Make last parameter of openssl_seal optional

[phil-davis]

Fixes issue phpstan/phpstan#3896

Which seemed to have been broken by #327

Where are the tests for this? It seems to be untested, so we won't know if there is a regression in this file of function signature definitions.

[ondrejmirtes]

Where are the tests for this? It seems to be untested

How would you test function signatures? It'd just duplicate information that's already in functionMap.php, and it wouldn't actually be the guarantee the signatures are correct :)

[ondrejmirtes]

You can add a test case in CallToFunctionParametersRuleTest, but I definitely don't want that for all the described functions in that file :)

[ondrejmirtes]

Thank you!

[ondrejmirtes]

Alright, I just tested this exact situation and found a couple more violations: d2f443c

[phil-davis]

Where are the tests for this? It seems to be untested

How would you test function signatures? It'd just duplicate information that's already in functionMap.php, and it wouldn't actually be the guarantee the signatures are correct :)

Yes, I agree and realise that - what to do for tests of things that are just "constant input data". The test criteria is that "the function signatures match correctly the signatures defined by the PHP documentation/spec". But the real implementation will probably be a nonsense comparison of the content of functionMap with an "expected map", and the "expected map" will be just as likely to have errors in its definition!

If there is somewhere that has an easily machine-readable-parseable "official specification" of all those functions, then a test could parse all that and check against functionMap - but that's a job for for when someone is stuck for 2 weeks in quarantine...

[ondrejmirtes]

easily machine-readable-parseable "official specification" of all those functions

This is notoriously messy. If it wasn't messy then we wouldn't need to manually maintain this functionMap at all :)