Development discussions (i.e., software architecture and project management) about Antora, the static site generator for making documentation sites based on AsciiDoc.
I'll give lodash a try too. Thanks for the suggestion!
we use lodash in Antora (though I try not to use dependencies if possible)
the one minor exception to that is how the canonical URL is handled for pages that are missing. that probably should have gone into 2.1.0, but hindsight
2.1.0 will have some improvements, but no pipeline changes. those will either happen in a later minor or in the next major
I'm done with my research, currently creating typescript defs for the site-generator process and everything it knows about. I'm going to do them within our codebase first- then I'll ping you to review their content for accuracy, and work out where we want to put them etc before I create a PR.
I think the timing is good, because we really need to have a discussion about the API surface area in Antora
I think we were tentative at first because we weren't sure exactly how it would be used. but it's pretty clear now that the API will become one of Antora's best assets
they are in each package
they may need some updating, but that was what we knew to be true at 1.0.0
@sophistifunk I think it was Node JS from the very start by design. I think it was a good choice FWIW as node is really gaining a following in server side and having server and client side languages the same is helpful. I am curious as to specific reasons @mojavelinux had for choosing it though
there are two key reasons for it being written in Node.js
- Node.js fits best with the web-based ecosystem, which is the primary output target
- Node.js is much easier to get buy-in / adoption in the enterprise than Ruby (and I say that as a person who wishes that were true for Ruby)
the resistance to Ruby has been incredibly strong (and again, I say that as a person who absolutely loves the language). Node.js is shockingly easy to get companies to adopt (for better or for worse)
but (1) is really the key reason
the alternative choice was Java, but a) I wanted more runtime flexibility (monkey patching) and b) i feared it would pin us too closely to that ecosystem (we would be a "java" tool)
and yet, Node.js still allows us to run in the JVM (one day, some how, we'll get there)
though I hope that Antora is an architecture that can be implemented again and again
I should say, thirdly, we want to try something different. there's always that "let's just go for it" aspect to a decision ;)
;)
No rush to respond, because I'm knocking off for the day, but here's my typedefs: https://gist.github.com/sophistifunk/7fd33bfeca88cd297d1cdbc23dd0e24d
it should just be "VirtualFile" or "File"
pub is how it is referred to on the public internet
that's an important distinction (which honestly most site generators get wrong)
I need to refine my defs for VirtualFile, then I'm going to put them into fork of Antora with a few code simplifications so I can run TSC as a JS-checker against them, in order to make sure they actually match what the code is doing. Then I'll move just the definitions and not the code changes into a new fork, so I can send in a PR
Hi, I just tested Netlify's lambda feature and I'm very impressed how easy is to link static content with serverless code. First idea comes to mind is to have a plugin who encrypt/hide some blocks of documentation and after identify the user decrypt/show it executing some aws function
What do you think ? can be interesting ? are there some similar plugin ?
What do you think ? can be interesting ? are there some similar plugin ?
Antora is just the HTML-generator. It doesn't provide any logic after that. You can create your own UI that incorporates additional functionality, and you can use various Javascript frameworks to do so. It's just that Antora is not involved in the operational portion, where a user is browsing your documentation. If you can make your authentication/display operations work with a hand-crafted web page, then you incorporate it into your Antora UI.