Re: [Bug-tar] HFS+ support in tar

[Paul Eggert]

Blair Zajac <address@hidden> writes:

> 1) Is there any interest in getting this into tar?

I think so, in the long run.  However, it should be done using the
POSIX 1003.1-2001 mechanism for extended headers.  The next version of
GNU tar is planned to use this mechanism, so it's probably better to
wait until after it is out before proceeding.

>    Normally we test if the system provides a specific capability that
>    the code can make use of.  The hfstar code makes use of
>    getattrlist() and setattrlist() and other OS specific calls.
>    Should configure just test if it's running on Darwin and enable the
>    entire suite of HFS+ code, say with HAVE_HFS_PLUS?

That depends on whether the set of calls is really a unit.  If some
non-Darwin systems (or different versions of Darwin) have some calls
and not others, then the configuration should use whatever calls are
available.

>    What's the best way of testing this, say checking for getattrlist()

Yes.

>    or just doing a match on the output of config.guess?

No.  (That's not the Autoconf way.  :-)