Closed dokmic closed 2 days ago
Since checking empty old values in the AlwaysPresentNotUserProvided helper makes no sense in the first place, we can safely remove this condition. The removal didn't cause any regressions.
@dokmic does it make sense to remove this check? It allows us to make a fast decision at the first step when checking a large number of new values.
does it make sense to remove this check? It allows us to make a fast decision at the first step when checking a large number of new values.
@vaerh I don't think keeping it in the first place is correct. The property should be skipped if it's not present in the user configuration, regardless of the stored value. From what I see in the git history, this check was only for such optimization, but it may cause some side effects if we use this helper more.
For RouterOS, it's common to return empty strings for unset properties. This means that empty strings will end up in the state and, hence, in the old
value. The aaa.password-format
property is just our first case.
Yeah, I agree.
:tada: This PR is included in version 1.56.0 :tada:
The release is available on GitHub release
Your semantic-release bot :package::rocket:
This PR is mostly about adding newly added properties in the 7.15 release:
sfp_ignore_rx_los
torouteros_interface_ethernet
;enable_ipv6_accounting
torouteros_ppp_aaa
;require_message_auth
torouteros_radius
;require_message_auth
torouteros_user_manager_settings
;reselect_interval
torouteros_wifi_channel
;current_caps_man_address
torouteros_wifi_cap
;current_caps_man_identity
torouteros_wifi_cap
.On top of that, RouterOS 7.15 started always returning an empty value in
routeros_wifi_configuration
(/rest/interface/wifi/configuration
):Since checking empty old values in the
AlwaysPresentNotUserProvided
helper makes no sense in the first place, we can safely remove this condition. The removal didn't cause any regressions.