[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [bug-inetutils] Re: DNS SRV support (patch)

From: Simon Josefsson
Subject: Re: [bug-inetutils] Re: DNS SRV support (patch)
Date: Mon, 13 Dec 2004 00:23:07 +0100
User-agent: Gnus/5.110003 (No Gnus v0.3) Emacs/21.3.50 (gnu/linux)

"Alfred M. Szmidt" <address@hidden> writes:

>    The code isn't part of gnulib, yet,
> On a slightly offtopic note, I don't think that gnulib is the right
> place for these kind of things.  It seems that with time it is
> becoming a dumping ground for all sorts of stuff.
> Seems to me that gnulib has two goals nowadays, one is to provide a
> portable implementation of various GNU extentions (which I think was
> the inital goal), and the other as a place to provide for usable
> functions that aren't in glibc (x* for example) but many programs use.
> For the first goal, then having to duplicate the code all over the
> place is kinda a must.  But for the second, a shared library would
> make life so much easier for everyone.
> Maybe my information about gnulib is a bit outdate, maybe this is
> already possible.  Is it possible to install gnulib as a library onto
> the system and then use one or more headers to get the prototypes?

You could run 'gnulib-tool --import' and import every module, and then
modify the generated Makefile to install the library and header files,
but other than that, I don't think what you are asking for is

I wouldn't agree that a shared library would make things much easier.
It may make some things simpler, but not all things.  For me, the
benefits of having more shared stuff in gnulib rather than in a shared
library in many cases outweigh the disadvantages.

Note that you won't get non-glibc stuff like x* from gnulib unless you
ask for it.  So even if you don't like the x* stuff, I don't see what
harm it is in having those modules be part of gnulib.  If you don't
want it, don't use it.

As for InetUtils, there seem to be three options: 1) use RULI (or
another shared library) as proposed, 2) incorporate a SRV
implementations not from gnulib, or 3) use the alleged gnulib module.
I think 2) would be a bad idea (who would maintain it?).  Which of 1)
and 3) is the best depend on the goals of InetUtils.

IMHO, InetUtils should have as few external dependencies as possible,
to be small and easily ported, which I think would argue for 3).


reply via email to

[Prev in Thread] Current Thread [Next in Thread]