when to split a package into architecture: all and architecture: any halves
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Hi,
I have a package of bup - a git-based backup tool - sitting in the NEW
queue. bup is 99% python, with a small .so module for some
speed-critical code.
I decided to have an architecture: all bup-common package for the
majority of the program, and a small architecture: any package for the
.so module, which depends on the -common package. I did this mostly to
see how hard it was, and to see whether it would save mirror space.
On my system, the -common package ends up being 76K and the architecture
dependent package 12K. By my calculation:
12K * 12 release architectures + 76K = 225K total
As appose to,
(76K + 12K) * 12 release architectures = 1M total
So the savings are quite significant, in terms of the total package size
(which is relatively small).
Is this worth the added complexity of two binary packages? Are there
other advantages/disadvantages to consider?
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.10 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/
iEYEARECAAYFAkv6etMACgkQFotOcXAy8jjFBQCfUGN+tDNgqIMr+ycfbaegBlMT
ejYAoIsARMvB4NqwN115pi3MKyc4LWef
=RZ5F
-----END PGP SIGNATURE-----
Reply to: