scheduler: task failure not logged when retries are configured

[oliver-sanders]

A small logging bug where task failures can be omitted from the logs when execution retries are configured.

We are currently logging all task state changes, however, when retries are configured, the a failed task does not change state until all retries have been exhausted.

Reproducible Example

[scheduling]
    [[graph]]
        R1 = foo

[runtime]
    [[foo]]
        execution retry delays = PT1S
        script = """
            if [[ $CYLC_TASK_SUBMIT_NUMBER -gt 1 ]]; then
                true
            else
                false
            fi
        """

$ cylc play -N <workflow>

 ▪ ■  Cylc Workflow Engine 8.3.0.dev
 ██   Copyright (C) 2008-2024 NIWA
▝▘    & British Crown (Met Office) & Contributors

...
INFO - [1/foo:waiting(runahead)] => waiting
INFO - [1/foo:waiting] => waiting(queued)
INFO - [1/foo:waiting(queued)] => waiting
INFO - [1/foo:waiting] => preparing
INFO - [1/foo/01:preparing] submitted to localhost:background[91088]
INFO - [1/foo/01:preparing] => submitted
INFO - [1/foo/01:submitted] => running
INFO - [1/foo/01:running] => waiting
WARNING - [1/foo:waiting] retrying in PT1S (after 2024-06-18T09:56:09+01:00)
INFO - xtrigger satisfied: _cylc_retry_1/foo = wall_clock(trigger_time=1718700969.3399053)
INFO - [1/foo:waiting] => waiting(queued)
INFO - [1/foo:waiting(queued)] => waiting
INFO - [1/foo:waiting] => preparing
INFO - [1/foo/02:preparing] submitted to localhost:background[91380]
INFO - [1/foo/02:preparing] => submitted
INFO - [1/foo/02:submitted] => running
INFO - [1/foo/02:running] => succeeded
INFO - Workflow shutting down - AUTOMATIC
INFO - DONE