[Xastir-Dev] Re: Package or not?

[Jack Twilley]

WARNING: Unsanitized content follows.
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

>>>>> "Gerry" == Gerry Creager <N5JXS <gerry.creager at tamu.edu>> writes:

Gerry> Realizing that, as a RedHat user I'm a little
Gerry> discontinuous... because I don't mind building from source, and
Gerry> keep the 'make uninstall' capability around, I'd like to
Gerry> politely disagree.

Not all RedHat users are package weenies.  This I know -- the
distribution I used to run when I ran Linux was RedHat.

Gerry> What we need to do is provide a makefile designed to unpack a
Gerry> set of tarballs, sequentially configure them, make then and run
Gerry> a chrooted make install of the interim libs.

There are pluses and minuses to this approach.  These are the ones
that come to mind at the moment:

Plus - we *always* have the right version for all the right libraries
Plus - we can correct any problems the libraries have ourselves

Minus - developers now have to maintain tens of thousands of lines of
        source code which only has a peripheral relationship to xastir
Minus - the binary (and thus any packages thereof) becomes huge since
        we can't use shared libraries anymore
Minus - changing libraries (adding, subtracting, upgrading) is a PITA
Minus - possible licensing issues

It's a tough call on what should be done in our case.  Even if we only
include the chaotic libraries (ImageMagick and shapelib and geotiff,
oh my) in the source tree and ignore well-behaved stuff like libcurl,
developers will have a whole lot more stuff to worry about.  

There are some applications which do this.  GDAL, for instance,
includes internal versions of a number of libraries, including geotiff
and xpm.  However, there are very good reasons that the vast majority
of applications out there do not include their own versions of library
software, and we should keep this in mind.

In addition, it will not entirely solve this problem.  There will
always be OS/platform-specific issues that need to be handled.  The
only way to completely solve this problem would be to develop the
software package for an embedded system -- and wouldn't that be cool.
Imagine a cheap PC running on 12V running embedded Xastir.  Goodness.

Jack.
(really liking that last idea.  hmmm.)
- -- 
Jack Twilley
jmt at twilley dot org
http colon slash slash www dot twilley dot org slash tilde jmt slash
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.2 (FreeBSD)

iD8DBQE+xV6rGPFSfAB/ezgRAnpMAKCAU9EXaYAszf+uxZaruob5fnHVHgCgnIkL
DE2ZfFk1GvN1sX/GIuGkpwM=
=H0BJ
-----END PGP SIGNATURE-----