Re: Manual improvement suggestion

On 3/29/26 12:47, James Youngman wrote:
> On Sat, Mar 28, 2026 at 11:46 PM raf <[email protected]> wrote:
> 
>> On Sat, Mar 28, 2026 at 03:22:05PM +0000, James Youngman <[email protected]>
>> wrote:
>>
>>> What about "directory with no children" or "directory with no
>> descendants"?
>>
>> There's nothing ambiguous about "empty directory". Everyone knows
>> that all directories contain "." and "..".
>>
> 
> FWIW, POSIX doesn't require that and not all file systems have them.
> 
> POSIX distinguishes the meaning of dot and dot-dot in pathname resolution
> <https://pubs.opengroup.org/onlinepubs/9799919799/basedefs/V1_chap04.html#tag_04_16>
> (where they act as I'm sure you expect) and in the output of readdir
> <https://pubs.opengroup.org/onlinepubs/9799919799/functions/readdir.html#tag_17_477>
> (from which they may be absent without affecting pathname resolution).

Thanks for the pointers.

> IOW to be strictly correct we would need to say:
> 
> Zero-length file or a directory containing no entries (other than "." and "..").

I agree that this would be the more precise yet most concise wording.

> (I unilaterally suggested "zero-length" file because IMO "zero-byte" file
> invites nitpicking about confusion with a non-empty file containing bytes
> whose values are all zero.)

yes, that's better, indeed.

> Based on our discussion so far it seems some or perhaps all  suggested
> clarifications (including my own) have provoked corrections and quibbles.
> Whereas the existing wording has survived a long time and generated no
> confusion or bug reports to date, so far as I can recall.  Perhaps that
> indicates we should keep the current wording.

I agree here as well.  So let's leave it as it is.

Thanks & have a nice day,
Berny