Re: Another rfe: "cp" this time

[Jim Meyering]

Bruce Korb wrote:
> On 04/27/12 19:00, Bruce Korb wrote:
>>> Link? no. Copy? yes. But it is a long way from...
>>
>> Not only not ready for prime time, but
>
> Link? Yes.
>
> I needed another file, so I finished the basic work.
> Reflecting on the special config needs tho (throttling,
> user entertainment [progress info], etc.) maybe not
> a candidate for incorporation into "cp", yielding
> yet another utility?  The shell wrapper around "dd"
> ought to have worked, but it did not.  All the dd's
> with a non-zero "seek" option quit without transferring
> any data at all.  The "rsync" solution decided that
> since the source file was local, it would read through
> the entire file.  Apparently it is not network mount
> aware.  Anyway, with a reasonably recent "libopts"
> installation, you can clone this, unroll the .tar
> file and build:
>
> git://github.com/brkorb/pcopy.git

Hi Bruce,

I took a quick look: one nit: you should either change
the three strdup uses, s/strdup/xstrdup/, or arrange to
handle memory allocation failure, so that the post-strdup
code doesn't deref NULL.

I did try to build, but got this w/F17's
autogen-libopts-5.12-2.fc17.x86_64:

    autogen -MTstamp-pcopy-opts -MFpcopy-opts.d -MP pcopy-opts.def
    Error in template /usr/share/autogen/optcode.tlib, line 523
            DEFINITIONS ERROR in /usr/share/autogen/optcode.tlib line 523 for 
pcopy-opts.c:
            invalid value for 'no-xlate'
    Failing Guile command:  = = = = =

    (error "invalid value for 'no-xlate'")

    =================================
    make: *** [stamp-pcopy-opts] Error 2

So far, it seems like you're the only one to find this useful.
But now that you've identified the problem and have a solution,
who knows... maybe a lot of people will come forward and say your
tool makes a big difference for them.

What do you think about distributing it separately for now,
and we reconsider if/when there's more demand?