Re: sendmail masquerading question
On 9/9/07, Bill McGonigle <[EMAIL PROTECTED]> wrote: >/etc/postfix/transport > to have the lines: >root: >* : smtp:saturn.syslang.net I take it that means, in essence, "no special rule for user 'root', so do what is normally done (deliver locally); send everything else to SATURN"? Does that override everything else? What would be needed to configure Postfix on SATURN to rewrite "From" addresses (masquerade, in Sendmail terms) for Steven's machines to be <@syslang.net>? Don't forget it has to rewrite mail from PLUTO as well as locally-originated mail. > I tend to run sendmail with simple configs and postfix for complex > stuff, mostly because I don't ever want to take the bat book off the > shelf again. ;) I regard sendmail.cf as I do object code: Opaque, intended for machine and not human consumption, and best generated by other programs. I understand there are those who actually manipulate sendmail.cf directly. There are also those who can write real programs in assembly directly. I am not one of either group. sendmail.mc is pretty reasonable, and well-documented. The hardest part is usually mapping one's abstract needs and wants into configuration directives. That's a universal problem, although I suspect Sendmail's terminology makes this harder than it could be. I would guess Postfix, lacking Sendmail's 30 or so years of historical baggage, does this better. I haven't been able to understand Postfix without RTFM. If I could, I'd be using Postfix today. But I've already RTFM for Sendmail and haven't RTFM for Postfix, so I'm still using Sendmail. It's easier for me (in the short term) to expend minimal effort maintaining my Sendmail knowledge than it is learn an entirely new software system (which would likely yield better long-term rewards). I find this is a common scenario in the IT world. -- Ben ___ gnhlug-discuss mailing list gnhlug-discuss@mail.gnhlug.org http://mail.gnhlug.org/mailman/listinfo/gnhlug-discuss/
Re: sendmail masquerading question
Wow! Impressive reply, Ben. At the risk of incurring rotten fruit, alternately one could configure: /etc/postfix/transport to have the lines: root: * : smtp:saturn.syslang.net (I think, untested) and, include it in: /etc/postfix/main.cf with: transport_maps = hash:/etc/postfix/transport On a redhat-ish box: /usr/bin/sudo /usr/bin/system-switch-mail takes care of the switch back and forth among MTA's, so it's pretty easy to try out. I tend to run sendmail with simple configs and postfix for complex stuff, mostly because I don't ever want to take the bat book off the shelf again. ;) -Bill - Bill McGonigle, Owner Work: 603.448.4440 BFC Computing, LLC Home: 603.448.1668 [EMAIL PROTECTED] Cell: 603.252.2606 http://www.bfccomputing.com/Page: 603.442.1833 Blog: http://blog.bfccomputing.com/ VCard: http://bfccomputing.com/vcard/bill.vcf ___ gnhlug-discuss mailing list gnhlug-discuss@mail.gnhlug.org http://mail.gnhlug.org/mailman/listinfo/gnhlug-discuss/
Re: sendmail masquerading question
After an off-list conversation with Steven Orr , I have a
better idea of what he's after.
SYMPTOMS
Steven has hosts SATURN and PLUTO, and a registered domain name
. He wants SATURN to be responsible for most mail
delivery. The MX records in the public DNS effectively tell the world
to deliver mail to SATURN, and that works fine. From SATURN, he can
send mail to the outside world, and that works fine. On PLUTO, he can
send mail to the outside world, and PLUTO relays the mail though
SATURN, as he wants it to. On SATURN, he can send mail to addresses
<@syslang.net>, and that works fine.
The problem arises when he sends mail on PLUTO to addresses
<@syslang.net>. He wants PLUTO to hand off such mail to SATURN.
Instead, PLUTO is (attempting to) deliver the mail locally, to
mailboxes on PLUTO.
COMPLICATIONS
The term "domain" gets interchangeably used (by everybody) to mean
"a particular entry in DNS" or "all hosts with the given name as a
parent domain". In other words, saying "the syslang.net domain" might
mean "anything ending in *.syslang.net", or it might mean whatever
records I get when I ask my DNS resolver about , and not
the various hosts within. I have tried to keep my terminology clear
by context in this discussion, but this is still something to remain
conscious of.
There are multiple Sendmail configurations involved in this
environment: The Sendmail on PLUTO and the Sendmail on SATURN. There
is also "Sendmail in the general case". Because Steven wants
different behavior on different hosts, we will have the Sendmail
configurations on each host behaving differently. It is thus
essential to maintain awareness of *which* Sendmail is being discussed
at any given point.
As a result of these complications, this discussion is rather
verbose, even by my standards. Bear with me, folks.
CAUSE
This is almost certainly being caused by the presence (on PLUTO) of
in Sendmail's list of local host names (in terms of
Sendmail internals, the {w} class). Any Sendmail will consider itself
the destination host for any such name. If the list contains
, Sendmail will consider itself responsible for that
domain, and will attempt local delivery for <[EMAIL PROTECTED]>.
Typically, the local name list contains the system's canonical name
(usually the same as the output of "hostname -f"), along with any
names listed in a "sendmail.cw" or "local-host-names" file (the exact
name and location of the file depends on Sendmail version, distro
config, and local config, but they typically exist under /etc/ or
/etc/mail/).
SMART_HOST won't affect the local name list, since a smart host is
only used for mail *not* destined for the local system. Since that
list means "I am the destination host for these names", Sendmail
thinks no further relaying is necessary, and SMART_HOST never gets
involved.
CORRECTION
Steven tells me Sendmail on PLUTO is configured with
MASQUERADE_AS(`syslang.net'). During the off-list conversation, I had
stated that I suspected doing so put in the local name
list. I now believe I was mistaken on that point. The docs don't say
it does, and the more I think about it, the less that makes sense.
RESOLUTION
So I now suspect may be listed in one off the files
mentioned above ("sendmail.cw" or "local-host-names"). Or it may be
that the OS on PLUTO thinks it's fully-qualified host name is
("hostname -f" command). I would advise Steven to check
those.
FURTHER SUGGESTIONS
Moving on to more general advice toward what the "best" solution is
for Steven. I suspect masquerading on PLUTO isn't really the "right"
solution for him.
NULL CLIENT
My first thought is that a null client configuration might be what
Steven really wants. Configuring Sendmail as a null client causes
Sendmail to just hand *all* mail off to a (specified) host. That
would be SATURN, in this case. The advantage would be that Steven
would only have to worry about mail administration on SATURN. Mail on
PLUTO becomes uninteresting.
However, Steven also tells me that he would like root's mail on
PLUTO to deliver locally. Because a null client is such a minimal
configuration, you cannot make exceptions. Sendmail doesn't do any
interpretation at all; it just passes everything on to the specified
host. So the null client idea is out.
LOCAL RELAY
My next idea is that, since Steven wants *almost* all mail routed
though SATURN, is that on PLUTO, he should define SMART_HOST and
LOCAL_RELAY (both targeting SATURN), omit the masquerading directives
on PLUTO entirely, and define LOCAL_USER(`root'). That should result
in just about all mail being relayed though SATURN. Mail to root
would still deliver locally. Any mail explicitly addressed to
<[EMAIL PROTECTED]> would also deliver locally, but presumably that's
intended, since it was asked for.
(I believe one can additionally override that last behavior by
setting FEATURE(`stickyhost') and defining MAIL_HUB to be SATURN as
well. But I'm not
Re: sendmail masquerading question
On Sep 9, 2007, at 16:04, Steven W. Orr wrote: > When sending mail from pluto, I want the addresses that the message is > sent to, to be changed so that *any* address that is bound for > syslang.net > does not get sent to pluto.syslang.net. All addresses on saturn are > valid > on pluto, but I don't want any mail sent from pluto to not be sent > through > saturn. (I do already have EXPOSED_USER(`root')) So, if you're on pluto, you want to be able to do: mail steveo rather than: mail [EMAIL PROTECTED] and have the mail get relayed through saturn, right? If you have only a few users, .forward files would do it, assuming pluto knows about syslang.net's MX record from its network vantage point. -Bill - Bill McGonigle, Owner Work: 603.448.4440 BFC Computing, LLC Home: 603.448.1668 [EMAIL PROTECTED] Cell: 603.252.2606 http://www.bfccomputing.com/Page: 603.442.1833 Blog: http://blog.bfccomputing.com/ VCard: http://bfccomputing.com/vcard/bill.vcf ___ gnhlug-discuss mailing list gnhlug-discuss@mail.gnhlug.org http://mail.gnhlug.org/mailman/listinfo/gnhlug-discuss/
Re: sendmail masquerading question
On 9/9/07, Steven W. Orr <[EMAIL PROTECTED]> wrote: > When sending mail from pluto, I want the addresses that the message is > sent to, to be changed so that *any* address that is bound for syslang.net > does not get sent to pluto.syslang.net. I'm not sure I follow what you mean on the above. Does Sendmail on pluto believe its canonical name is "pluto.syslang.net"? If so, mail sent from pluto would normally be delivered to whatever is behind TCP port 25 on 207.172.210.41. That's what the MX and A records for say to do. I assume that is the machine "saturn". If pluto believes it's canonical name is , well, it's going to attempt to handle mail for itself. So don't do that, then. Change pluto's canonical to be what it should be (which is *not* , per your description). If your problems arise because you want the "From" address of mail sent from pluto to be set to something the outside world can understand (I presume you're running a private NAT LAN like most everybody is these days), you've got a couple options. You could actually set a public MX record for and tell saturn to handle mail for pluto as well. This would be desirable if you want accounts on pluto to have an identity discrete from to the outside world. Otherwise, use a Sendmail generics table to rewrite "From" addresses to use <@sysland.net> (or whatever). For example, my user account on the box I'm typing this on is "bscott". The box's canonical name (for Sendmail) is . But my email account is <[EMAIL PROTECTED]>. So I need to have my "From" address rewritten so the outside world understands it. To do that, I have the following lines in my /etc/mail/sendmail.mc file: FEATURE(genericstable, `hash -o /etc/mail/genericstable') GENERICS_DOMAIN_FILE(`/etc/mail/genericsdomains') and the following in my /etc/mail/genericsdomains file: blackfire blackfire.bscott.local dragonhawk.dnsalias.org and the following in my /etc/mail/genericstable file: bscott [EMAIL PROTECTED] If I'm totally misunderstanding your question, maybe the above will help you clarify what you're after. :) > All addresses on saturn are valid on pluto, but I don't want any mail sent > from pluto > to not be sent through saturn. You may also want to explore the MAIL_HUB and LOCAL_RELAY options in addition to SMART_HOST. They permutate in ways which can be tricky. See: http://www.ilkda.com/sendmail/Relay.htm http://www.sendmail.org/m4/masquerading_relaying.html -- Ben ___ gnhlug-discuss mailing list gnhlug-discuss@mail.gnhlug.org http://mail.gnhlug.org/mailman/listinfo/gnhlug-discuss/
Re: Spreadsheets and precision?
"Bill Ricker" <[EMAIL PROTECTED]> writes: > Every spreadsheet since VisiCalc has done floating point division and > other basic numeric formuli in formulaic form as > > =3329.10/10 > > and used NamedFunctions() only for things without conventional (to > business folks) infix notation. As I said: >> I will admit to being a complete and utter idiot when it comes to >> spreadsheets :) I also said: >> I tried the obvious of increasing the cell format to go out to >> several decimal places. I guess I just didn't try the *really* obvious of "just plugging in the numbers" to see if the "obvious" would work. I'm not sure if that makes it obvious or not :) But, I do feel like a real dope now! -- Seeya, Paul ___ gnhlug-discuss mailing list gnhlug-discuss@mail.gnhlug.org http://mail.gnhlug.org/mailman/listinfo/gnhlug-discuss/
sendmail masquerading question
I have a linux box "saturn" running sendmail which is the "hub" where the MX record for my domain points. I have a box inside the house called "pluto" that sends all its mail to saturn using SMART_HOST for delivery. All is good in the universe. When sending mail from pluto, I want the addresses that the message is sent to, to be changed so that *any* address that is bound for syslang.net does not get sent to pluto.syslang.net. All addresses on saturn are valid on pluto, but I don't want any mail sent from pluto to not be sent through saturn. (I do already have EXPOSED_USER(`root')) The MASQ stuff that I see seems to only affect the From address and not the To or Cc etc addresses. Anyone know how to doo dees? -- Time flies like the wind. Fruit flies like a banana. Stranger things have .0. happened but none stranger than this. Does your driver's license say Organ ..0 Donor?Black holes are where God divided by zero. Listen to me! We are all- 000 individuals! What if this weren't a hypothetical question? steveo at syslang.net ___ gnhlug-discuss mailing list gnhlug-discuss@mail.gnhlug.org http://mail.gnhlug.org/mailman/listinfo/gnhlug-discuss/
Re: Spreadsheets and precision?
On 9/9/07, Paul Lussier <[EMAIL PROTECTED]> wrote: > So, how to people do actual division in > spreadsheets? The same way as you state Emacs does: With the slash (/). I'm just sayin'... ;-) -- Ben ___ gnhlug-discuss mailing list gnhlug-discuss@mail.gnhlug.org http://mail.gnhlug.org/mailman/listinfo/gnhlug-discuss/
Re: Spreadsheets and precision?
On Sun, 9 Sep 2007 14:01:56 -0400 "Bill Ricker" <[EMAIL PROTECTED]> wrote: > > Oh. I do see that now that I look. This strikes me as completely > > counter-intuitive. So, how to people do actual division in > > spreadsheets? Does Excel suffer from this as well? > > Yes. > > Every spreadsheet since VisiCalc has done floating point division and > other basic numeric formuli in formulaic form as > > =3329.10/10 > > and used NamedFunctions() only for things without conventional (to > business folks) infix notation. > > VisiCalc and Lotus 1-2-3 allowed >+3329.10/10 > and might figure out that > 3329.10/10 > was NOT a date and work, but OO.o.C and "Excel" insist on = prefix for > any non-constant formula Just to elaborate, floating point (doubles) uses 1 sign bit, 11 exponent bits and 53 mantissa bits. Note that the high order mantissa bit is hidden except in denormals which is why 1 + 11 + 53 = 65. The exponent is based on power of 2 which is why you might see a lot of 9 digits. You can use the format cell menu item to expand the visible precision, but you are limited about 17 digits. Some chips use 80 bits in their coprocessor. -- Jerry Feldman <[EMAIL PROTECTED]> Boston Linux and Unix user group http://www.blu.org PGP key id:C5061EA9 PGP Key fingerprint:053C 73EC 3AC1 5C44 3E14 9245 FB00 3ED5 C506 1EA9 signature.asc Description: PGP signature ___ gnhlug-discuss mailing list gnhlug-discuss@mail.gnhlug.org http://mail.gnhlug.org/mailman/listinfo/gnhlug-discuss/
Re: Spreadsheets and precision?
> Oh. I do see that now that I look. This strikes me as completely > counter-intuitive. So, how to people do actual division in > spreadsheets? Does Excel suffer from this as well? Yes. Every spreadsheet since VisiCalc has done floating point division and other basic numeric formuli in formulaic form as =3329.10/10 and used NamedFunctions() only for things without conventional (to business folks) infix notation. VisiCalc and Lotus 1-2-3 allowed +3329.10/10 and might figure out that 3329.10/10 was NOT a date and work, but OO.o.C and "Excel" insist on = prefix for any non-constant formula -- Bill [EMAIL PROTECTED] [EMAIL PROTECTED] ___ gnhlug-discuss mailing list gnhlug-discuss@mail.gnhlug.org http://mail.gnhlug.org/mailman/listinfo/gnhlug-discuss/
Re: Spreadsheets and precision?
Paul Lussier <[EMAIL PROTECTED]> writes: > Hi All, > > Does anyone know how to get any kind of decent precision out of either > oocalc or gnumeric? Dope slap! The obvious answer is: use emacs (M-x ses-mode), where formulas are elisp functions! (/ 3329.10 10) works, and makes a whole lot more sense than =QUOTIENT(3329.10;10) any day, *and* actually results in a correct answer! My mantra of "If emacs can't do it, then I don't need to!" seems to continue to be correct :) -- Seeya, Paul ___ gnhlug-discuss mailing list gnhlug-discuss@mail.gnhlug.org http://mail.gnhlug.org/mailman/listinfo/gnhlug-discuss/
Re: Spreadsheets and precision?
[EMAIL PROTECTED] (Lee A. Newcomb) writes: > This is where reading the doc. helps (marks added) :) > > QUOTIENT > This function is only available if Analysis AddIn is installed. For some reason, I don't see this statement anywhere, but since I have the function available, I must have that AddIn. > Returns the _integer_ result of a division operation. Oh. I do see that now that I look. This strikes me as completely counter-intuitive. So, how to people do actual division in spreadsheets? Does Excel suffer from this as well? /me goes back to emacs, which operates with expected/desired/useful behavior :) -- Seeya, Paul ___ gnhlug-discuss mailing list gnhlug-discuss@mail.gnhlug.org http://mail.gnhlug.org/mailman/listinfo/gnhlug-discuss/
Re: Spreadsheets and precision?
This is where reading the doc. helps (marks added) :) QUOTIENT This function is only available if Analysis AddIn is installed. Returns the _integer_ result of a division operation. Syntax QUOTIENT(Numerator;Denominator) Example =QUOTIENT(11;3) returns 3. _The remainder of 2 is omitted._ > > Hi All, > > Does anyone know how to get any kind of decent precision out of either > oocalc or gnumeric? > > I'm trying a simple formula: > =QUOTIENT(3329.10;10) > > Which leads to the answer 332.00 rather 332.90. Which is really just > wrong, especially since what I *really* want to do involves numbers > requiring much higher precision: > > =QUOTIENT(301.52;3329.10) > > Should not yield an answer of 0.00 :) > > I will admit to being a complete and utter idiot when it comes to > spreadsheets, but I tried the obvios of increasing the cell format to > go out to several decimal places. > > Any ideas? > > Thanks. > -- > Seeya, > Paul > ___ > gnhlug-discuss mailing list > gnhlug-discuss@mail.gnhlug.org > http://mail.gnhlug.org/mailman/listinfo/gnhlug-discuss/ > ___ gnhlug-discuss mailing list gnhlug-discuss@mail.gnhlug.org http://mail.gnhlug.org/mailman/listinfo/gnhlug-discuss/
Spreadsheets and precision?
Hi All, Does anyone know how to get any kind of decent precision out of either oocalc or gnumeric? I'm trying a simple formula: =QUOTIENT(3329.10;10) Which leads to the answer 332.00 rather 332.90. Which is really just wrong, especially since what I *really* want to do involves numbers requiring much higher precision: =QUOTIENT(301.52;3329.10) Should not yield an answer of 0.00 :) I will admit to being a complete and utter idiot when it comes to spreadsheets, but I tried the obvios of increasing the cell format to go out to several decimal places. Any ideas? Thanks. -- Seeya, Paul ___ gnhlug-discuss mailing list gnhlug-discuss@mail.gnhlug.org http://mail.gnhlug.org/mailman/listinfo/gnhlug-discuss/
