lmi
[Top][All Lists]
Advanced

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

Re: [lmi] 'cp' fails with "permission" error


From: Vadim Zeitlin
Subject: Re: [lmi] 'cp' fails with "permission" error
Date: Thu, 9 Nov 2017 21:29:20 +0100

On Thu, 9 Nov 2017 16:45:52 +0000 Greg Chicares <address@hidden> wrote:

GC> Vadim--On a new machine in the office (configured by corporate IT)
GC> we observe:
GC> 
GC> cp --preserve /cache_for_lmi/downloads/md5sum.exe /opt/lmi/third_party/bin
GC> cp: failed to preserve ownership for '/opt/lmi/third_party/bin/md5sum.exe': 
Permission denied
GC> make: *** [install_miscellanea.make:219: md5sum_msw] Error 1
GC> 
GC> How would we go about debugging this? I'm guessing the answer
GC> involves something like 'icacls', but I know nothing about
GC> msw permissions since the msw-xp era.
GC> 
GC> (I thought this was supposed to be an msw-10 machine, but cygwin
GC> diagnostics identify it as msw-7.)

 Hello,

 First of all, I wonder if using the "--preserve" cp option is really
crucial here? POSIX ownership/permissions maps only imperfectly on MSW
ACL-based security model, so why insist on propagating them? I'd say that
if everything works fine without "--preserve" on this machine, then nothing
would be lost by just dropping it.

 If things don't work when doing just a simple copy, e.g. if the file loses
its "executable" status, then you do need to use icacls (or cacls, which is
deprecated, but still works fine under MSW 7 AFAIK) to gather more
information. I'd start by running just "icacls c:/opt/lmi/third_party/bin"
to see if there are any unusual ACLs associated with this directory and,
maybe, do the same for its parent directories, recursively. I expect the
current user doesn't have full control ("(F)" in icacls output) over that
directory, but I can't say much more than that and I don't know what fix
would be appropriate as this depends on the local security policies that I
also don't know anything about.

 Sorry if this is not particularly helpful,
VZ


reply via email to

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