Re: inet_addr? inet_aton?

[Russ Allbery]

Austin Schutz <address@hidden> writes:

> So why not make that part of an autoconf library of standardized
> functions?

It's been discussed from time to time, and if someone wanted to spearhead
a project, I think it could be made to happen.  libiberty already tries to
do this in part.

One of the issues that that person would need to think about, though, is
licensing.  Within the GNU community there's an understandable bias
towards using the GPL, but the nature of such a replacement function
library is that putting it under the GPL drastically limits the number of
projects that are able to use it and therefore limits the number of
contributors.  For example, I'd be happy to help contribute to such a
project if I could use it, and I have a half-dozen functions suitable for
such a library (at least), but many of the projects that I work on are
covered under other licenses and I don't want the overall package to be
put under the GPL.  (And in some cases, for reasons outside of my control,
the licenses are incompatible; for example, INN still has the advertising
clause and I've not been successful yet in dropping it.)

So such a project would be welcome to swipe my functions, as they're all
in the public domain, but I'm not interested in actively contributing to
it or feeding any updates to it unless the license is suitable (which
means LGPL at a bare minimum, and for this sort of project I'd personally
prefer to release it *all* into the public domain).

Of course, on the flip side, not allowing GPL'd code means you can't use
some things like getopt_long or the GNU getloadavg code, which would be
very useful for this sort of project.

-- 
Russ Allbery (address@hidden)             <http://www.eyrie.org/~eagle/>