Re: Delay On Failed Send

Giganews Newsgroups
Subject: Re: Delay On Failed Send
Posted by:  Remy Lebeau \(Indy Team\) (re…@lebeausoftware.org)
Date: Fri, 19 Nov 2010

"Tom Hahn" <tha…@hahntech.com> wrote in message
news:535677473311859502.501624thahn-hahntech.c…@news.atozed.com...

> After noticing that there was a lengthy delay when a Send fails because
> of an authentication problem, I added some time-stamped logging. I
> found that the time between the Send and the Exception being raised
> could be as long as a minute. Is this time being taken by the server or
> is there some timeout factor that I have not found?

It depends.  What does your logging actually show?

> Additional  questions - Should all of the SASL components be used
> or is there a recommended set?

It depends on the capabilities of the server(s) you are connecting to.
Different servers support different algorithms.  Some support NTLM, some
CRAM-MD5 or CRAM-SHA1, some none at all.  The more SASL components you
enable in your code, the more server combinations you will be able to
support.

> Is the list to be used dependent upon the server being accessed?

Yes.  TIdSMTP looks at the server's capabilties (from the EHLO reply, so
make sure the UseEHLO property is set to True), picks the enabled SASL
components that match, then attempts to login with them one at a time until
either one succeeds or they all fail.

> I assume each component adds to the delay mentioned above

You have to look at your log to determine that.  Most SASL algorithms don't
have much delay at all.

--
Remy Lebeau (TeamB)

Replies

In response to

Delay On Failed Send posted by Tom Hahn on Fri, 19 Nov 2010