Two Postgres unicode order_by tests fail locally on some machines

GregoryTravis commented 7 months ago

These tests fail consistently for me. @radeusgd suggests checking the collation setting in the database to skip the test.


[PostgreSQL] Table.order_by:  [17/19, 246ms]
(...)
    - [FAILED] should behave as expected with Unicode normalization, depending on the defaults settings [16ms]
        Reason: [Nothing, 'śb', 'śa', 'śc'] did not equal [Nothing, 'śa', 'śb', 'śc']; first difference at index 1  (at /Users/gmt/dev/enso/enso/test/Table_Tests/src/Common_Table_Operations/Order_By_Spec.ens>
    - [FAILED] should support case insensitive ordering [5ms]
        Reason: ['Aleph', 'alpha', 'Beta', 'bądź'] did not equal ['Aleph', 'alpha', 'bądź', 'Beta']; first difference at index 2  (at /Users/gmt/dev/enso/enso/test/Table_Tests/src/Common_Table_Operations/Ord>
    - [PENDING] should support natural and case insensitive ordering at the same time
    - ```

radeusgd commented 7 months ago

Yes, we could inspect the result of the query SELECT datcollate FROM pg_database WHERE datname = current_database();. The tests should be marked as pending if the collation is not en_GB.UTF8 (I assume other values are OK too, we may only ignore it if it is C.UTF-8).

enso-bot[bot] commented 7 months ago

Greg Travis reports a new STANDUP for today (2024-01-26):

Progress: Reviews for Nothing PRs; review for collation fix PR; determining what collation setting is used on the CI machines; trying to find logging in the CI logs It should be finished by 2024-01-29.

Next Day: Collation

enso-org / enso

Two Postgres unicode order_by tests fail locally on some machines #8827