Intra-RC regression: User dashboard: show most recent contributions, not earliest

[agh1]

This is the same as #13901, just for the 5.12 RC.

Overview

A recent change in how contributions are listed in the User Dashboard has caused a regression where it displays the 12 contributions with the lowest IDs (generally the earliest) rather than the most recent ones.

Before

A test user makes 15 contributions, one on each day in March. The first 12 appear.

After

Same scenario: 15 contributions, one on each day in March. The most recent 12 appear.

Technical Details

The listing was switched to use the API in #13584. There isn't a good way in APIv3 to list the most recent things starting with the oldest. In this case, I had to sort by most recent and then reverse the array.

Comments

I might wonder whether the header should say "Your Most Recent Contribution(s)", but that's a bit outside scope, and the user dash appears to have always just been the dozen most recent contributions.

[civibot]

(Standard links)

• If this is your first pull-request for CiviCRM, please browse CONTRIBUTING.md for information about the development and testing processes.
• If you are reviewing this pull-request, you may wish to consult the test sites and the Review Standards (long template, short template).

[eileenmcnaughton]

thanks @agh1 looks good to me. It took me a while to understand your sort comments - but I got there.

[agh1]

I just updated the test that was failing on #13901: the CRM_Contact_Page_View_UserDashBoardTest::testDashboardContentContributionsWithInvoicingEnabled test (whether intentionally or not) was checking that the first contribution created appeared first in the listing when all three contributions were set with the same date. I updated it so that the contributions have different dates, and the first one created is second chronologically, so the test now confirms that the second one listed is ID 1.

[eileenmcnaughton]

Those tests were written prior to switching over to ensure the switch didn't cause regressions so it will have been testing 'what was happening at the time it was written' - obviously having the dates the same was a pretty limited data set & your change sounds like an improvement