Re: ACL patch feedback

[James Youngman]

On Thu, Nov 24, 2005 at 11:32:16PM +0100, Jim Meyering wrote:
> James Youngman <address@hidden> wrote:
> > On Thu, Nov 24, 2005 at 02:31:12PM +0100, Jim Meyering wrote:

> >> [assuming there's a good reason for --attributes=REGEX, ]
> >> With the XATTR changes, copy.c's copy_extended_attributes
> >> exits if it encounters an invalid xattr regexp.  
[...]

> Are you suggesting that cp exit immediately upon the first
> failure to copy?

No, I had assumed (as I said, without readin ghte patch, sorry) that
the "invalid xattr regexp" could be detected at initialisation time,
before the first output file has been opened.  

> If either cp or mv encounters an error while processing a command-line
> argument, it must report it (usually), note the failure so as to ensure
> a nonzero exit status, and continue processing any remaining arguments.

Indeed.

> You don't really want to change such a long-standing tradition, do you?

No, I had misunderstood the point at which the error you are referring
to is detected.  I had thought your comment was about an error that
can be detected while cp is still parsing its command line arguments.

> I'm not sure what you're suggesting, since you admit that
> you like it when `cp a no-such c dest-dir' copies `c', even
> though it failed to copy a preceding nonexistent `no-such' argument --
> since it exits nonzero.

Yes, once the first output file has been created it's better to go
forward and complete as much as we can rather than fail immediately on
an error.

Sorry about any confusion caused.

James.