Sendmail currently ships with being a relay by default turned off. Also, all BSDs ship with sendmail set up that way. And they're not ancient versions anyway. (8.12.x, last time I checked). Of course NetBSD ships with postfix, but I harly use it. Sendmail performs well enough, and m4 isn't the hassle everyone thinks it is.
Like some other poster says, postfix is actually pretty fussy when it comes to virtual domains. In sendmail you use a sendmail.cw, plonk all your recieving domains in there an be done with it. And there's milter.
Sendmail is good enough for me, the same as postfix would be, but I don't see a solid reason to switch.
Oh, I haven't seen a compromise through sendmail in YEARS. Yeah sure there were bugs, but if you keep your world upt o date with cvsup or cvs, the holes get plugget VERY fast.
Of course now I get al the exim, qmail and postfix fanboys blasting at me, but sendmail works well. Works good enough for most. Heck, if sendmail were so insecure, why is OpenBSD still including it in it's base?
Don't get me wrong, postfix is a nice MTA. Yes, it is easier to set up depending on what you think is "easy", but still, it's a nice MTA, but no reason to not use Sendmail if you can help it.
ncreased efficiency due to productivity tools not available on the target platform. This is particularly evident if your target platform is less powerful, and you can afford a much better development computer.
Sun is just listening to their customers and giving them what they want. The customers want linux. So Sun messes around with it. The customer is king (well, usually it is).
Since Sun is a hardware shop, they don't really care much. As long as you use their hardware, they don't really care what the fsck you run on it.
in every source file so it displays with every source file you compile. One can cruft together a check for autoconf that sets this #define, with an underscore so it won't clash with other stuff.
Then why aren't you migrating? *nix is a lot easier to maintain in large numbers.
Case in point: say, you need a security patch roled out on all the production machines. Say your platform of choice is FreeBSD. I'd do it like this:
A central build farm builds updated worlds and kernels for all the types of machines. That is it's sole task. The kernels get distributed to the right machines which use rsync to sync their kernels, modules and worlds when they need to.
The same buildfarm builds updated packages which get pkg_added from a central repository. This too can be done in a scheduled fashion.
Then, the configurations of all the machines get synced up with cfengine if need be.
There you go... Almost hassle free corporate wide rollouts of updated software/patches. I've built such a system once to maintain 20+ FreeBSD machines in a webserver and build farm. And the way it worked scaled beautifully. The FreeBSD boxes never had a ports or a world/kernel source tree.
Sure, it's a few days work to set up, and it'll probably take some weeks of initial testing, but once you get it working, you hardly have to touch it anymore. Heck, with a little elbow grease you could muscle it to work with other *nixen.
Although I like the idea, I have one gripe at SpamAssassin. It is way to resource hungry to process every incoming mail for my taste. At my mail site, mail gets screened by dnsbls and greylisting first, then mimdefang weeds the junk out (like html in attachements) and only _then_ does spamassassin get a cut of the action.
Oh, and I never open html mail as html, always in the text/plain view. 99% of all html mail I get is spam anyway.
Slashdot Mirror
False:
Well well well, ancient huh? Whatever. Yes, that's openbsd's default sendmail as of version 3.3
Like some other poster says, postfix is actually pretty fussy when it comes to virtual domains. In sendmail you use a sendmail.cw, plonk all your recieving domains in there an be done with it. And there's milter.
Sendmail is good enough for me, the same as postfix would be, but I don't see a solid reason to switch.
Oh, I haven't seen a compromise through sendmail in YEARS. Yeah sure there were bugs, but if you keep your world upt o date with cvsup or cvs, the holes get plugget VERY fast.
Try better. I'm not convinced.
Milter is one of the things that's keeping me with sendmail.
Of course now I get al the exim, qmail and postfix fanboys blasting at me, but sendmail works well. Works good enough for most. Heck, if sendmail were so insecure, why is OpenBSD still including it in it's base?
Don't get me wrong, postfix is a nice MTA. Yes, it is easier to set up depending on what you think is "easy", but still, it's a nice MTA, but no reason to not use Sendmail if you can help it.
but promising! Clearly, Sun has cooped something that looks good. Let's hope they'll be a nice player and release this vor x86 as well.
Huh?.
Oh, you said "May"... never mind...
Ah hah! The sign of a true sociopath. :)
The bastards.
I agree with everyone here. If you don't want to end up like a lays crisp, stay away from the electric stuff.
(ObAdvice) DON'T TOUCH UNLESS YOU WISH TO SHORTEN YOUR LIFESPAN WITH A SCARY AMOUNT.
yeah, "Redundant", maybe, but it cannot be repeated enough times.
And what if you zip it up uncarefully so the foreskin... yeech... ow...
You _do_ know that this is exactly the reason why slashdot is called slashdot, right?
Yah, she was all torn apart about it.
FreeBSD runs Linux Oracle just fine through the Linux ABI if need be. So you still have support, What's the problem?
Switch to BSD! I hear FreeBSD is nice. Also in the enterprise. And a license that does not make $neckties nervous.
Gosh...
Since Sun is a hardware shop, they don't really care much. As long as you use their hardware, they don't really care what the fsck you run on it.
One can turn the thing off, yielding a black image.
Let's keep it on topic :)
Ah yessiree matey, she's going under.
(yes, a joke. sorry. couldn't help it. no go patch your winders box)
Nah, even better would be something like this:
#ifdef _SCO
#warn Please pay us $699!
#endif
in every source file so it displays with every source file you compile. One can cruft together a check for autoconf that sets this #define, with an underscore so it won't clash with other stuff.
Case in point: say, you need a security patch roled out on all the production machines. Say your platform of choice is FreeBSD. I'd do it like this:
A central build farm builds updated worlds and kernels for all the types of machines. That is it's sole task. The kernels get distributed to the right machines which use rsync to sync their kernels, modules and worlds when they need to.
The same buildfarm builds updated packages which get pkg_added from a central repository. This too can be done in a scheduled fashion.
Then, the configurations of all the machines get synced up with cfengine if need be.
There you go... Almost hassle free corporate wide rollouts of updated software/patches. I've built such a system once to maintain 20+ FreeBSD machines in a webserver and build farm. And the way it worked scaled beautifully. The FreeBSD boxes never had a ports or a world/kernel source tree.
Sure, it's a few days work to set up, and it'll probably take some weeks of initial testing, but once you get it working, you hardly have to touch it anymore. Heck, with a little elbow grease you could muscle it to work with other *nixen.
Unix has the way in(tm)
Right... This proves it. HTML needs a tag. :)
Excuse me? Slashdot unbiased? Oh well, whatever you say,... Care to invest in a sailing trip around zwitserland?
Oh, and I never open html mail as html, always in the text/plain view. 99% of all html mail I get is spam anyway.