feat: add Quarkus CLI commands

[metacosm]

• chore: initial work on CLI plugin
• wip: first commands layout
• feat: first implementation of API creation command

[csviri]

LGTM

[mkouba]

Like I mentioned, this is not ideal in the sense that the template content is parsed every time the Qute#fmt() method is called. If you need to render the template multiple times then either use Qute.enableCache() or Qute.engine().parse(...) together with Qute.engine().putTemplate() and Qute.engine().getTemplate()...

[metacosm]

I couldn't get it to work in that context, though… where all the templates are retrieved from the jar and processed the same way. If you have a better way to do, I'll gladly consider it :)

[mkouba]

What I'm saying is: if you only render the template once, then Qute.fmt(templateAsString) is OK. If multiple times, then either use Qute.enableCache() - this enables cache for Qute.fmt() by default (templateAsString will be parsed once); or use something like:

// This should be called once
Engine engine = Qute.engine();
engine.putTemplate("myTemplate", engine.parse(templateAsString));
...
// And then to render a template:
Qute.engine().getTemplate("myTemplate").render();

[metacosm]

Ah, I think I was missing the putTemplate part, I thought that I could call getTemplate directly and that wasn't working. Thanks! That said, since these things won't probably be run very often, I'm not sure it matters but I might still make the change (it might help when called from the dev console, maybe?). Then again, you could make a point that if it's not going to be called often, caching a template might actually be worse because then you commit memory to something that will only be used infrequently and possibly only once.

[mkouba]

Yes, it really depends on the use case 🤷 .