If your segemnts are broken or Mailtrain complains about missing 20190726150000_shorten_field_column_names.js, run the following in `server/setup/knex/fixes`:
```NODE_ENV=production node fix-20190726150000_shorten_field_column_names.js```
Fix - message sent to a list not associated with a campaign couldn't be shown in archive - to know which message to show even if the list is not at the campaign, we store test messages in table test_messages
- More options how to send test emails.
- Fixed problems with pausing a campaign (#593)
- Started rework of transactional sender of templates (#606), however this contains functionality regression at the moment because it does not interpret templates as HBS. It needs HBS option for templates as described in https://github.com/Mailtrain-org/mailtrain/issues/611#issuecomment-502345227
TODO:
- detect sending errors connected to not able to contact the mailer and pause/retry campaing and queued sending - don't mark the recipients as BOUNCED
- add FAILED campaign state and fall into it if sending to campaign consistently fails (i.e. the error with sending is not temporary)
- if the same happends for queued email, delete the message
The namespace filter in campaigns was dropped (i.e. "Work with campaign's namespace"). Instead, we need a universal solution. For instance a namespace slector somewhere in the top-right corner, which should apply to everything (not just campaigns).
Nevertheless, I kept the ...-by-namespace rest endpoints and related functions in models because they will be useful for implementing the universal namespace selection feature.
It is now possible to have hierarchical merge tags (separated by "."). The merge tags are now case sensitive.
Mailtrain allows passing element "mt:entries-json" in the RSS item feed. If this is detected, it parses the json structure and makes it available through RSS_ENTRY_CUSTOM_TAGS. Then it can be used as [RSS_ENTRY_CUSTOM_TAGS.field_quote_text.rendered]
