Improve cursor performance

[hakastein]

1. Updated composer deps
2. mb_substr replaced to mb_str_split
3. Node version updated to actual
4. phpunit.xml.dist updated to actual

Parser bench before:

+------+-------------+-------------------------+-----+------+------------+---------------+--------------+----------------+
| iter | benchmark   | subject                 | set | revs | mem_peak   | time_avg      | comp_z_value | comp_deviation |
+------+-------------+-------------------------+-----+------+------------+---------------+--------------+----------------+
| 0    | ParserBench | benchFirefoxPreferences |     | 10   | 4,622,672b | 368,879.500μs | -0.65σ       | -0.88%         |
| 1    | ParserBench | benchFirefoxPreferences |     | 10   | 4,622,672b | 366,646.000μs | -1.09σ       | -1.48%         |
| 2    | ParserBench | benchFirefoxPreferences |     | 10   | 4,622,672b | 370,244.100μs | -0.38σ       | -0.51%         |
| 3    | ParserBench | benchFirefoxPreferences |     | 10   | 4,622,672b | 373,931.200μs | +0.35σ       | +0.48%         |
| 4    | ParserBench | benchFirefoxPreferences |     | 10   | 4,622,672b | 381,093.800μs | +1.77σ       | +2.40%         |
+------+-------------+-------------------------+-----+------+------------+---------------+--------------+----------------+

After:

+------+-------------+-------------------------+-----+------+------------+--------------+--------------+----------------+
| iter | benchmark   | subject                 | set | revs | mem_peak   | time_avg     | comp_z_value | comp_deviation |
+------+-------------+-------------------------+-----+------+------------+--------------+--------------+----------------+
| 0    | ParserBench | benchFirefoxPreferences |     | 10   | 8,330,040b | 23,806.100μs | -0.93σ       | -1.03%         |
| 1    | ParserBench | benchFirefoxPreferences |     | 10   | 8,330,040b | 23,670.400μs | -1.43σ       | -1.60%         |
| 2    | ParserBench | benchFirefoxPreferences |     | 10   | 8,330,040b | 24,274.300μs | +0.82σ       | +0.91%         |
| 3    | ParserBench | benchFirefoxPreferences |     | 10   | 8,330,040b | 24,171.900μs | +0.44σ       | +0.49%         |
| 4    | ParserBench | benchFirefoxPreferences |     | 10   | 8,330,040b | 24,349.000μs | +1.10σ       | +1.22%         |
+------+-------------+-------------------------+-----+------+------------+--------------+--------------+----------------+

Its looks like we got memory usage x2, but adwantage in speed is more than 15 times. And on bigger files it becomes more significant. 204KB file in example:

Old:

+------+-------------+-------------------------+-----+------+------------+-----------------+--------------+----------------+
| iter | benchmark   | subject                 | set | revs | mem_peak   | time_avg        | comp_z_value | comp_deviation |
+------+-------------+-------------------------+-----+------+------------+-----------------+--------------+----------------+
| 0    | ParserBench | benchFirefoxPreferences |     | 10   | 7,377,128b | 4,960,811.400μs | -1.53σ       | -1.07%         |
| 1    | ParserBench | benchFirefoxPreferences |     | 10   | 7,377,128b | 4,997,625.000μs | -0.48σ       | -0.34%         |
| 2    | ParserBench | benchFirefoxPreferences |     | 10   | 7,377,128b | 5,062,494.200μs | +1.37σ       | +0.96%         |
| 3    | ParserBench | benchFirefoxPreferences |     | 10   | 7,377,128b | 5,011,520.200μs | -0.09σ       | -0.06%         |
| 4    | ParserBench | benchFirefoxPreferences |     | 10   | 7,377,128b | 5,040,271.900μs | +0.73σ       | +0.51%         |
+------+-------------+-------------------------+-----+------+------------+-----------------+--------------+----------------+

New:

+------+-------------+-------------------------+-----+------+-------------+--------------+--------------+----------------+
| iter | benchmark   | subject                 | set | revs | mem_peak    | time_avg     | comp_z_value | comp_deviation |
+------+-------------+-------------------------+-----+------+-------------+--------------+--------------+----------------+
| 0    | ParserBench | benchFirefoxPreferences |     | 10   | 20,932,768b | 84,057.700μs | +0.94σ       | +1.66%         |
| 1    | ParserBench | benchFirefoxPreferences |     | 10   | 20,932,768b | 84,720.500μs | +1.39σ       | +2.46%         |
| 2    | ParserBench | benchFirefoxPreferences |     | 10   | 20,932,768b | 82,237.600μs | -0.31σ       | -0.55%         |
| 3    | ParserBench | benchFirefoxPreferences |     | 10   | 20,932,768b | 80,988.300μs | -1.16σ       | -2.06%         |
| 4    | ParserBench | benchFirefoxPreferences |     | 10   | 20,932,768b | 81,439.900μs | -0.86σ       | -1.51%         |
+------+-------------+-------------------------+-----+------+-------------+--------------+--------------+----------------+

Execution time reduced by more than 60 times. And in our project its not biggest file.

All tests still passed

[jrmajor]

Wow, this is a significant improvement! Thank you, @hakastein.

[hakastein]

Thx. There will be more. Ill preparing units support