Add missing one-click login to OMP 3.4

[ajnyga]

In OMP 3.4 the one click reviewer access seems to be broken. The email contains a link that looks ok, there is also corresponding line in access_keys. However, the link leads to the login page. In OJS 3.4 the one-click links works.

Seems that it is not handled at all in OMP 3.4
https://github.com/pkp/omp/blob/stable-3_4_0/pages/reviewer/ReviewerHandler.php

OJS 3.4 has the required handling
https://github.com/pkp/ojs/blob/stable-3_4_0/pages/reviewer/ReviewerHandler.php

This pr introduces the similar logic to OMP 3.4.

[asmecher]

@defstat, could you have a look at this one, both the PR for stable-3_4_0 and to check whether this needs attention in stable-3_5_0 and main, where it'll be relying instead on the invitation toolset?

[ajnyga]

Getting failed tests but with errors that seem unrelated.

[ajnyga]

Any information about this? Thinking mainly whether the changes in overall seem good since I would like to have this in our production site.

[defstat]

@defstat, could you have a look at this one, both the PR for stable-3_4_0 and to check whether this needs attention in stable-3_5_0 and main, where it'll be relying instead on the invitation toolset?

@asmecher @ajnyga sorry for the delay on this.

This does not need porting to stable-3_5_0 and main branches, in my opinion - the one-click feature is handled by invitations after 3.5.0, and the respective code is already in the public function authorize($request, &$args, $roleAssignments) function of the ReviewerHandler class.

[ajnyga]

Thanks! Agreed that no porting needed, it was rewritten and also OMP had the same logic in 3.5+.
@asmecher I restarted the tests, not sure what caused them to fail the last time since I do not think one-click is used in the tests.

[asmecher]

Sorry for the wait, @ajnyga!

I did catch an issue that's also present in the OJS code. Would you mind changing it here, and opening a PR in OJS as well? (No good deed goes unpunished.)

[asmecher]

The $reviewSubmission->getId() != $reviewAssignment->getSubmissionId() will always be false here, since we're getting the submission from the review assignment.

However, there's no check to make sure that the review assignment and the review are from the current journal!