I just removed LINK_TYPES. For the regex, I'm willing to remove it and use a python library to validade but, since the system runs on xenial, it's hard to find a good library for it. I can't use python3-validators for example. There is urllib that can parse the url but don't validate it unless you use it for a query and check for the returning error. There is also the requests library but you also needs to do a request in order to check the result. About the regex, why too fragile? Maybe there would be cases that it will fails but than we can manually approve, right?
I just removed LINK_TYPES. For the regex, I'm willing to remove it and use a python library to validade but, since the system runs on xenial, it's hard to find a good library for it. I can't use python3-validators for example. There is urllib that can parse the url but don't validate it unless you use it for a query and check for the returning error. There is also the requests library but you also needs to do a request in order to check the result. About the regex, why too fragile? Maybe there would be cases that it will fails but than we can manually approve, right?