Config Framework Feedback

I bumped into a limitation that was a little frustrating to work around with the config framework.

It is inconvenient that values stored in files read by adding to Config::config_files are not available within the bro_init event. After reviewing how the Config framework works, I understand why this is the case, but it means that if you want to use configuration values on startup, you're not guaranteed to be working with the anticipated value.

I think the introduction of an event that ensures that configuration files have been read by the time that the event fires might be worthwhile. I was wondering if anyone had any thoughts on how to resolve/work-around this issue.




module Known;

redef Config::config_files += {"Known.dat"};

export {
    option KnownServers: set[addr] = set();

event bro_init() {
        print KnownServers;

event bro_done() {
        print KnownServers;

======= output =======

bro -r stream-1.pcap ./try-config.bro







This is an issue, but probably not the one you are thinking of. even if configuration files were fully read by the time bro_init was ran, you'd still need to handle the value changing at runtime.

If you need to run code when an option changes, you can use the Option::set_change_handler to make bro raise an event. Using this instead of bro_init would fix the startup problem, and handle it changing at any time.

There's a test that uses this in testing/btest/core/option-priorities.bro and a few other files.

I agree that more complex code on my side would definitely make this a non-issue.

In my case, I was setting up something in bro_init that would be used repeatedly throughout execution, and you're right--my code never handled option updates.

In this case, I opted to just use old redefable consts, as they will perform exactly as expected for any downstream users (changes require a bro restart).

Honestly, I only brought this up at all because I can definitely see a person new to scripting wondering why their options do not have the expected values at bro_init time and troubleshooting for hours.

Maybe it would be worthwhile to just add a note in the documentation for the config framework about it and leave it at that?

Hi folks,

I would agree that it takes a bit of experimentation to figure out exactly when a change handler fires and how to reliably initialize or update things based on an option's value.

Consider this:

   module Foo;

   export { option foo = F; }

   function foo_handler(ID: string, foo_new: bool): bool
           print fmt("New foo: %s", foo_new);

           # Update stuff here based on foo's value
           # ...

           return foo_new;

   event bro_init() {
           Option::set_change_handler("Foo::foo", foo_handler);

... foo_handler doesn't get called when you simply run the script without redefing Config::config_files. When you do redef it, the handler fires both when the config file sets foo to T, and when it sets it to F.

So you have to make sure that your initialization happens even when the handler doesn't get called, and you cannot write your handler assuming that the new value is actually different from the old one.

These arguably aren't bugs, but imo they do take getting used to.


The oberservations / thoughts in this thread seem worth a ticket I'd
say. We can refine this over time if the current semantics aren't
quite ideal yet.


Okay Robin, I've created for this.