feat: use regex for ignoring hosts

[silviumarcu]

Allow to use regexes for ignore_hosts to easily ignore hosts that are frequently changed e.g. pipelinesghubeus*.actions.githubusercontent.com

[hartwork]

I would like to note that this change would break API backwards-compatibility (e.g. because . no longer matches just literal dots) and hence would need a major version bump if merged.

[silviumarcu]

@hartwork do you have any suggestion on how to modify it to make it API backwards compatible?

[jairhenrique]

The signature of ignore_hosts could be like this:

ignore_hosts: set[str | re.Pattern]

In the _build_ignore_hosts you can check the type of each item of set and do the validation for strings and regex.

[silviumarcu]

Correct me if I am wrong, but that will work only if the regex pattern is compiled, re.compile("regex"). If it's passed as a normal string r"regex" then it's just a plain string when checking the type so it will use the non-regex implementation.

[vEpiphyte]

Even if the function signature is changed, that still has an interaction with the ignore_localhost option which adds multiple items to the (potentially empty) hosts_to_ignore set. If the handling of that that option was changed to just add its own filter function, then the function _build_ignore_hosts could change to the following:

    @staticmethod
    def _build_ignore_hosts(hosts_to_ignore):

        if isinstance(re.Pattern):
            def filter_ignored_hosts(request):
                # DISCUSS: Would need to decide if search() or match() is best here.
                # search() tends to be more user friendly behavior
                # match() is more specific though.
                if hasattr(request, "host") and hosts_to_ignore.search(request.host):
                    return
                return request

        else:
            hosts_to_ignore = set(hosts_to_ignore)
            def filter_ignored_hosts(request):
                if hasattr(request, "host") and request.host in hosts_to_ignore:
                    return
                return request

        return filter_ignored_hosts

Then in the config.py code:

        if ignore_localhost:
            filter_functions.append(self._build_ignore_hosts(("localhost", "0.0.0.0", "127.0.0.1")))
        if ignore_hosts:
            filter_functions.append(self._build_ignore_hosts(ignore_hosts))

This should be backwards compatible, at the cost of splitting out the localhost filter function from the existing host filters.