Proofpoint

OMG, how retrorse. Don't build your identity on such as you might get lost.
No, identity not the issue, just custom. In principle it is no problem to use alternatives, to use other OS, but it is nice to have an OS at hand and not to continuously waste time learning other OS/Programs.
 
Alternatives? What kind of problem?
Can you name a maintained MTA with BSD license, capable to relay a big number of Emails?

There is the tendence to put in the base systems programs like OpenSMTP or dma (dragonfly mail agent), but they are not substitutes for sendmail.

What kind of problem? Again, just used that the base system has such MTA and that it is not a specialised program that must be installed.
 
No, identity not the issue, just custom. In principle it is no problem to use alternatives, to use other OS, but it is nice to have an OS at hand and not to continuously waste time learning other OS/Programs.
It's not only that. It's about roots. Reading in /usr/share/sendmail/cf/README like this gives me a smile:
One normally masquerades as one of one's own subdomains (for example, it's unlikely that Berkeley would choose to masquerade as an MIT site)
From there you understand a piece of the motivation why the thing was developed - and why it was developed the way it is. And you wlll need such knowledge when something doesn't work as desired, to figure out and understand the root cause.

This has indeed to do with identity. Because there always is an identity; everything has an identity, even if you deny it. But understanding and appreciating that identity, You're usually better off considering the whole. Because when you do away with it, you do not know where you are. And when you do not know who you are, you cannot figure out the location of anything else. Then you will just take the manual, enter the commands as are written there, and wonder why it doesn't work as expected...

Sure, we can use or learn or do everything, and nowadays we're told this is only a matter of the necessary time&effort, which can be counted and measured. Like if we were machines, who only take a measurable amount of compute to do anything desired. But, are we? Are we indeed just machines, here to fulfil the desire of - yeah, of whom?
 
Can you name a maintained MTA with BSD license, capable to relay a big number of Emails?
Why does every Freebsd host need to relay a big number of emails? Most of mine don't.

There is the tendence to put in the base systems programs like OpenSMTP or dma (dragonfly mail agent), but they are not substitutes for sendmail.
DMA is not designed to be a full-featured MTA, sure, but what's the problem with Opensmtpd?
 
Ha! I knew that I'll get you with that. Using tradition wording is so revealing. :)
Yeah, but what's Your point in it?

See, I'm an old man now, and I start to look back onto my life. At times when we wanted to install emacs and sent an email in the morning, and at evening we got back hundred emails with slices of the source package. At these times I looked up at guys like Eric Allman, and I adored them. So this is part of my identity. What to do with it? Put it in a museum and forget about it?
 
Why does every Freebsd host need to relay a big number of emails? Most of mine don't.
Not every FreeBSD host need to do that, but (Free)BSD offers some software as part of it and such MTA
is part of it. It is not like Linux that is only the kernel and all the rest is a question of personal choice.

OpenSMTP is also not a replacement for sendmail, it is more like dma.
 
Are we indeed just machines, here to fulfill the desire of - yeah, of whom?
The use of "we" gets just a little too profane nowadays. We the people ... etc.
I get more and more repelled on what "we" may be.
Put it in a museum and forget about it?
On a personal stage I'd suggest to backup on a backup medium that survives. Then get someone to put a copy in there:

And yes, the sentimentality of men getting aged even hits me occasionally. But what bothers me most is the stupidity of the species I belong to, which implies a "we" I cannot deny. *)

This species is now at the stage having to prove "being human" by solving captchas. When it comes to this, it is a privilege being blind.

*) but some can as this joke (relevant to identity) suggests:
If a horse is neighing like a horse, and is galloping like a horse but is saying nowadays "I'm not a horse!", then you met a Russian horse.
 
Not every FreeBSD host need to do that, but (Free)BSD offers some software as part of it and such MTA
is part of it. It is not like Linux that is only the kernel and all the rest is a question of personal choice.
Why does base need a full-featured MTA?

OpenSMTP is also not a replacement for sendmail, it is more like dma.
This is not what the project claims
Be fast and efficient. OpenSMTPD must be able to handle large queues with reasonable performance.

Got a source for your claim that "it is more like dma?"
 
Why does base need a full-featured MTA?

Indeed I am wondering about that. There was the removal of named/BIND, only to then add unbound for local usage. So since the default configuration apparently uses sendmail only for local delivery, there is not really a need for sendmail to do that.
 
The use of "we" gets just a little too profane nowadays. We the people ... etc.
I get more and more repelled on what "we" may be.
*shrug* There is such a thing called human beings...

On a personal stage I'd suggest to backup on a backup medium that survives. Then get someone to put a copy in there:
Certainly. If people are no longer able to sustain, then at least the national value should be preserved.

And yes, the sentimentality of men getting aged even hits me occasionally. But what bothers me most is the stupidity of the species I belong to, which implies a "we" I cannot deny. *)

This species is now at the stage having to prove "being human" by solving captchas. When it comes to this, it is a privilege being blind.
So You seem to get the point. Or some of it.

*) but some can as this joke (relevant to identity) suggests:
If a horse is neighing like a horse, and is galloping like a horse but is saying nowadays "I'm not a horse!", then you met a Russian horse.
I entirely fail to get any sense out of that.
 
Can you name a maintained MTA with BSD license, capable to relay a big number of Emails?

There is the tendence to put in the base systems programs like OpenSMTP or dma (dragonfly mail agent), but they are not substitutes for sendmail.

What kind of problem? Again, just used that the base system has such MTA and that it is not a specialised program that must be installed.
Now that's really a nonsense criterium. 99.9% of the people don't have to relay a big number of emails (let's say >= 1 million/day); and these who do are doing it probably as their business and therefore don't worry about the software license too much.

And these who do care about this speed either are going to use Postfix, or if Postfix is too slow will switch over to Haraka.
 
Because FreeBSD is a full-featured operating system
Is it? Or is that marketing-speak?

I'd like to point to /usr/src/contrib where you find all "contributions" to what we call "base".

There is a lot in there but when is "full" fully featured? OpenBSD i.e. makes other selections.

As all these base contributions are more (OpenSSL) or less (tzdata) affected by security events. Because of that I happily "build my world" WITHOUT (see src.conf(5)) where possible. And I do prefer ports available instead, for some even alternative ports as this method opens a choice.

My point is, that those contributions are a selection. And selections imply that something is chosen while other is not. These decisions are rarely communicated transparently. I.e. the Mozilla Foundation gets revenues for providing Google a preferred place among the search engines. So sometimes there are reasons that are not technically or for compatibility reasons only.
 
There was the removal of named/BIND, only to then add unbound for local usage
I am against the removal, or at least for a substitution with a full featured DNS server.

Why doesn't that full-featuredness include a graphical desktop environment, an authoritative DNS server, or a Git client?
DNS server and Git client should be there, X11 perhaps also. The two first, as also sendmail, and perhaps X11, are not big programs. Why people get disturbed by their mere presence?

So as MacOS is a full featured OS, should also be (Free)BSD.
 
I am against the removal, or at least for a substitution with a full featured DNS server.
I agree.
DNS server and Git client should be there, X11 perhaps also. The two first, as also sendmail, and perhaps X11, are not big programs. Why people get disturbed by their mere presence?

So as MacOS is a full featured OS, should also be (Free)BSD.
Problem: how many product design artists are employed to work on MacOS?

This is a resource exhaustion problem: MacOS does not claim to be a suitable server OS (to some extent it can be, but then you have to adapt things youself). And also for the client experience the choices are limited: there is already a specific look+feel and you have more or less accept and use that. If you want a different window manager, again you're on your own.

It you begin with X11, I am certain there is a real mass of folks complaining that KDE or gnome or whatever is not readymade starting up. This approach does not find an end and is exponentially growing. And at the same time there is no personnel that could put all that together - nowhere. What we would get is something that might work and probably run a web-browser, but in real everyday use one will stumble over hundreds of loose ends and ill-conceived layouts.
I suppose it is no longer practical to have a server OS (something that can be rolled onto a cloud KVM) and a desktop/handheld OS in the same development line. And this will get worse as requirements diverge further.

Example:

This change makes rtsol run a bit earlier, so that people get their IPv6 configuration from DHCP faster (or whatever on that line).
At the same time it makes each single one of my jails take 15 seconds longer to startup. Because these solicitation requests do now run at a time when the firewall is not yet loaded, and run into timeout a couple of times.

So, what one side sees as an improvement, is a damage to the other.
 
In is terrible that this company bought sendmail:
It is irrelevant. As discussed in this thread, they bought the Sendmail Inc. company that had been owned by Eric and friends. They did not buy the source code, as it can not be bought, being under a freeware license. They are modifying and releasing the source code, and FreeBSD today uses one version of the source code modified by them. More about the license situation below.

It's from 2013
Exactly. Ancient news. But not terribly newsworthy.

If a software does strange things may not always be admins' responsibility. There are bugs, backdoors, vulnerabilities, zero-days etc. which all is beyond admins' responsibility.
Yes, but given that Eric is involved, and the sendmail source code is not overly complex or long, the probability of that is near zero.

any "hard working newbie" just cannot do this: review millions of lines of source code from all software installed. Even if the required knowledge were available on the personal level by a hard working professional, that individual would not even have the time for doing anything else. ...

The possibility of reading OSS is not an argument that is valid in admin's life experience. Except for some few bug bounty hunters there is no systematic review done for multiple reasons.
Yes, for one amateur to do it is probably too hard. But in the case of sendmail, my educated guess it that the FreeBSD project has sufficient manpower to review all changes to sendmail (here, man = Eric). And even for large software projects (such as freeware databases or kernels), the large industrial users are capable of doing an in-depth review. I'm quite sure that the big users do perform these full reviews, using teams of dozens or hundreds of engineers.

You don't need millions of people which use a software and have zero knowledge of understanding the actual source code. What you need is group of code review developers with deep understanding of that particular program which can approve the code changes ...
Exactly. And that happens among big users.

Unfortunately sendmail has not anymore BSD license, but we fall again in the problem of alternatives.
However, if you read the license (it's the file LICENSE at the top of the source tree), it sort of contains the BSD license as a special sub-case for open source usage.

Another question that was discussed in this thread: Why does FreeBSD still ship sendmail as the default MTU, given that 99.99% of all machines on the planet do not need to run a full-feature MTU which pairs with the open internet? There are two answers to that. The first one is that it is low risk, and great convenience to those FreeBSD users that have existing sendmail configurations. And it is no hassle to those people who want to use an alternative (simpler) MTU, they can install many of those from ports, disable sendmail, and instead use the MTU of their choice. It is what I do on my FreeBSD server.

Second argument: Tradition. I think sendmail has been used on *BSD for ~40 years; the author of sendmail is deeply involved in the BSD community.
 
Why doesn't that full-featuredness include a graphical desktop environment, an authoritative DNS server, or a Git client?
Because a complete operating system does not need one. Anything beyond the base system is forcing choice on the user which is not the path of FreeBSD. Forcing a graphic system then also forces those who do not want or need one to find a way to remove it.
 
sendmail_enable="NONE"
I don't think this is correct. Should be sendmail_enable="NO" as per /etc/defaults/rc.conf
Code:
# Settings for /etc/rc.sendmail and /etc/rc.d/sendmail:
sendmail_enable="NO"    # Run the sendmail inbound daemon (YES/NO).
The rest is fine.
 
I don't think this is correct. Should be sendmail_enable="NO" as per /etc/defaults/rc.conf
The function checkyesno() defines what is "correct" and generates error- and warning-messages.

Now test in /etc/rc.conf:

Code:
sendmail_enable="nonsense"
 
Back
Top