Re: the quality of Debian's diff.gz
On Sun, 1 Jun 2008 14:59:06 -0700, Don Armstrong <don@debian.org> said:
> On Sun, 01 Jun 2008, Manoj Srivastava wrote:
>> I said that a stacked diff was impossible to create m4echanically
>> from topic branches.
> Given the topic and the merged branch, though, there's no inherent
> reason why you couldn't create them. [Even if just by applying all the
> topic diffs which apply cleanly in some deterministic order, then
> generating a final diff from the merged branch to the tree from the
> applied topic diffs.[1]]
The topic diffs do not apply cleanly in some deterministic
order for all my packages. Often, the conflict resolution happened
several release cycles ago, and the patches and resolution have moved
around due to subsequent development, and it takes significant effort
to re-integrate stuff.
Now, if you think it is that simple, please just give me a tool
that will take my topic branches and do the right thing, and I'll see
about using it. Until then, this seems just a lot of hand waving, smoke
and mirrors to me.
> 1: This would obviously be ugly in the case where you had loads of
> diffs which conflicted, but I'd suspect this only occurs frequently in
> dead-upstream packages.]
No. It occurs also in packages where fedora differs wildl from
Debian, and you n4eed to get a security policy that actually
lets a Debian machine work.
Sure, refpolicy patches are soon accepted upstream, but they are
developed in Debian, and there are massive overlaps.
manoj
--
In Minnesota they ask why all football fields in Iowa have artificial
turf. It's so the cheerleaders won't graze during the game.
Manoj Srivastava <srivasta@debian.org> <http://www.debian.org/~srivasta/>
1024D/BF24424C print 4966 F272 D093 B493 410B 924B 21BA DABB BF24 424C
Reply to: