Re: Request to the git maintainer

[Erik Soderquist <ErikSoderquist@gmail.com>]

On Mon, Jul 10, 2017 at 7:17 PM, Eliot Moss  wrote:
> I'm not asking "every vendor".  I'm asking a limited number of
> cygwin port maintainers (and so far only one specifically).
>
> CrashPlan does not have "bizarre" limitations -- it quite
> naturally has some difficulty if a file's access does not grant
> it permission.

In my experience, that is a bizarre limitation; I have yet to work
with a Windows backup program that does not have the ability to bypass
permissions when performing a backup or restore when it has been
properly configured.  Even simple robocopy has this option.

> The permissions on these files is what is bizarre.  There is no
> good reason they should deny read access.

I work with a lot of customers where the opposite perspective is taken
for security and confidentiality reasons -- if you don't have a very
strong reason to allow access, you deny by default.

> Windows does not really have an equivalent of "root" that can
> access anything, so there is not really any privilege I can
> give the backup tool that is guaranteed to work in all cases.

Windows does have its backup operators set of permissions, allowing
the backup processes to archive and restore files the user would not
normally have access to without changing those permissions and even
with restoring the permissions as they were archived.  I would check
with the CrashPlan admins for why they are not using the backup
operators permissions.

> I can fix the access mode manually, but every time there is an
> update to the package, it will revert, so that's not a great
> solution.

You should not have to change the file permissions, regardless of what
they are; Windows based backup software should not be having a problem
with the existing permissions.

-- Erik

--
Problem reports:       http://cygwin.com/problems.html
FAQ:                   http://cygwin.com/faq/
Documentation:         http://cygwin.com/docs.html
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple