Make enable_starttls_auto opt-in in ActionMailer

Pushed with default to true.

Pratik,

I just let it false because some applications were breaking.

I wanted to avoid the situation that someone is complaining that something is not working and then we had to reply that he have to set something to false in order to make everything work.

jts,

The patch was already applied into core.

Try changing your configurations to:

config.action_mailer.smtp_settings = {

:domain               => 'domain.com',
:address              => 'localhost',
:port                 => 25,
:user_name            => 'user',
:password             => 'passwd',        
:authentication       => :login,
:enable_starttls_auto => false

}

The problem is that enable_starttls_auto comes true by default.

Tell us if that worked! :)

Hi Jose, thanks for your response, much appreciated.

Unfortunately, I'd already tried :enable_starttls_auto => false in the configuration seetings and no joy, I'm still receiving the same error message.

The production log file looks like this for the POST to the controller responsible for mailing Parameters: {"name"=>"test user", "message"=>"testing", "commit"=>"send message", "authenticity_token"=>"a550eb2ff258d7fef0018lfkdofijd3388961d5", "website"=>"test message", "email"=>"jts@gmail.com"} Sent mail to jts@gmail.com getaddrinfo: Name or service not known

The fact that I'm getting the message "Sent mail to jts@gmail.com" suggests that the call to UserMailer.deliver_contact_form is working correctly, but that something fails in communicating with the smtp service.

I've tried POP auth before SMTP too, no luck.. And my hosting provider support is useless. They've suggested trying localhost as the domain twice now !!!

Is there anything else I might check?

for your info, here's the versions of ruby & rails I have installed on production, dunno if it'll help..

ruby 1.8.6 (2008-08-11 patchlevel 287) [x86_64-linux] Rails 2.2.2 local gems:

actionmailer (2.2.2, 2.1.2) actionpack (2.2.2, 2.1.2) activerecord (2.2.2, 2.1.2) activeresource (2.2.2, 2.1.2) activesupport (2.2.2, 2.1.2) builder (2.1.2) cgi_multipart_eof_fix (2.5.0) daemons (1.0.10) fastthread (1.0.1) fcgi (0.8.7) gem_plugin (0.2.3) hpricot (0.6.164) login_generator (1.2.2) mislav-will_paginate (2.3.6, 2.2.3) mongrel (1.1.5) mongrel_cluster (1.0.5) mysql (2.7) radiant (0.6.9) rails (2.2.2, 2.1.2) rake (0.8.3) RedCloth (4.1.1) rmagick (1.15.17) sqlite3-ruby (1.2.4) tzinfo (0.3.12) will_paginate (2.2.2)

thanks for any advice you might be able to think of! L

Correct me if I am wrong, but it appears that the patch was applied for 2.3, so this option is not available in 2.2.2.

I got this working by monkey patching 2.2.2.

I added this to environment.rb:

if RAILS_GEM_VERSION == '2.2.2'
  module ActionMailer
    class Base

      def perform_delivery_smtp(mail)
        destinations = mail.destinations
        mail.ready_to_send
        sender = mail['return-path'] || mail.from

        smtp = Net::SMTP.new(smtp_settings[:address], smtp_settings[:port])
        smtp.enable_starttls_auto if smtp_settings[:enable_starttls_auto] && smtp.respond_to?(:enable_starttls_auto)
        smtp.start(smtp_settings[:domain], smtp_settings[:user_name], smtp_settings[:password],
              smtp_settings[:authentication]) do |smtp|
                smtp.sendmail(mail.encoded, sender, destinations)
              end
       end
    end
  end
else
  raise "you forgot to remove the ActionMailer monkey patch!"
end

I'm a bit apprehensive about upgrading to 2.2.3 on the production server - I dimly remember how much of a pain it was to upgrade to 2.2.2 (on bluehost), so I'd like to avoid that if possible.

I put the "monkey patch" at the end of my environment.rb and the application died, so I quickly removed it again!

Should I put monkey patch into or outside of the

Rails::Initializer.run do |config| block?

thanks for your input ken!

2.2.2 is the latest release, 2.3 (where the fix was committed) hasn't been released yet.

I put the patch at the end of environment.rb.

Thanks Ken, I've managed to "monkey patch" my environment.rb.

Alas, after this failed to fix the issue, I did a telnet session with my hosts SMTP service and it's failing to transmit a message with an SMTP status code 550.

So I've mailed support about this and I hope they'll (finally) be able to solve the issue...

thanks for all your help everyone! Will let you know if/when the issue is resolved!!

@jts

sorry, now reading through the description of your issue again, I'm not sure that this particular patch addresses your problem. Good luck in getting your mail problems solved!

I've found something quite strange.

I've got in my user_mailer.rb a function to fire off a mail to the user via the contact form...

def contact_form(name, email, website, message)

            recipients      "app@domain.com"
            from            "app@domain.com"
            subject         "your contact form"
            sent_on         Time.now
            content_type    "text/html"
            body            :name => name, :email => email, :website => website, :message => message
            template        "contact_form"
    end

If I specify the body variables as a hash, like the documentation states, my application fails (error code 500).

If I specify the body variables as I have above, the program works, but I'm getting this getaddrinfo: Name or service not known error.

The documentation on ActionMailer that I'm referring to is here: http://guides.rubyonrails.org/ac...

Could I be doing something wrong that's causing the program to return a 500 status code?

thanx! j

re: Ken Liu's monkey patch to get it working in 2.2.2

I'd like to verify Ken's monkey patch mentioned above. It fixed this issue on my shared host that was showing this problem.

I'll also confirm Ken's monkey patch as working; fixed this somewhat insidious problem when setting up email, using Redmine 0.8.x on Dreamhost.

The main indicator of the problem is the error "hostname was not match with the server certificate" - without trawling through the ActionMailer source, I'm guessing that there's a problem in AM when the SSL cert name doesn't match the SMTP server name [relatively common for wildcard SSL certs in shared hosting environments]. Potentially should be an option for ignoring certificate mismatch errors, but I guess that's open to abuse, and I'm not about to make a request for it since I'm not prepared to cobble a patch together myself.

Still, hopefully helps someone else stumbling through google results.