On Tue, Apr 09, 2013 at 07:49:13AM +0200, Petr Pisar wrote:
On Mon, Apr 08, 2013 at 03:31:19PM -0500, Derek Martin wrote:
As for the comment regarding the Mutt fork: in principle that's great,
but in practice IMO forks are bad: They divide the effort, and cause
problems potentially for both (or all) projects involved, trying to
track important changes between them. I don't like forks, and unless
there's a very very clear center of mass developing around one (and
there have been a few prominent examples in my lifetime), I don't ever
want to contribute to one.
I absolutly agree with you. But there is at least one serious reason for the
fork: Current upstream lacks sense for maintenance.
This can be demonstrated by bug <http://dev.mutt.org/trac/ticket/3300>
(mutt-1.5.20: mutt terminates with >=gpgme-1.2.0) which has been opened
4 years ago. It's pure bug which prevents from using gpgme encryption by
mutt. It's easy to fix, it has attached correct patch, the patch applied
virtualy by any distribution, yet the patch has been commited into development
brach only 4 months ago and none mutt release has contained the fix yet.
In other words, the mutt is 4 years affected and the burden of maintainance
lies on downstream distributors. This is the only but serious problem I can
see at mutt project.
I have little knowledge of how mutt actually is packaged for
various OS distributions, so could you give me an overview of the
process? I assumed that packagers would be pulling in bugfixes
from the mutt mercurial repository and overlaying any distribution
specific patches on top. How does a new release versus pulling
from the repository make a difference in your process?
Back in December I put out a call for patches
(http://marc.info/?l=mutt-dev&m=135586839208976&w=2) and I thought
I had pulled in patches from all the major OS distributions. If I
have missed some, please do let me know.
