win dirs don't handle lack of inherited rule(?): getfacl + tar dir Warning: Cannot acl_to_text: Invalid argument

L A Walsh
Sun Mar 10 16:38:00 GMT 2019

On 3/10/2019 6:27 AM, Andrey Repin wrote:
> Greetings, L A Walsh!
>> On 3/8/2019 4:15 AM, Corinna Vinschen wrote:
>>> On Mar  7 19:35, L A Walsh wrote:
>>>> I ran tar on another directory and got a huge number**
>>>> of these:
>>>> tar: rules: Warning: Cannot acl_to_text: Invalid argument
>>>> tar: adblockplus: Warning: Cannot acl_to_text: Invalid argument
>>>> tar: autopager: Warning: Cannot acl_to_text: Invalid argument
>>>> tar: bookmarkbackups: Warning: Cannot acl_to_text: Invalid argument
>>>> ---
>>> Can you please provide the cacls or icacls command creating
>>> a directory that allows to reproduce the issue?
>> I doubt that area of my disk has ever been manipulated by 
>> cacls or icacls. That's in my roaming profile.
> It's not about manipulation, it's about current state.
> icacls is a Windows equivalent of POSIX's setfacl/getfacl.
	I know.  I meant manipulate in the sense of 
handling something with dexterity and finesse. -- which
is a different connotation or sense of the word than when
talking about one person manipulating another.

	How would _you_ create these symptoms if you don't know
how they got that way -- just that they exist that way.

	Furthermore.  I'm pretty sure that a person would not
be able to create that symptom with icacls (or the deprecated cacls).

I would not doubt that icacls would refuse to create
mis-ordered ACL's, for example.  I.e. its likely a non-windows 
program or odd interaction between one and windows.

That's why I pointed out that besides cygwin creating acl's that
explorer will complain about, the roaming profile has also received
profiles from a samba server on linux (with a different permission
structure)..  Its unclear what permissions are copied then and how
they are mapped.


Problem reports:
Unsubscribe info:

More information about the Cygwin mailing list