On Sun, 18 May 2014, Daniel Stenberg wrote:

> > Can we have some fallback for it...like trying out "EHLO" again with a
> > dummy valid hostname in case of EHLO failure with "501 5.5.4 Invalid
> > Address". This will ensure email delievery without reinitiatiating
> > complete email sending process.
>
> I haven't gone back to study what the RFCs says and given that nobody
> else has responded, perhaps you can write up a patch and suggest a
> way as that might be easier for people to respond to and comment on?

Sorry I meant to respond but got wrapped up with other stuff :(

I would be interested to know what a valid "dummy valid hostname" might be
and a patch might help explain things a little more as I got a little
confused as to what the problem really is especially as I would always
recommend using a valid hostname.

It is, however, worth noting that Tom could separate the connection and
authentication process from the mail sending by using CURLOPT_CONNECT_ONLY
with the URL and hostname in a call to libcurl first and then once that is
successful, load up a second call to libcurl to perform the email sending -
this is what I do myself in my SMTP code.

Kind Regards

Steve
-------------------------------------------------------------------
List admin: http://cool.haxx.se/list/listinfo/curl-library
Etiquette: http://curl.haxx.se/mail/etiquette.html
Received on 2014-05-18