Op di 15-07-2003, om 01:01 schreef Craig Sanders: > On Mon, Jul 14, 2003 at 08:25:47PM +0200, Wouter Verhelst wrote: > > Op ma 14-07-2003, om 08:49 schreef Craig Sanders: > > > On Sun, Jul 13, 2003 at 10:31:51AM +0200, Joey Hess wrote: > > > > For sarge we have two options for the default MTA in base: > > > > > > > > a. replace exim with exim4 > > > > b. no MTA installed by default, add a MTA task > > > > > > why not: > > > > > > c) replace exim with postfix? > > > > Why? Do you have any compelling reasons against exim? > > it's more that postfix has numerous compelling advantages. it's small, it's > fast, it's very flexible, it's very easy to configure, it's secure, it scales > beautifully from the very smallest mail systems to the very largest, and more. Except for the 'small' part, your arguments apply perfectly to exim. wrt the 'small' part, I don't really think that any mailserver with as much features as both exim and postfix have can still be called 'small'. 'lots of features' and 'small' are two things that don't work well in the same sentence (and no, I didn't say you said postfix has lots of features, but it does). > > Debian has been installing exim by default for quite a while now, and it > > works well enough for most people[1]; and for those people for which it does > > not work well enough, it's easy enough to install another MTA -- we're not > > FreeBSD[2] ;-) > > > > I see no reason to change that, although exim4 as opposed to exim3 would be a > > good thing, of course. > > if we're going to change from one MTA (exim3) to another (exim4), then we might > as well change to a better MTA (postfix) at the same time. Unless you come up with reasoning as to why postfix would be better, compared to exim, I see no reason for that. And with reasoning, I mean 'technically fundamented reasoning'. Not just 'postfix is faster, smaller, more flexible, easier to configure, more secure, and scales better'. > exim was only chosen as default MTA, with the proviso that the choice would be > reviewed when circumstances changed, because postfix's licence (IBM public > licence) needed some clarification at the time to be truly DFSG free. that > clarification happened years ago. Sure, but 'review a choice' doesn't mean 'revert it'. > > [1] as in "for most people, the choice of MTA doesn't really matter, and exim > > is quite flexible" > > true, which is why we should install the best available by default. I still fail to see why exim would be better than postfix. Exim is a great MTA. > > [2] I have nothing against FreeBSD (in fact, my coworker is a FreeBSD > > developer), but their procedure to switch MTA's could have been made a > > little easier. > > nobody's proposing that existing installations of exim3/exim4 be replaced with > postfix. the proposal is that the default MTA on new debian installations is > postfix rather than exim. That wasn't my point with this paragraph. My point was that installing a different MTA is an operation as simple as '$PKGMANAGER install $MTA_OF_CHOICE', it's not like hours of work or so. Even more so, you don't even have to install our chosen default on new systems, you can just select the 'right' one when the installer is still running. The same isn't true for FreeBSD (although it's not difficult once you know what to do, just installing your MTA of choice on a FreeBSD system isn't enough to get rid of sendmail). As such, what we (the Debian Developers) choose as our default MTA doesn't really matter, as long as we don't pick the worst one out there, which exim definately isn't. You, personally, may not like exim; but the fact remains that exim, especially exim 4 with its debconf configuration, is quite easy to configure and flexible enough to support most installations (as in: 99% of them -- I don't believe there's any MTA out there that will work in _any_ remotely possible situation). I support exim for our default as opposed to any other MTA since it's been our default before. Think "if it ain't broke, don't fix it".
Attachment:
signature.asc
Description: Dit berichtdeel is digitaal ondertekend