How to find the files that are the slowest to analyze with PHPStan?

README.md

How to find the files that are the slowest to analyze with PHPStan?

For us, PHPStan became a bit slower with every release. We have a very large codebase with 10.000+ classes. There seem to be a few known issues related to big arrays.

See: phpstan/phpstan#8353 phpstan/phpstan#8146

To understand which files are problematic we run the following command:

vendor/bin/phpstan analyze --memory-limit=-1 --debug -vvv | tee phpstan.log

Then we run:

php parse.php

parse.php

<?php

declare(strict_types=1);

$log = new SplFileObject("phpstan.log");

$logs = [];
$file = null;
while (! $log->eof()) {
    $line = trim($log->fgets());
    if ($line === '') {
        continue;
    }
    if ($file === null) {
        $file = $line;
        continue;
    }
    preg_match('/took (?<seconds>[\d.]+) s/', $line, $matches);

    $logs[] = [(float) $matches['seconds'], $file];
    $file = null;
}

usort($logs, fn(array $left, array $right) => $right[0] <=> $left[0]);
$logs = array_slice($logs, 0, 100);

echo "Slowest files" . PHP_EOL;
foreach ($logs as $log) {
    echo sprintf("%.2f seconds: %s", $log[0], $log[1]) . PHP_EOL;
}

Thank you very much @ruudk.

I was able to shave off 2 minutes phpstan runtime (8 minutes to 6 minutes) by excluding module/*/config/module.config.php from the analysis in an Laminas MVC project.

@bcremer Happy to hear that!

PHPStan will soon get an option that displays the slow files too.

For those who could use a shell one-liner:

vendor/bin/phpstan analyze --memory-limit=-1 --debug -vvv | sed -e 'N; s#\n# #; s#^\(\S\+\) .* \(\S\+\) s$#\2 \1#; /^ /d' | sort -n

I made a rust version as part of a learning exercise.

(I hope this self-plug is OK)

I got an undefined array key 'seconds' error, fixed by adding this piece of code after the preg_match call :

    if (!array_key_exists('seconds', $matches)) {
        continue;
    }