Date: Thu, 27 Oct 2005 04:56:55 +0300 From: Ion-Mihai Tetcu <itetcu@people.tecnik93.com> To: Ertan Kucukoglu <ertank@ozlerplastik.com> Cc: ports@freebsd.org Subject: dspam-3.6.0 on FreeBSD (was: Re: dspam for FreeBSD) Message-ID: <20051027045655.2e3a435b@it.buh.tecnik93.com> In-Reply-To: <435F4E0C.8000209@ozlerplastik.com> References: <435F4E0C.8000209@ozlerplastik.com>
next in thread | previous in thread | raw e-mail | index | archive | help
[ ports@ cc'ed as a safety measure for my inbox ] On Wed, 26 Oct 2005 12:36:12 +0300 Ertan Kucukoglu <ertank@ozlerplastik.com> wrote: > Hi, > > I've found your e-mail address in Makefile for dspam port. > > I know that there is a recent new version of dspam. I wonder if it > will be available in ports anytime soon. Would you advise please? 3.6.0 is already available on FreeBSD as dspam-devel port. (As a note it was added in our Ports Collection the same day the official release got out - this time we beat all other OSes, for example debian still hasn't it :) If you would be reading dspam-users mailing list (which is something I recommend) you'd already know this (I'm active on dspam-* lists and I've posted there on this subject); You an safely: portupgrade -R -o mail/dspam-devel dspam-3.4\* Reasons I haven't MFC it yet in mail/dspam: - there are a lot of changes (like dynamic back-end support) which I feel it's better to have them tested for a while before declaring them safe (esp. the new hash_drv, esp. on more "exotic" arch like sparc where on solaris and linux I have crash reports and a fix in cvs) - there are a few bugs (in CGI) which I haven't have time to patch - MaxMessageSize from dspam.conf also applies to clamav virus scanning (if WITH_CLAMAV= yes) which IMO is a nice way for self-shooting; given the way this is implemented now (see dspam.c) patching this elegantly is not trivial. (- there also is a memory leak report which I haven't been able to reproduce yet.) I'm waiting for a reply from Jonathan about the points above to know if I should go ahead with local patches or we will have a 3.6.1 release which will address them soon. Jonathan does his devel work on linux and MacOSX. There have been cases when a version of dspam was working perfectly on linux and crashing on FreeBSD due to bugs in either dspam or FreeBSD (thread libs come in mind, grrrr). Jonathan Zdziarski's (and the rest of dspam contributors) ideas and work have made dspam (one of) the best anti spam applications both from accuracy and administrations point of view. Dspam was and is a very active project. Maybe because of this the quality of releases wasn't the best (small bugs had a way to creep in in the releases) and it has always been a challenge to provide the latest features _and_ a stable product. Producing a diff to upgrade a port and making the canonical tests (porlint, package, ...) is a relatively easy process but a as port maintainer I think I my _obligation_ to actually test on my semi-production test server the new versions before offering them to the users via the Ports. Given the number of possible configurations this takes (a lot of) time. The first step though better releases was made when I created mail/dspam-devel port as a way for daring people to easily run (and hopefully help with) the latest version. A second step, with much help from Andreas Klemm (one of our commiters and aspfileter maintainer), was taken about 1 1/2 or 2 years ago when we pushed Jonathan to have all devel done in a separate branch and only bug fixes imported in the stable branch (see how dspam is versioned on dspam home page) and so stable and devel releases came out. 3.6.0 was the first release to have a QA snapshot (this was a RC3) so I can say that we've made a long way from the beginning. Still, as you see, it isn't perfect. But I have reasons to hope the quality of the releases will continue to improve <rant> We are used on FreeBSD with a level of quality (both in the base system and in thirds party apps) unseen on many other OSes out there; when installing a port we expect it to work, not merely install and deinstall clean; and this takes time for the maintainer and port commiters. I _do not_ think having a port upgraded a week sooner is worth renouncing this quality. If you or anyone else want to see dspam ports upgraded sooner you can: run -devel port and sens me problem reports and patches; respond to my requests for testing new version/features (indeed, being conservative, I estimate the number of mail/dspam* users to be more that 1000 but when I ask for testers I receive 1 -3 replies; but request for upgrades - hehe). </rant> -- IOnut Unregistered ;) FreeBSD "user" "Intellectual Property" is nowhere near as valuable as "Intellect"
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20051027045655.2e3a435b>