From 6aa64d77e0f8ef2e42bdb90bc5ed3f6557e3ed81 Mon Sep 17 00:00:00 2001
From: Case Duckworth
Date: Sun, 17 Jul 2022 23:10:22 -0500
Subject: Initial commit

---
 scripts/get-feed.sh | 59 +++++++++++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 59 insertions(+)
 create mode 100755 scripts/get-feed.sh

(limited to 'scripts')

diff --git a/scripts/get-feed.sh b/scripts/get-feed.sh
new file mode 100755
index 0000000..f000fee
--- /dev/null
+++ b/scripts/get-feed.sh
@@ -0,0 +1,59 @@
+#!/bin/sh
+
+main() {
+	# get-feed.sh URL(str) => DIRECTIVE(feed_directive)
+	url="$1"
+	wp="$(mktemp /tmp/get-feed.XXXXXX)"
+	curl -sL "$url" >"$wp"
+	case "$url" in
+	*html) # We know it's a webpage
+		type=html
+		;;
+	*xml) # We know it's a feed
+		type=xml
+		;;
+	*) # Not sure
+		type="$(head -n1 "$wp")"
+		;;
+	esac
+	case "$type" in
+	*xml*) # a feed
+		title="$(get_title_xml <"$wp")"
+		output_feed "$title" "$url"
+		;;
+	*html*) # a webpage
+		cat "$wp" | sfeed_web | cut -f1 |
+			while read u; do
+				title="$(curl -sL "$u" | get_title_xml)"
+				output_feed "$title" "$u"
+			done
+		;;
+	*)
+		echo >&2 "Don't know type \"$type\"."
+		exit 1
+		;;
+	esac
+}
+
+output_feed() {
+	## output_feed TITLE(str) URL(str) => FEED_DIRECTIVE(str)
+	printf "feed \"%s\" '%s'\n" "$1" "$2"
+}
+
+get_title_xml() {
+	## get_title_xml < FILE => TITLE(str)
+	awk '
+/<channel>/ { channel = 1; }/<item>/{ channel = 0; }
+channel && $0 ~ /<title>/ { title = 1; }
+title {
+      if (match($0,/<\/title>/)) title = 0;
+      gsub(/<\/?title>/,"");
+      sub(/^[ \t]*/,"");
+      sub(/[ \t]*$/,"");
+      print;
+}
+channel && $0 ~ /<\/title>/ { title = 0; }
+'
+}
+
+main "$@"
-- 
cgit 1.4.1-21-gabe81