Re: coreutils-9.4.170-7b206 ls/removed-directory test failure

[Paul Eggert]

On 3/26/24 11:35, Pádraig Brady wrote:
> Actually the FreeBSD system ls(1) does _not_ show the error in this case
> (it doesn't use readdir I think).

If I'm reading the source code aright, FreeBSD ls uses fts, which does 
use readdir. It's not clear to me whether it's intended that the readdir 
errno is swallowed up by fts's maze of twisty code paths.


> So I'm not leaning towards GNU ls
> just eating the ENOENT which simplifies the test and makes both
> ls implementations behave the same in this regard.

I assume you meant "now" not "not". That sounds OK to me, thanks.