My thought here was not thoroughly considered. I was thinking that something like LOCAL isn't possible in a regex, which seemed like a pity. But you are right that I want finer control than that anyway. Perhaps a solution wherein one can define multiple custom-named lists of trusted subnets (eg. Androids, Servers, Workstations) and then refer to those names somehow in the ABE rulesets, including in regular expressions. ABE looks a bit like a firewall and some firewalls allow for this sort of thing. That said, I can achieve what I need without that in this case, so consider it a low priority wishlist itemThrawn wrote:I'm not sure what you mean by this, apart from the existing 'LOCAL' keyword? If you really want to apply this to multiple LANs, and if you're very confident that no-one on those LANs has access to sensitive sites by virtue of their network location (router admin pages, local servers, etc), then you could just modify the built-in rule. But you've already indicated that you don't want non-Android systems to be included.@Giorgio: It would be much appreciated if there were a way to specify a regular expression <resource> utilizing a token of some kind to represent a generic local network. Specifying the subnet explicitly as I have done means that a simple rule will not work for multiple offices.
Similar to what I'm thinking, but I'd like to see the user given the option of multiple lists to use as she sees fit. I may trust example.com enough to let them run scripts in my browser, but not enough to let them do AirDroid-like activities.You know, this discussion has reminded me of a thought that I've had before, about the idea of having a keyword available in ABE rules, eg 'TRUSTED', to refer to the regular site whitelist, so you can have rules that take into account whether you trust a site. Then you could have a rule such as:Code: Select all
Site LOCAL Accept from LOCAL Accept GET HEAD OPTIONS from TRUSTED Deny
