[Gambas-user] SMTP Errors

[Rolf-Werner Eilert]

Hi Marty,


> error (authentication failed). I have tried everything including Rolf's
> suggestion of leaving off the user and password. Still authentication

That was Willy's suggestion, and it's only possible when you're directly 
connected to the provider. When you want to use another provider, they 
will insist on you having a name/password for authentication to avoid 
spammers.

> fails. Looking at the documentation for password and user it says that
> only plain authentication is supported. I think the send process breaks
> down when the password and user get to the smtp host. It may be looking
> for an SSL encoded string not a plain string; therefore authentication
> fails. Rolf and Willy let me know what you think. Is this possible?
> Could openSSL be at fault? That's the other thought I had.

Actually, that brings back to my mind an idea I had last week. If we 
could produce a string that is what name/password would be in SSL and 
store it and send it instead, such as

sUser = File.Load(Path &/ "encryptedString.user")
sPass = File.Load(Path &/ "encryptedString.pass")

'and with SMTP client:

	.User = sUser
	.Password = sPass


could that work? I don't know what the SSL function does with these 
values, and if it runs anyway. If I was to make a guess, the answer 
would be NO just because the SSL part should mingle these two strings 
and build ONE single string out of them, so pre-building them would 
produce mere garbage.

On my machines (both at home and in the office), SSL returns no messages 
at all, only TLS returns error messages. So it might be, SSL isn't 
started properly or even isn't started at all. Is there a way to find 
out? One of my providers (my main provider) uses STARTTLS (which in my 
understanding is a special variation of TLS), and there is no hint in 
the docu whether Gambas supports this or not.

BUT Gambas is open source :-) So why not look into the code and judge 
ourselves?

>
> Rolf I've tried the routine you suggested and it seems to work. I may
> follow your advice and use it.

Don't know which one you mean, but nice to hear :-)