GNU bug report logs - #10468
BUG: Severe or critical - deletes existing files and leaves nothing. (cp)

Previous Next

Full log

Message #24 received at 10468 <at> debbugs.gnu.org (full text, mbox):

From: Eric Blake <eblake <at> redhat.com>
To: Linda Walsh <coreutils <at> tlinx.org>
Cc: 10468 <at> debbugs.gnu.org
Subject: Re: bug#10471: Severe or critical - deletes existing files and leaves
	nothing. (cp)
Date: Mon, 09 Jan 2012 17:57:58 -0700

[Message part 1 (text/plain, inline)]

On 01/09/2012 05:46 PM, Linda Walsh wrote:
> reopen 10468
> thanks
> =============
> 
> You didn't answer my question.. I explained why it was a bug.

You explained that the cygwin binary for cp didn't meet your
expectations, but I in turn explained that the cygwin cp binary has code
that is not present upstream, and that it is this distro-specific code
which is the primary suspect.  Until we can confirm that upstream
coreutils has the same issue without cygwin in the mix, then the bug
should be discussed on the cygwin list.

> 
> You didn't explain why it is copying links.

That's correct, because I haven't reproduced it on my cygwin machine
yet.  But I also haven't discarded that fact, merely isolated it to the
right place to discuss it.  My replies on that topic will be local to
the cygwin list, unless I can prove that the problem really does affect
upstream cp.

> If you think it is the cause of cygwin...
> I asked you to explain why.

I think I did a fair job of that.  The cygwin pre-built binary contains
additional patches not present in upstream coreutils, as well as
interacting with a rather complex cygwin1.dll that must deal with both
settings of the Windows case-sensitivity registry key, either of which
is more likely to have an issue in corner cases of case sensitivity than
upstream coreutils.

> 
> The fact that there is no behavior in cygwin that would activate forced
> link coping doesn't seem to bother you.

Wrong again - it does bother me, but without further evidence, it only
bothers me concerning the cygwin build, and not coreutils in general.
Until we have evidence to the contrary that upstream coreutils is also
buggy, I'd rather focus on fixing the bug where it was observed - on the
cygwin list.

-- 
Eric Blake   eblake <at> redhat.com    +1-919-301-3266
Libvirt virtualization library http://libvirt.org

[signature.asc (application/pgp-signature, attachment)]

Previous Next

GNU bug tracking system
Copyright (C) 1999 Darren O. Benham, 1997,2003 nCipher Corporation Ltd, 1994-97 Ian Jackson.