfs_dir_read returns 1
• If at end of directory, lfs_dir_read returns 0

After:

• If an entry is found, lfs_dir_read returns 0
• If at end of directory, lfs_dir_read returns LFS_ERR_NOENT

---

For some background:

lfs_dir_read is a bit of an odd function. It's supposed to match lfs_file_read, but instead of reading bytes, it reads directory entries.

Its POSIX equivalent, readdir, indicates whether or not we reached the end of the directory by returning a NULL pointer:

struct dirent *readdir(DIR *dirp);

But this API needed to change for littlefs, since we don't use errno, instead returning errors as signed integers through all API functions, including lfs_dir_read. So, in an effort to match lfs_file_read, lfs_dir_read returned the "number" of directory entries read, limited to either 0 or 1.

Perhaps unsurprisingly, this turned out to be confusing to users, and a better API was proposed @hathach to instead return either 0 or LFS_ERR_NOENT, an error which can't occur for other reasons in lfs_dir_read.

Suggested by @hathach
Should fix #114