chore: temporarily disable cascade relationship combinations

[nikophil]

in order to make the CI pass again since sebastianbergmann/phpunit#6213, let's temporarily disable this behavior, until we find a solution.

[kbond]

Benchmark?

[nikophil]

@kbond

Benchmark?

you can see here the benchmark result:

+--------------------------------+-------------------+-----+------+-----+-----------------+------------------+-----------------+
| benchmark                      | subject           | set | revs | its | mem_peak        | mode             | rstdev          |
+--------------------------------+-------------------+-----+------+-----+-----------------+------------------+-----------------+
| CategoryFactoryBench           | bench_create      |     | 10   | 5   | 17.969mb +0.01% | 5.084ms +0.25%   | ±4.35% -11.61%  |
| CategoryFactoryBench           | bench_create_many | 1   | 10   | 5   | 17.976mb +0.00% | 5.992ms +18.43%  | ±4.42% +4.06%   |
| CategoryFactoryBench           | bench_create_many | 10  | 10   | 5   | 21.416mb +0.00% | 38.614ms +0.14%  | ±0.83% -17.15%  |
| CategoryFactoryBench           | bench_create_many | 50  | 10   | 5   | 36.218mb +0.00% | 179.635ms -0.23% | ±0.84% +5.95%   |
| ContactFactoryBench            | bench_create      |     | 10   | 5   | 17.752mb +0.00% | 2.117ms -5.23%   | ±5.88% +17.51%  |
| ContactFactoryBench            | bench_create_many | 1   | 10   | 5   | 17.756mb +0.00% | 2.467ms +8.96%   | ±3.42% -3.52%   |
| ContactFactoryBench            | bench_create_many | 10  | 10   | 5   | 18.638mb 0.00%  | 10.672ms -2.27%  | ±3.69% +96.23%  |
| ContactFactoryBench            | bench_create_many | 50  | 10   | 5   | 22.791mb 0.00%  | 46.658ms -0.70%  | ±2.30% +188.71% |
| PersistentDocumentFactoryBench | bench_create      |     | 10   | 5   | 16.265mb 0.00%  | 781.903μs +0.32% | ±2.66% +14.50%  |
| PersistentDocumentFactoryBench | bench_create_many | 1   | 10   | 5   | 16.269mb 0.00%  | 831.145μs -1.88% | ±4.11% +34.89%  |
| PersistentDocumentFactoryBench | bench_create_many | 10  | 10   | 5   | 16.405mb 0.00%  | 4.811ms +0.56%   | ±1.89% -6.21%   |
| PersistentDocumentFactoryBench | bench_create_many | 50  | 10   | 5   | 18.001mb 0.00%  | 23.281ms +1.05%  | ±1.48% +73.92%  |
+--------------------------------+-------------------+-----+------+-----+-----------------+------------------+-----------------+

almost all bench are equals to the previous version. The only one which is above the 10% threshold is:

+--------------------------------+-------------------+-----+------+-----+-----------------+------------------+-----------------+
| benchmark                      | subject           | set | revs | its | mem_peak        | mode             | rstdev          |
+--------------------------------+-------------------+-----+------+-----+-----------------+------------------+-----------------+
| CategoryFactoryBench           | bench_create_many | 1   | 10   | 5   | 17.976mb +0.00% | 5.992ms +18.43%  | ±4.42% +4.06%   |

this one mesures the time taken by CategoryFactory::createMany(1) when it is run 10 times (10 revs). It performs 5 iterations (its) and I think it takes the average of those iterations.

This average is 5.992ms and is +18% above than the bench on 2.x. Both ran on the exact same code, so the difference comes from some "versatility" of the code (maybe because of the db I/O I think). I could reduce this "versatility" (not sure this is the good term 😅 ). Maybe this benchmark needs some fine-tuning, but I couldn't do it without a significant increase in the duration of the benchmark.

As a comparison here is a benchmark result with a "real" performance problem (when enabling the global event system)

| benchmark                      | subject           | set | revs | its | mem_peak         | mode              | rstdev         |
+--------------------------------+-------------------+-----+------+-----+------------------+-------------------+----------------+
| CategoryFactoryBench           | bench_create      |     | 10   | 5   | 17.941mb +3.36%  | 6.961ms +42.50%   | ±3.15% -46.07% |
| CategoryFactoryBench           | bench_create_many | 1   | 10   | 5   | 17.953mb +3.35%  | 7.537ms +36.44%   | ±4.48% -16.12% |
| CategoryFactoryBench           | bench_create_many | 10  | 10   | 5   | 21.384mb +24.87% | 52.092ms +35.94%  | ±2.06% +80.35% |
| CategoryFactoryBench           | bench_create_many | 50  | 10   | 5   | 36.153mb +73.85% | 243.340ms +35.30% | ±1.75% +96.80% |
| ContactFactoryBench            | bench_create      |     | 10   | 5   | 17.723mb +0.85%  | 3.178ms -4.61%    | ±2.42% -15.81% |
| ContactFactoryBench            | bench_create_many | 1   | 10   | 5   | 17.727mb +0.91%  | 3.185ms +13.37%   | ±6.07% -12.84% |
| ContactFactoryBench            | bench_create_many | 10  | 10   | 5   | 18.606mb +7.91%  | 14.116ms +20.28%  | ±4.41% +96.76% |
| ContactFactoryBench            | bench_create_many | 50  | 10   | 5   | 22.726mb +31.28% | 60.565ms +23.66%  | ±0.58% -36.64% |
| PersistentDocumentFactoryBench | bench_create      |     | 10   | 5   | 16.236mb +0.47%  | 836.236μs +10.56% | ±0.97% -17.80% |
| PersistentDocumentFactoryBench | bench_create_many | 1   | 10   | 5   | 16.241mb +0.47%  | 895.347μs +15.70% | ±0.71% -8.13%  |
| PersistentDocumentFactoryBench | bench_create_many | 10  | 10   | 5   | 16.306mb +3.79%  | 5.586ms +13.52%   | ±1.29% -2.29%  |
| PersistentDocumentFactoryBench | bench_create_many | 50  | 10   | 5   | 17.939mb +13.91% | 26.747ms +15.04%  | ±0.72% -31.87% |
+--------------------------------+-------------------+-----+------+-----+------------------+-------------------+----------------+

you can see almost all benchs result exploded.