AppImage / AppImage/appimage.github.io
Replace Jekyll with a more efficient rendering system
- Lenguaje dominante
- Shell
- Estrellas
- 374
- Forks
- 769
- Merge medio
- 4 d 10 h
- PR fusionados (30 d)
- 8
Descripción
As the site currently uses Jekyll, it is quite limited in what data formats the data can be stored in so that it is readable to Jekyll. These "Markdown" pages containing nothing but these tables are a good example for this, they're hard to read for people who want to make use of the data in this repository. As a primary goal is machine readability, this is a no-go.
IMO, it is mandatory to switch to our own software for managing the data, and rendering the static pages. This allowed for extracting the actual AppImageHub as being the central data storage An approach I could imagine to work well is the following:
* Write a library that implements the data format (e.g. on top of JSON, but I could as well imagine something more abstract that can then be converted to JSON, XML, AppStream XML etc., whatever works best for third party software). This library would contain a full specification of these formats, which allows for writing AppImageHub client libraries that could then be used by other applications.
* Write a library for rendering the AppImageHub data to something I'd call the "reference client". This is what would be deployed on `https://hub.appimage.org` as a web frontend, while AppImageHub could be deployed on `https://hub.appimage.org/api`.
The latter library could really improve the workflow for adding software to AppImageHub. I could imagine to implement a little interactive web server based on Flask, which renders the pages on demand, allowing a "change, then press F5 in the browser" workflow.
But I assume that the biggest advantage would be that the time for rendering the whole AppImageHub pages could be greatly reduced. We don't require most of the features it provides, but which make it somewhat slow. A simple loop calling a templating engine with the same template should suffice at the moment. And, we're not limited any more to the data format that is fed into this renderer.
Guía de contribución
Evaluación
Este issue todavía no se ha evaluado.