Initial modifications to rebuilt only changed files based on mod date, performance updates

1 files changed, 179 insertions, 126 deletions

diff --git a/barf b/barf
index c53ed7b..c19ec94 100755
--- a/barf
+++ b/barf
@@ -1,158 +1,211 @@
 #!/bin/sh
 
+start_time=$(date +%s)
 domain="https://btxx.org"
-
-# Check the operating system
-os_name=$(uname -s)
-
-if [ "$os_name" = "OpenBSD" ]; then
-    alias sed=gsed
-    alias date=gdate
-    alias rsync=openrsync
-elif [ "$os_name" = "Darwin" ]; then
-    alias sed=gsed
-    alias date=gdate
-fi
+timestamp_file="timestamps.tsv"
+
+# Check the operating system and set aliases
+case $(uname -s) in
+    OpenBSD)
+        alias sed=gsed
+        alias date=gdate
+        alias rsync=openrsync
+        ;;
+    Darwin)
+        alias sed=gsed
+        alias date=gdate
+        ;;
+esac
 
 set -eu
 MARKDOWN=lowdown
-IFS='	'
-
-# Create tab separated file with filename, title, creation date, last update
-index_tsv() {
-	for f in "$1"/*.md
-	do
-		title=$(sed -n '/^# /{s/# //p; q}' "$f")
-		printf '%s\t%s\t%s\t%s\n' "$f" "${title:="No Title"}"
-	done
+IFS=$'\t'
+
+# Precompute values used in multiple places
+header_title=$(sed -n '/^# /{s/# //p; q}' index.md)
+git_user_name=$(git config user.name)
+first_commit_date=$(git log --pretty='format:%ai' . | cut -d ' ' -f1 | tail -1)
+now=$(date +%FT%TZ)
+now_rss=$(date -u +"%a, %d %b %Y %H:%M:%S %z")
+
+# Update the timestamp for a markdown file
+update_timestamp() {
+    local filename=$1
+    local timestamp=$(date -r "$filename" +%s)
+    # Remove old timestamp and add new
+    awk -v file="$filename" -v ts="$timestamp" '$1 != file {print $0}' "$timestamp_file" > "${timestamp_file}.tmp"
+    echo "$filename $timestamp" >> "${timestamp_file}.tmp"
+    mv "${timestamp_file}.tmp" "$timestamp_file"
 }
 
-index_html() {
-	# Print header
-	title=$(sed -n '/^# /{s/# //p; q}' index.md)
-	sed "s/{{TITLE}}/$title/" header.html
-
-	# Intro text
-	$MARKDOWN index.md
-
-	echo "<div>"
+# Check if the markdown file needs rebuilding
+needs_rebuilding() {
+    local filename=$1
+    local last_mod=$(date -r "$filename" +%s)
+    local last_build=$(awk -v file="$filename" '$1 == file {print $2}' "$timestamp_file")
+
+    if [ "$last_mod" -gt "${last_build:-0}" ]; then
+        return 0  # needs rebuilding
+    else
+        return 1  # does not need rebuilding
+    fi
+}
 
-	# Posts
-	while read -r f title created; do
-		link=$(echo "$f" | sed -E 's|.*/(.*).md|\1/|')
-		created=$(echo $(head -3 "$f" | tail -1))
-		echo "<span>$created &middot; <a href=\"$link\">$title</a></span><br>"
-	done < "$1" | sort -r
+# Define write_page function
+write_page() {
+    local filename=$1
+    local title=$2
+    local category=$3  # Category determines if it's a post or a page
+
+    if needs_rebuilding "$filename"; then
+        local directory
+        if [ "$category" = "posts" ]; then
+            directory="build/posts/$(basename "$filename" .md)"
+        else
+            directory="build/$(basename "$filename" .md)"
+        fi
+        mkdir -p "$directory"
+        local target="${directory}/index.html"
+        local created=$(head -3 "$filename" | tail -1)
+
+        {
+            sed "s|{{TITLE}}|$title|" header.html
+            $MARKDOWN "$filename"
+            cat footer.html
+        } > "$target"
+        update_timestamp "$filename"
+        
+        echo "Rebuilt: $filename"
+    fi
+}
 
-	echo "<div>"
+# Create tab separated file with filename, title
+index_tsv() {
+    for f in "$1"/*.md; do
+        title=$(sed -n '/^# /{s/# //p; q}' "$f")
+        printf '%s\t%s\n' "$f" "${title:="No Title"}"
+    done
+}
 
-	# Print footer after post list
-	cat footer.html
+# Generate index.html
+index_html() {
+    sed "s/{{TITLE}}/$header_title/" header.html
+    $MARKDOWN index.md
+    echo "<div>"
+
+    while read -r f title; do
+        link="${f%.md}/"
+        created=$(head -3 "$f" | tail -1)
+        echo "<span>$created &middot; <a href=\"$link\">$title</a></span><br>"
+    done < "$1" | sort -r
+
+    echo "</div>"
+    cat footer.html
 }
 
+# Generate atom.xml
 atom_xml() {
-	uri=$(sed -rn '/atom.xml/ s/.*href="([^"]*)".*/\1/ p' header.html)
-	first_commit_date=$(git log --pretty='format:%ai' . | cut -d ' ' -f1 | tail -1)
-
-	cat <<EOF
-<?xml version="1.0" encoding="utf-8"?>
-<feed xmlns="http://www.w3.org/2005/Atom">
-	<title>$(sed -n '/^# /{s/# //p; q}' index.md)</title>
-	<link href="$domain/atom.xml" rel="self" />
-	<updated>$(date +%FT%TZ)</updated>
-	<author>
-		<name>$(git config user.name)</name>
-	</author>
-	<id>$domain,$first_commit_date:default-atom-feed/</id>
-EOF
-
-	while read -r f title created; do
-
-		content=$($MARKDOWN "$f" | sed 's/&/\&amp;/g; s/</\&lt;/g; s/>/\&gt;/g; s/"/\&quot;/g; s/'"'"'/\&#39;/g')
-		post_link=$(echo "$f" | sed -E 's|posts/(.*).md|\1|')
-		basic_date=$(echo $(head -3 "$f" | tail -1))
-		published_date=$(date -d $basic_date -u +%Y-%m-%dT10:%M:%SZ)
-
-		cat <<EOF
-	<entry>
-		<title>$title</title>
-		<content type="html">$content</content>
-		<link href="$domain/$post_link"/>
-		<id>$domain/$post_link</id>
-		<updated>$published_date</updated>
-		<published>$published_date</published>
-	</entry>
-EOF
-	done < "$1"
-
-	echo '</feed>'
+    echo "<?xml version=\"1.0\" encoding=\"utf-8\"?>
+<feed xmlns=\"http://www.w3.org/2005/Atom\">
+    <title>$header_title</title>
+    <link href=\"$domain/atom.xml\" rel=\"self\" />
+    <updated>$now</updated>
+    <author>
+        <name>$git_user_name</name>
+    </author>
+    <id>$domain,$first_commit_date:default-atom-feed</id>"
+
+    while read -r f title; do
+        content=$($MARKDOWN "$f" | sed 's/&/\&amp;/g; s/</\&lt;/g; s/>/\&gt;/g; s/"/\&quot;/g; s/'"'"'/\&#39;/g')
+        post_link="${f%.md}"
+        basic_date=$(head -3 "$f" | tail -1)
+        published_date=$(date -d "$basic_date" -u +%Y-%m-%dT10:%M:%SZ)
+        echo "    <entry>
+        <title>$title</title>
+        <content type=\"html\">$content</content>
+        <link href=\"$domain/$post_link\"/>
+        <id>$domain/$post_link</id>
+        <updated>$published_date</updated>
+        <published>$published_date</published>
+    </entry>"
+    done < "$1"
+    echo '</feed>'
 }
 
+# Generate RSS
 rss_xml() {
-	uri=$(sed -rn '/rss.xml/ s/.*href="([^"]*)".*/\1/ p' header.html)
-	first_commit_date=$(git log --pretty='format:%ai' . | cut -d ' ' -f1 | tail -1)
+    uri=$(sed -rn '/rss.xml/ s/.*href="([^"]*)".*/\1/ p' header.html)
+    first_commit_date=$(git log --pretty='format:%ai' . | cut -d ' ' -f1 | tail -1)
 
-	cat <<EOF
+    cat <<EOF
 <?xml version="1.0" encoding="utf-8"?>
 <rss version="2.0">
-	<channel>
-		<title>$(sed -n '/^# /{s/# //p; q}' index.md)</title>
-		<link>$domain/rss.xml</link>
-		<description>Feed description here</description>
-		<lastBuildDate>$(date -u +"%a, %d %b %Y %H:%M:%S %z")</lastBuildDate>
-		<pubDate>$(date -u +"%a, %d %b %Y %H:%M:%S %z")</pubDate>
-		<generator>Custom RSS Generator</generator>
-		<ttl>1800</ttl>
+    <channel>
+        <title>$(sed -n '/^# /{s/# //p; q}' index.md)</title>
+        <link>$domain/rss.xml</link>
+        <description>Feed description here</description>
+        <lastBuildDate>$(date -u +"%a, %d %b %Y %H:%M:%S %z")</lastBuildDate>
+        <pubDate>$(date -u +"%a, %d %b %Y %H:%M:%S %z")</pubDate>
+        <generator>Custom RSS Generator</generator>
+        <ttl>1800</ttl>
 EOF
 
-	while read -r f title created; do
-		content=$($MARKDOWN "$f" | sed 's/&/\&amp;/g; s/</\&lt;/g; s/>/\&gt;/g; s/"/\&quot;/g; s/'"'"'/\&#39;/g')
-		post_link=$(echo "$f" | sed -E 's|posts/(.*).md|\1|')
-		basic_date=$(echo $(head -3 "$f" | tail -1))
-		published_date=$(date -d "$basic_date" -u +"%a, %d %b %Y %H:%M:%S %z")
-
-		cat <<EOF
-		<item>
-			<title>$title</title>
-			<description>$content</description>
-			<link>$domain/$post_link</link>
-			<guid isPermaLink="false">$domain/$post_link</guid>
-			<pubDate>$published_date</pubDate>
-		</item>
+    while read -r f title created; do
+        content=$($MARKDOWN "$f" | sed 's/&/\&amp;/g; s/</\&lt;/g; s/>/\&gt;/g; s/"/\&quot;/g; s/'"'"'/\&#39;/g')
+        post_link=$(echo "$f" | sed -E 's|posts/(.*).md|\1|')
+        basic_date=$(echo $(head -3 "$f" | tail -1))
+        published_date=$(date -d "$basic_date" -u +"%a, %d %b %Y %H:%M:%S %z")
+
+        cat <<EOF
+        <item>
+            <title>$title</title>
+            <description>$content</description>
+            <link>$domain/$post_link</link>
+            <guid isPermaLink="false">$domain/$post_link</guid>
+            <pubDate>$published_date</pubDate>
+        </item>
 EOF
-	done < "$1"
+    done < "$1"
 
-	echo '</channel>'
-	echo '</rss>'
+    echo '</channel>'
+    echo '</rss>'
 }
 
-write_page() {
-	filename=$1
-	directory=$(echo $(basename "$filename" .md))
-	$(mkdir -p build/$directory)
-	target=$(echo "$filename" | sed -r 's|\w+/(.*).md|build/\1/index.html|')
-	created=$(echo $(head -3 "$filename" | tail -1))
-	title=$2
-
-	$MARKDOWN "$filename" | \
-		cat header.html - |\
-		sed "s|{{TITLE}}|$title|" \
-		> "$target" && cat footer.html >> "$target"
+# Initialize timestamps for all Markdown files if not already done
+initialize_timestamps() {
+    if [ ! -f "$timestamp_file" ] || [ ! -s "$timestamp_file" ]; then
+        echo "Initializing timestamps..."
+        for f in index.md posts/*.md pages/*.md; do
+            update_timestamp "$f"
+        done
+        echo "Timestamps initialized for all Markdown files."
+    fi
 }
 
-rm -rf build && mkdir build
+# Main operations
+mkdir -p build/posts build/pages
+initialize_timestamps
 
-# Blog posts
-index_tsv posts | sort -rt "	" -k 3 > build/posts.tsv
-index_html build/posts.tsv > build/index.html
-atom_xml build/posts.tsv > build/atom.xml
-rss_xml build/posts.tsv > build/rss.xml
-while read -r f title created; do
-	write_page "$f" "$title" "$created"
+index_tsv posts > build/posts.tsv
+index_tsv pages > build/pages.tsv
+
+while read -r f title; do
+    write_page "$f" "$title" "posts"
 done < build/posts.tsv
 
-# Pages
-index_tsv pages > build/pages.tsv
-while read -r f title created; do
-	write_page "$f" "$title" "$created"
+while read -r f title; do
+    write_page "$f" "$title" "pages"
 done < build/pages.tsv
+
+if needs_rebuilding "index.md"; then
+    index_html build/posts.tsv > build/index.html
+    echo "Rebuilt index.html due to changes in index.md"
+    update_timestamp "index.md"
+fi
+
+atom_xml build/posts.tsv > build/atom.xml
+rss_xml build/posts.tsv > build/rss.xml
+
+end_time=$(date +%s)
+elapsed_time=$((end_time - start_time))
+echo "Total time: ${elapsed_time}s"