>>> MAIL FROM:<invalid_email_at_somedomain.com>
>> < 250 2.1.0 OK l6sm88100403oeq.0 - gsmtp

> But nobody except you know that it is in fact invalid, and you know that due to some out of band information. The server cannot know it is invalid.

True. But my point is at least after libcurl tried sending the email (using the invalid email address), the server at the other end (on receiving it) should have rejected it, since it is not from a valid email id.
But since the server did not do so, isn't it likely that libcurl did not send the email using the invalid address??

>> Kindly let me know if any additional curl option needs to be included
>> to ensure that the value passed to CURLOPT_MAIL_FROM is in sync with
>> that passed to CURLOPT_USERNAME

> How do you mean in sync? There's nothing that says they need to be even remotely similar or use any similar patterns or names. How would libcurl be able to check or enforce something like that?

Pardon me. But my point here is that, since CURLOPT_MAIL_FROM takes the sender's email id, should it not be a valid email id AND also correspond to the same email address whose username and password are being passed to the options: CURLOPT_USERNAME and CURLOPT_PASSWORD respectively?

Since otherwise, regardless of whatever invalid value is passed to CURLOPT_MAIL_FROM, libcurl would still proceed to send the mail, only as long as the CURLOPT_USERNAME and CURLOPT_PASSWORD have the correct/valid values (which is what is happening currently).
In fact, it does look like libcurl is ignoring the value given to CURLOPT_MAIL_FROM in the presence of CURLOPT_USERNAME and CURLOPT_PASSWORD.

Thanks,
Naveen

-------------------------------------------------------------------
List admin: http://cool.haxx.se/list/listinfo/curl-library
Etiquette: http://curl.haxx.se/mail/etiquette.html
Received on 2014-06-11