[comp.mail.sendmail] Embedded Spaces in Usernames

rpcfod@uarthur.UUCP (Robert Patt-Corner) (02/15/91)

Apologies if this has come up often ...

We're experimenting with an SMTP<===>Digital Mailbus gateway from WIN/TCP
to move mail back and forth from the Internet to a local All-In-One
system.  Our All-In-One has a username convention of 

FirstName<SPACE>Lastname

which is quite common in the A1 world.  Our Sun sendmail is very unhappy with the embedded space in the username -- obvious efforts like enclosing that
portion of the address in single or double quotes fail.

Does anyone out there have experience with this, a portion of a SENDMAIL
script that deals with embedded spaces, or the like?  If so, please let me
know.

The syntax for the "whole" address, for example is:

   Robert Patt-Corner%A1.MROUTER@MailBusGatewayHost

Obviously something has to be done with the "Robert Patt-Corner" portion
to make this go.

Thanks ...

jch@dyfed.rdg.dec.com (John Haxby) (02/18/91)

It is quite reasonable to put parts of the address that have spaces in them in quotes, so, in your example,

	Robert Patt-Corner%A1.MROUTER@MailBusGatewayHost

becomes

	"Robert Patt-Corner"%A1.MROUTER@MailBusGatewayHost

although getting the quotes in place without support from the gateway might be a little tricky. Where the quoting falls flat on its face, however, is when you try to use something like UCB mail that doesn't believe in quoted atoms: in this case, the address is split into

	"Robert
and
	Patt-Corner"%A1.MROUTER@MailBusGatewayHost

which sendmail takes a dim view to (unbalanced "s, it says).  There are a selection of ways to get around this, the best ones involve throwing UCB mail away and using MH :-) which does believe in quoted atoms. Another solution involves using an undocumented feature of Message Router--which I don't think I should tell you about, sorry.

ULTRIX Mail Connection (V1.1) which is supplied by us (Digital) chooses to quote atoms with spaces in them, at cost of not working with UCB mail--a possible future product may fix this.

[Personally, I think UMC is far better than the WIN/TCP product, but then I wrote it!]
-- 
John Haxby, Definitively Wrong.
Digital				<jch@wessex.rdg.dec.com>
Reading, England		<...!ukc!wessex!jch>