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

Re: Lintian ERROR saying dpatch is obsolete



On Mon, Nov 28, 2011 at 12:54:57PM +0100, Alexander Wirt wrote:
> Its simple and things like dpatch-edit-patch are just great. I now use dpatch
> for round 8 years and it worked every time. I don't see any reason to move
> away.
> 
> And I still like the "never touch a running system" approach. If dpatch works
> without problems, why deprecate it?

Well, there is also the cost of diversity to take into account. I don't
doubt that for you, right now, dpatch is better than quilt. You are used
to it and you're happy with it. But in a project as large as Debian
diversity has a cost.

Think about learning packaging (which is an important use case, given
that we often lament we don't have enough people power in Debian). The
cost of package learning is proportional to the number of tools
involved. Multiplicating the number of tools that do the same thing adds
up to that number, for anyone who has to deal with packages maintained
by diverse teams with potentially different habits.

To name another use case, we have learned in the past release cycles
that the only way to keep up with Debian releases is to have a
significant number of people that do NMUs. Given that we need those
people, we should also try to apply a principle of least surprise from
one package to another. For the Squeeze release I've NMU-ed packages
maintained in yada. Not. Fun.

The maintainer surely had the right to maintain them in yada, but that
choice induced a cost on the release cycle of others who had to learn
yada in the unfortunate case the maintainers stopped doing her job
properly.

We have a tradition in Debian on standardizing on interfaces, which is
good. But also standardizing on tools has value, because it reduces the
cost of diversity throughout the archive. If standardizing on tools is
considered to be too much, we should at least encourage uniformity.
That, I believe, is what Gergely is doing, and I applaud the effort.

(The technical merits of quilt vs dpatch is a totally different matter,
 but Colin & Ian argument "thou-shall-not-exec-stuff-while-patching"
 alone make me score quilt way higher than dpatch, for sanity of the
 archive sake.)

Cheers.
-- 
Stefano Zacchiroli     zack@{upsilon.cc,pps.jussieu.fr,debian.org} . o .
Maître de conférences   ......   http://upsilon.cc/zack   ......   . . o
Debian Project Leader    .......   @zack on identi.ca   .......    o o o
« the first rule of tautology club is the first rule of tautology club »

Attachment: signature.asc
Description: Digital signature


Reply to: