DI-tags documentation should contain 'swiftmailer.default.plugin' tag instead of 'swiftmailer.plugin' to work

[mgrinko]

SwiftMailer plugin works only if its service defined with tag 'swiftmailer.default.plugin' and does not work with tag 'swiftmailer.plugin' specified in documentation

[wouterj]

Hmm, I'm not sure how we should document this. If the tag is swiftmailer.default.plugin, the plugin will be registered on the mailer with name default. If the tag is swiftmailer.other_mailer.plugin it'll be registered on the mailer with name other_mailer.

So, just saying swiftmailer.default.plugin doesn't work. Maybe swiftmailer.mailer_name.plugin in the table and some text saying that mailer_name is the name of the mailer will solve the issue?

[stof]

actually, having different tags for the different mailers was a mistake IMO (and I said it on the PR but it was merged without the change). Thus, it was breaking BC. I'm considering changing it back.

[mgrinko]

I have just found out that current tag does not work and found the solution for my situation.
I don't know common solution. So I am only asking to fix this in documentation.

[makasim]

The doc could have a notice\warning that describes what default means in the tag name.

[wouterj]

@stof yeah, I was also thinking why they didn't create an attribute to determine the mailer. I'm +1 for that (but that will break BC again... :(

@mgrinko you're fix is correct. The only thing is that it only works when your mailer is called 'default'. That will confuse other people when they named their mailer to something else. Can you please add a note, saying that default is the name of the mailer? Something like this:

.. note::

    ``default`` in this tag is the name of the mailer. You should change it to the
    name of your mailer in order to use this tag.

[mgrinko]

@wouterj I added your explanation as a note.

[wouterj]

this tag line should be as long as the headline itself

[mgrinko]

Done

[wouterj]

thank you @mgrinko !

[wouterj]

👍

[weaverryan]

Great work everyone! I agree that the way this feature was implemented was a little silly and the BC break didn't make it back to the docs. But, all good now!

Thanks!