chore: bypass slow password validation in integration tests

[icewind1991]

From early testing it looks like we spend 50%-60% of the runtime of integration tests doing password validation 🙈

This enables the security_softening, which bypasses this by caching the password validation.

[nickvergessen]

From early testing it looks like we spend 50%-60% of the runtime of integration tests doing password validation

Did you also do:

config:system:set hashing_default_password --value=true --type=bool

locally when testing your improving? Otherwise that would explain the non-existing difference on CI :P

[nickvergessen]

I guess this should be moved to

server/build/integration/run.sh

Lines 26 to 39 in 65135bb

	if [ "$INSTALLED" == "true" ]; then
	# Disable appstore to avoid spamming from CI
	$OCC config:system:set appstoreenabled --value=false --type=boolean
	# Disable bruteforce protection because the integration tests do trigger them
	$OCC config:system:set auth.bruteforce.protection.enabled --value false --type bool
	# Disable rate limit protection because the integration tests do trigger them
	$OCC config:system:set ratelimit.protection.enabled --value false --type bool
	# Allow local remote urls otherwise we can not share
	$OCC config:system:set allow_local_remote_servers --value true --type bool
	# Allow self signed certificates
	$OCC config:system:set sharing.federation.allowSelfSignedCertificates --value true --type bool
	# Allow creating users with dummy passwords
	$OCC app:disable password_policy
	else

so that people running it locally benefit as well

[icewind1991]

hashing_default_password doesn't seem to make much of a difference for me locally.

It seems to have just been a case of ACPu not being enabled before, the performance improvements seem to work now: 8m/15m for sharees_feautures/sharing_features compared to 13m/24m before

The various failures are probably a side effect of the added caching, will look into those