Re: binary NMUs and version numbers
Andreas Metzler wrote:
Anthony Towns <aj <at> azure.humbug.org.au> writes:
Hrm, why isn't this 1.2+20041208-1 ? Isn't the date describing the
upstream version? Or "1.2-20041208-1", or "1.2+cvs20041208-1" or whatever.
-rw-rw-r-- 16 katie debadmin 2908273 May 2 2004
-rw-rw-r-- 16 katie debadmin 412082 Nov 17 10:17
It seems to result in rather large diffs, and I can't really see the
It is a payoff, larger diff for less frequent orig.tar.gz uploads. Instead
of uploading a 3MB mutt_1.5.6-20040915.orig.tar.gz the mutt maintainers can
upload a 400KB mutt_1.5.6-20040915+1.diff.gz when updating to CVS 20040915.
Hrm, there seem to have been nine versions since the 1.5.6 orig was
uploaded in May. The total sizes of the version (source, and all debs
for all architectures), and the size of the .diff.gz's seem to be:
1.5.6-1: 17109678 59577
1.5.6-20040523+1: 1382297 102196 +3M - 40k
1.5.6-20040523+2: 15671819 102236 - 40k
1.5.6-20040722+1: 14633689 154874 +3M - 100k
1.5.6-20040722+2: 11890401 155978 - 100k
1.5.6-20040803+1: 14647304 301927 +3M - 250k
1.5.6-20040818+1: 14940217 313439 +3M - 250k
1.5.6-20040907+1: 15118784 409724 +3M - 350k
1.5.6-20040907+2: 15131102 412082 - 350k
Uploading a new 3M .orig.tgz for each CVS update would introduce 3M
uploads every now and then, but reduce the .diff down to 60k (or less?),
so doing it that way would cost 15MB in orig.tgz uploads, and save
something like 1.5MB in diffs. So 13.5MB extra from a total of 120MB, so
that's about an extra 11%.
Dunno that that's really worth it. Pity we don't support more than two
files to build up the source we want in a more fine grained way.