Cloud 5 - REPOINIT

James, have you ever heard the term repo init used outside of the Sling AEM world? You mean like a sweater nit? Sure. I’ve heard that. No, no, no, no, no. What about repository initialization? Repository initialization. Hmm. Might be something. I can’t really think of anything offhand though. What about Sling? Oh, yeah. That’s a very common word. In fact, I just cut my arm out of a sling if you wanted to. Yeah. No, no, no, no. I mean, okay. What’s your favorite TV show? What is something you wouldn’t miss an episode of? Oh, man. That’s easy. No hesitation at all. Peppa Pig. What does this have to do with repo init? Nothing. I mean, I just wanted to get it on record that your favorite show is Peppa Pig. RepoNet, or the Sling repository initializer, has been available out of the box in AEM since several versions of AEM 6.5. Alongside the JCR API, RepoNet has become a very powerful tool for setting up and configuring repositories in our mutable, immutable world that we now live in. It’s a pretty handy tool to have in your kit for creating service users and their associated ACLs as well as the required content tree structures that your app requires. There’s a number of ways to use RepoNet, but I’d like to focus on the way that the latest AEM archetype handles configuration as it really sets that best practices precedent. Many of the AEM sum components like sites, assets, and so on already use RepoNet internally to initialize their structures and system requirements within the AEM system. Yeah, Darren, this sounds particularly useful for AEM as a cloud service. Talk to me a little bit about some use cases. Why would I actually want to use this in my AEM as cloud service project? Well, not every project needs something like this. For something like creating site paths and other things, you can still use the old school way of hacking together a BLT content XML files that ultimately turn into installable packages. But if you want to create users and assign permissions to those users, RepoNet is perfect for that scenario since it executes early in the build image step as an elevated system user. Yeah, that’s really fantastic. I can really think of some great use cases for this. How would I get started using RepoNet? So the Adobe suggested format for your RepoNet configs are the osgi.cfg.json format. The latest architect will build one of those in that format. However, I’m not a big fan of this format for longer and larger scripts. The JSON format won’t let you put multiple lines in the script, so you’ll end up with some really long lines with escape slash ends for new lines and other things. Here’s an example of that. I personally prefer the .config format, which will make the scripts much more readable since it respects those new lines and formatting options. There are a few things that need to be escaped with .config files, and we’re including those down below. Here’s an example of that same .config.json file reformatted to the .config file format. The grammar for those script languages, or the RepoNet script language, is largely self-explaining, and we’ve included the Sling appendix down below for reference. One thing you can do, James, is you can generate a new project using the latest AEM archetype and look at the config located in the UI config module under the OSGI directory. That should really get you started. Yeah, really appreciate you sharing the best practices here, Darren. Human readable is always the best. Do you have any other important considerations you want to tell me in the audience about using RepoNet? I have three. First is the scripts execute every time the system starts up. Every time. So you don’t have to create a bunch of items that are transient or temporary. The good news is that if an item has already been created, the system will just no opt and move on and not do anything. Number two, the same rules apply about mutable and immutable pads, so don’t make RepoNet change application level nodes as it’ll break the product or make your cloud manager pipeline fail and that’ll result in all sorts of trouble. Number three, RepoNet is very granular and prescriptive. You can’t have it query statements that change a large group of things. All the steps you want to happen have to be listed in order. All of them, one by one. Okay, that makes sense, Darren. This is really great and really useful. It’s not quite as great as an episode of Peppa Pig, but you know what? It comes close. Thanks again. RepoNet.