kevinsuner / bloggo Goto Github PK

The markdown parser that Bloggo is using yuin/goldmark already supports this with its Built-in extensions so it should be pretty trivial to implement.

At this moment, the way the Bloggo theme is switching between dark and light mode is by using a custom script located at the header, and using the window.onload event to trigger a function that determines which theme and theme-icon need to be set. Prior to this, the script was located in the footer template and worked just fine, but since now we are waiting for the whole page to load (including resources such as stylesheets, scripts, iframes, and images), the content gets displayed using the previous theme, and then after some milliseconds switches to the correct one.

When creating the bloggo-demo repository, I noticed that the process of getting changes from upstream is not as ideal as I will wanted it to be, as I had to delete pretty much everything manually minus some specific files, and then re-generate the public folder.

To solve this problem I'm thinking of using Cobra to implement three separated commands:

• bloggo update Performs a git clone of Bloggo's master branch in a temporal folder, removes all folders minus user installed themes (bloggo theme comes by default, so it will get deleted too) and content.
• bloggo generate Generates the public directory and converts every Markdown file to HTML using go templates
• bloggo test Spins an HTTP server to serve the content contained in the public directory

Add a section in the README.md that showcases how to create a Github Action to serve the static HTML generated by Bloggo on Github Pages

At the moment the way the archive map is being created is by using the [year] part in the metadata.PublishedTime struct field as the key, and appending to it a post struct. This is happening inside a for-loop that iterates over a previously sorted []post array, so the inserts that are occurring on the archive map are being performed on a ordered way, but when printing out the map itself, the order appears to have been reversed and instead of it being [2018, 2017, 2016] is [2016, 2017, 2018]. This could be due to how maps are implemented in Go, or me not understanding them, so I'm thinking that it would be great to stop using a map and implement something along this lines Iterate over map.

Currently the way directories from the theme folder are being copied is by using a function called copyDir and passing it the directory that needs to be copied. Instead a better idea would be to let the user define which directories from the theme folder must be copied.

The way Bloggo is currently generating the index, about and archive pages under the hood is by having a specific receiver function per each one of them. Like this:

package main

import "fmt"

type Generator struct {
    BaseURL string
}

func (g *Generator) Index() {
    fmt.Println(g.BaseURL)
}

func (g *Generator) About() {
    fmt.Println(g.BaseURL)
}

func (g *Generator) Archive() {
    fmt.Println(g.BaseURL)
}

Although this solution works, from an extensibility and maintainability standpoint it wouldn't be wise to keep it as it is. In-order to tackle this problem the current generator must be replaced by a proper render package that accomplishes the aforementioned things.

The inner workings of the renderer are as straightforward as they can be, it profits from the user provided configuration and the metadata contained in the .md file to determine which options must be used when rendering.

flowchart LR
    1[config]-->2["render(...opts)"]
    3[markdown]-->4[frontmatter]-->2
    2-->5[html]

And it does not go against the KISS principle that Bloggo follows, because there will be a limit in the amount of options that users can provide in the configuration file. So, in the end index, about and archive will be the only pages that get rendered, but each one of them will be greatly customizable.

Replace the first part of the paragraph "Bloggo is a simplistic" by "Bloggo is a free and simplistic"