Add front_matter crate

We might want to move the blog to Zola. In that case, it will still be
useful to have a crate that validates and even fixes our front matter.

This will be especially important if we decide to duplicate the date of
a post in the two front matter fields `date` and `path`, which may be
necessary to preserve the existing URL scheme.

While the custom static site generator is still present, it can reuse
the front matter parsing logic.

Author: senekor

Cargo.lock

@@ -1,5 +1,5 @@
 [workspace]
-members = ["serve"]
+members = ["front_matter", "serve"]
 
 [workspace.package]
 edition = "2024"
@@ -10,6 +10,7 @@ chrono = "=0.4.40"
 color-eyre = "=0.6.3"
 comrak = "=0.36.0"
 eyre = "=0.6.12"
+front_matter = { path = "front_matter" }
 insta = "=1.42.2"
 rayon = "=1.10.0"
 regex = "=1.11.1"
@@ -32,6 +33,7 @@ chrono.workspace = true
 color-eyre.workspace = true
 comrak = { workspace = true, features = ["bon"] }
 eyre.workspace = true
+front_matter.workspace = true
 rayon.workspace = true
 regex.workspace = true
 sass-rs.workspace = true

Cargo.toml

@@ -0,0 +1,9 @@
+[package]
+name = "front_matter"
+version = "0.1.0"
+edition.workspace = true
+
+[dependencies]
+eyre.workspace = true
+serde = { workspace = true, features = ["derive"] }
+toml.workspace = true

front_matter/Cargo.toml

@@ -0,0 +1,29 @@
+use eyre::bail;
+use serde::{Deserialize, Serialize};
+
+/// The front matter of a markdown blog post.
+#[derive(Debug, PartialEq, Serialize, Deserialize)]
+pub struct FrontMatter {
+    pub title: String,
+    pub author: String,
+    #[serde(default)]
+    pub release: bool,
+    pub team: Option<String>,
+    pub layout: String,
+}
+
+/// Extracts the front matter from a markdown file.
+///
+/// The remaining normal markdown content is returned as the second element of
+/// the tuple.
+pub fn parse(markdown: &str) -> eyre::Result<(FrontMatter, &str)> {
+    if !markdown.starts_with("+++\n") {
+        bail!("markdown file must start with the line `+++`");
+    }
+    let (front_matter, content) = markdown
+        .trim_start_matches("+++\n")
+        .split_once("\n+++\n")
+        .expect("couldn't find the end of the front matter: `+++`");
+
+    Ok((toml::from_str(front_matter)?, content))
+}

front_matter/src/lib.rs

@@ -1,22 +1,12 @@
 use super::blogs::Manifest;
-use eyre::eyre;
+use front_matter::FrontMatter;
 use regex::Regex;
-use serde::{Deserialize, Serialize};
+use serde::Serialize;
 use std::{
     path::{Path, PathBuf},
     sync::LazyLock,
 };
 
-#[derive(Debug, PartialEq, Deserialize)]
-struct TomlHeader {
-    title: String,
-    author: String,
-    #[serde(default)]
-    release: bool,
-    team: Option<String>,
-    layout: String,
-}
-
 #[derive(Debug, Clone, Serialize)]
 pub struct Post {
     pub(crate) filename: String,
@@ -52,23 +42,17 @@ impl Post {
         let filename = split.next().unwrap().to_string();
 
         let contents = std::fs::read_to_string(path)?;
-        if contents.len() < 5 {
-            return Err(eyre!(
-                "{path:?} is empty, or too short to have valid front matter"
-            ));
-        }
 
-        // toml headers.... we know the first four bytes of each file are "+++\n"
-        // so we need to find the end. we need the fours to adjust for those first bytes
-        let end_of_toml = contents[4..].find("+++").unwrap() + 4;
-        let toml = &contents[4..end_of_toml];
-        let TomlHeader {
-            author,
-            title,
-            release,
-            team: team_string,
-            layout,
-        } = toml::from_str(toml)?;
+        let (
+            FrontMatter {
+                author,
+                title,
+                release,
+                team: team_string,
+                layout,
+            },
+            contents,
+        ) = front_matter::parse(&contents)?;
 
         let options = comrak::Options {
             render: comrak::RenderOptions::builder().unsafe_(true).build(),
@@ -81,8 +65,7 @@ impl Post {
             ..comrak::Options::default()
         };
 
-        // Content starts after "+++\n" (we don't assume an extra newline)
-        let contents = comrak::markdown_to_html(&contents[end_of_toml + 4..], &options);
+        let contents = comrak::markdown_to_html(contents, &options);
 
         // finally, the url.
         let mut url = PathBuf::from(&*filename);