Codecov Report

All modified and coverable lines are covered by tests ✅

Comparison is base (6804fbe) 46.70% compared to head (abcd82f) 48.01%.

@@            Coverage Diff             @@
##           master   #20015      +/-   ##
==========================================
+ Coverage   46.70%   48.01%   +1.30%     
==========================================
  Files         445      445              
  Lines       43896    43890       -6     
==========================================
+ Hits        20503    21075     +572     
+ Misses      23393    22815     -578     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

That's a good refactor but it's a cache and one more function call may matter on many-many cache calls. Are you willing to compare performance?

I performed a performance test to check the impact of the new method. I ran the test in a loop with 1 million iterations:

        $start = microtime(true);

        for ($i = 0; $i < 1000000; $i++) {
            $cache->flush();

            $cache->multiSet([
                'string_test' => 'string_test',
                'number_test' => 42,
                'array_test' => ['array_test' => 'array_test'],
            ], 0);
        }

        $end = microtime(true);

        echo "Old method elapsed time: " . ($end - $start) . " seconds\n";

I ran the tests three times:

• ...New method elapsed time: 8.5871288776398 seconds

• ...New method elapsed time: 8.7551159858704 seconds

• ...New method elapsed time: 8.8505420684814 seconds

• ...Old method elapsed time: 8.3703300952911 seconds

• ...Old method elapsed time: 8.3308370113373 seconds

• ...Old method elapsed time: 8.4228310585022 seconds

I think the refactored method is slightly slower, but the difference is minimal and likely won't significantly impact real-world performance.

These tests were done using an array cache. In other cache configs, the time taken for network or disk I/O operations would be much greater than in-memory array cache. Therefore, the overhead of an additional function call is likely negligible compared to the time spent on these I/O operations.

@samdark

Alright. Looks good to me.

@samdark any plan for this PR?