pfgrep bundling with Code for IBM i

[NattyNarwhal]

Prompted by the recent office hours and Zulip thread; obsoletes the #2429 effort.

We think the best path forward would be bundling a pfgrep binary with Code for i. The advantage is we wouldn't need to have it available in Yum or have the user install it from there, it would "just work" when Code for i engages itself with the system.

Concerns:

• Native code: Not sure what the story with bundling native code with Code for i is in general, in terms of how we do it, OS targets, etc.
• Licensing: pfgrep is GPL, but I don't think this will pose a problem. The virality part ends at the pfgrep binary, so it wouldn't spread the GPL into Code for i, it's just that changes made for Code for i would need to be shared to meet the GPL requirements (which I assume we'd do anyways). If it really does pose a problem, we can look into perhaps figuring out a relicense (as the copyright holder).
• Dependencies: pfgrep needs json-c (for data structures) and pcre2 (for regexp) at minimum, as well as libzip for building archives in pfzip. Would want to figure out if we'd be bundling these or just using the provided copies from Yum.

As the pfgrep developer, let me know what I can help with here.

[NattyNarwhal]

After more discussion on Zulip,it seemed better to just use the component system with an RPM-installed pfgrep.

To make it easier by being packaged in the IBM repo, I'm putting the specfile here.