[TYPO3-core] RFC #16610: Provide a SwiftMailer transport that only saves send mails to files

[Jeff Segars]

On 11/29/10 4:04 PM, Ernesto Baschny [cron IT] wrote:
> Hi,
>
> This is a SVN patch request.
>
> Type: Feature
>
> BT reference: http://bugs.typo3.org/view.php?id=16610
>
> Branches: trunk
>
> Problem:
> We introduced the SwiftMailer interface in t3lib_mail in 4.5 alpha2. So
> we can now generate RFC conforming mails and have a powerful mail engine
> available which also supports different transport mechanisms (like
> mail(), sendmail and direct SMTP).
>
> To ease debugging mail sending in a TYPO3 installation, it might be
> useful for a complete installation not really *send* mails, but just
> save them to a file.
>
> This RFC provides this feature, allowing the admin to configure the new
> transport "mbox", which saves all mails to a mbox-conformant file
> (simple text file, see RFC 4155). This file can be viewed with a regular
> text editor or a mail client which can open a "mbox" file.
>
> Together with Jigal's RFC #16591 which hooks into all other mail() usage
> we are able to log every outgoing mail to a file!
>
>
> How to test:
> Set transport to "mbox" in the install tool and configure the desired
> output file. Install Jigal's patch (manually fix the conflict in the
> core_autoloader file..) or use the attached patch which adds t3lib_mail
> support to the scheduler test task and just let TYPO3 send some mails!
>
>
> Additional info:
> This feature is a follow-up the the new mailing API and will be included
> according to the rules for "new features" in this phase (see
> http://forge.typo3.org/projects/typo3v45-projects/wiki/Feature_Freeze,
> where "t3lib_mail" is listed as potential candidates for follow-ups).
>
>
> Cheers,
> Ernesto
>


Hey Ernesto,
Nice feature! +1 on reading and testing with 2 small comments

* Currently transport_mbox_file must be an absolute path. It would be 
nice to either support a relative path or mention that its absolute 
since it varies throughout the Install Tool.

* In getReversePath() there's a little CGL cleanup needed on the elseif 
conditions.

Otherwise, all looks good.  Thanks!

Jeff