Domain Access and newletters

That is a very interesting question, and there may not be a good answer.

It depends entirely on how the newsletter module is storing its data and on how you wish to segregate that data for delivery.

Based on the very long discussion we had here: http://drupal.org/node/197488, there may be a need to write an additional module to handle these rules.

In the short term, understand this: In Domain Access HEAD (and possibly in rc3) there is an option to disable Domain Access rules during cron runs.

Looking at the SimpleNews code, here is the general issue:

When cron runs, it will run from whichever domain is specified. That is, if you run from test1.example.com, then Domain Access will only allow content from test1. SimpleNews respects this rule.

When cron runs, SimpleNews sets the send status to FALSE, indicating that the content has already been sent. So even if you were to run cron from multiple domains (which is possible), you would not send the newsletter again.

One option -- which would need thorough testing -- would be to use Domain Prefix to create new SimpleNews tables for each domain, but that might cause all kinds of end-user confusion.

Now, understand that Domain Access implementation of hook_node_grants() can be modified by reference using hook_domaingrants(), so we have the option of writing a module that changes the default behavior IFF:

-- Cron is running AND SimpleNews is active.

You could, in theory, also create domain-specific newsletters, since SimpleNews itself creates its newsletters as nodes. If the nodes in your newsletter for test1 all came from test1, then disabling Domain Access on cron would work as expected.

Does that make any sense?

I think the third option would work. If you use SimpleNews, and the "newsletter" nodes are assigned to specific domains, then you add content to each newsletter that either a) is viewable from that domain or b) is viewable to all domains. Then you disable Domain Access on cron.

It may also work if we force the newsletter to write absolute URLs to the correct domain -- using the "strict SEO" trick.

I'd be willing to experiment after we get a stable release.

By end of month, if not sooner. We have caught the major bugs in rc3 and are just doing some cleanup now.