Database upgrade problem for 2.4.8-2 to 3.1.0-1

Software Support
7

After making the above changes to the review_assignments table and then upgrading, there are still problems. Before describing them, I should mention that when I tried the 3.1.0-1 code straight out of the package the upgrade gave "ERROR: type ‘serial’ does not exist”. I applied the patch given by Alec Smecher in this thread – only the first of the two chunks worked but that message went away.

My software is Debian 9.3, Apache 2.4.25, PHP 5.6.34 and Postgres 9.6.7. I’m using the default theme at the moment.

Now the problems I noted, some of which may be features. I log in as a Journal Editor and click Submissions. I see four links “My Queue”, “Unassigned”, “All Active” and “Archives”.

  1. An article was submitted in 2013 under 2.3.6. The editor requested three reviews but rejected the paper before they came (or maybe they came to him directly by email). In 3.1.0-1 it appears under “Archives” with “Declined” in orange, but “A review is overdue” is written below it. The filter “Overdue” accepts it, but the filter “Incomplete” rejects it. The filter “Review” accepts it, but the other stage filters reject it. This information is contradictory: since the submission has been declined, it is not overdue and not at the review stage. At least one reviewer of this long-rejected submission got email from the Acron plugin complaining that his review was 4 years overdue.

  2. Lots of submissions in the “Archives” list have “Waiting for reviewers to be assigned” written below them even if they have been declined. But editors often reject submissions without assigning reviewers, and that shouldn’t raise an alarm.

  3. Selecting the Archives list and then the filter Overdue, a lot of Published submissions appear in addition to Declined submissions. How can Published submissions be overdue?

  4. In the Review panel of most rejected or published submissions (I tried about 20) there are active Accept Submission and Reject Submission buttons. Is this intentional? It seems dangerous to provide such buttons for already-finalised submissions.

  5. The All Active list begins with lots of partial submissions that are not brought up by the search engine in 2.4.8 at all. (Nor does the ArticleReport plugin list them.) Moreover, clicking on those search results takes one to some stage in the Submit an Article workflow, as if the editor is the author. I even found one which took me to Step 5, “Submission Complete” yet it wasn’t completed according to its 2.4.8 record. Is that intentional?

Thanks. Brendan.