bug-gnu-emacs
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

bug#29455: backup-by-copying ACL Operation not permitted with Windows 7


From: Stefan Kangas
Subject: bug#29455: backup-by-copying ACL Operation not permitted with Windows 7 and Samba
Date: Sun, 29 Sep 2019 02:11:45 +0200

Eli Zaretskii <eliz@gnu.org> writes:

>> From: Shuguang Sun <shuguang@gmail.com>
>> Date: Tue, 28 Nov 2017 18:43:53 +0800
>> Cc: 29455@debbugs.gnu.org
>>
>>  Thanks, but that's not what I asked.  I asked you to create a local
>>  file (which I call "SOME-FILE" below, replace that with the actual
>>  name of the file), and then type this inside Emacs:
>>
>>    M-: (set-file-acl "SOME-FILE" 
>> "O:S-1-22-1-79077G:S-1-22-2-108D:P(A;;0x1e01ff;;;S-1-22-1-79077)
>>  (A;;FR;;;S-1-22-2-108)(A;;FR;;;WD)") RET
>>
>> It raises the same error message:
>> (file-error "Setting ACL" "Operation not permitted" 
>> "c:/Users/username/Documents/base/subtitle.txt")
>>
>> or
>> eval: Setting ACL: Operation not permitted, 
>> c:/Users/username/HOME/.emacs.d/autosave/.bashrc.~1~
>
> OK, I think I understand what causes this: the problem is with setting
> the owner of the file to a user whose SID the local system doesn't
> recognize, because it's not a local user.  If you omit the
> O:S-1-22-1-79077G:S-1-22-2-108 part from the argument, the
> set-file-acl call will most probably succeed.
>
> I think that Emacs works correctly in this case: it tries to preserve
> the ACLs of the original file in the backup file, and when that fails,
> falls back to preserving only the read-write mode bits.  Your original
> problem was caused by debug-on-error being non-nil, which shouldn't be
> done except when debugging something.
>
> I don't see how we could avoid showing the error message, because
> Emacs cannot possibly know which user SID values are known to the
> local system.  When you configure your backups of files from networked
> drives to be stored on a local disk, you should realize that you run
> the risk of having these problems with preserving extended
> permissions.
>
> So I think we see Emacs working as designed, and this bug should be
> closed.

According to the above, this bug should be closed.  I'm doing so now.

Best regards,
Stefan Kangas





reply via email to

[Prev in Thread] Current Thread [Next in Thread]