Re: Fwd: New datasource implementation

[Guy Harris <guy () alum mit edu>]

On Dec 27, 2011, at 12:58 PM, Michael Richardson wrote:

> > > > > > "Akos" == Akos Vandra <axos88 () gmail com> writes:
>
>    Akos> P.S. Why are configure and config.h.in tracked? Shouldn't
>    Akos> these be generated by autoconf and autoheader?
>
> In theory, yes.  In practice, end users never have the exact *right*
> version of those utilities installed, and only people hacking that part
> of the code need it.

The distribution tarballs should definitely include configure and config.h.in - but they could be generated by the 
process of making the release tarball; that's how Wireshark does it.  That way, end users don't need to have autoconf 
or autoheader installed at all, much less having the right version installed; as you note, the people hacking on the 
configure scripts would need to have them installed anyway.

> If you "touch .devel" in the directory, configure does some more things,
> including running "make depend".  Hmm. I thought it also assumed
> autoconf existed, and would regenerate things, but I don't see that in ./configure.

Yes, we'd want Makefile rules to generate the configure script, and make the release tarball depend on configure so 
that those rules are run when you do "make releasetar".-
This is the tcpdump-workers list.
Visit https://cod.sandelman.ca/ to unsubscribe.