Remove memoization

[joemcgill]

An alternative approach to #5642 is to remove the memoization completely, which reverts these functions to their 6.3 behavior.

Trac ticket: https://core.trac.wordpress.org/ticket/59847

---

This Pull Request is for code review only. Please keep all other discussion in the Trac ticket. Do not merge this Pull Request. See GitHub Pull Requests for Code Review in the Core Handbook for more details.

[felixarntz]

@joemcgill Code-wise this looks good. Could you do a profiling / benchmarking check on whether/how much this slows things down? If it's negligible, I think we should go with this change as it would be the safer one to make over #5642. Otherwise, we can try exploring whether #5642 is enough.

[joemcgill]

@felixarntz I can do some benchmarks, but to be clear, this just restores these functions to their previous behavior, so it should have no impact when compared with 6.3 and previous versions.

[felixarntz]

@joemcgill Yes but part of why the memorizing was introduced was because these functions are now being called more often (rather than checking a constant). So there's a chance this change could make performance worse than pre-WP6.4. I would assume the impact is negligible but we should verify.

[joemcgill]

@felixarntz the performance test summary for this PR shows very little change between this and trunk. However, given that those results can fluctuate, I ran local benchmarks of this branch vs trunk and got the following server timing numbers:

metric	trunk	PR	diff	diff%
wp-total (p10)	311.71	313.89	2.18	0.70%
wp-total (p25)	313.7	315.34	1.64	0.52%
wp-total (p50)	317	317.86	0.86	0.27%
wp-total (p75)	319.7	320.75	1.05	0.33%
wp-total (p90)	326.99	324.57	-2.42	-0.74%

Seems like a safe change to me.

[mukeshpanchal27]

LGTM! Left some nit-pick

[joemcgill]

Added some unit tests to this PR that demonstrate the problems in trunk but now pass with this PR. The second set are tricky to show in unit tests because you have to switch to set the application state to where a theme is active from an alternate theme root directory that is not yet registered when the functions are called the first time, so there's a bit of filter mocking going on in the tests. Would be happy for feedback on ways of improving those tests, if anyone has ideas.

[felixarntz]

Thanks @joemcgill, tests look solid!

[joemcgill]

Merged in https://core.trac.wordpress.org/changeset/57129.