From 264f8a4a94fb366f90166e66300a82f513aaf515 Mon Sep 17 00:00:00 2001
From: Jaidyn Ann <10477760+JadedCtrl@users.noreply.github.com>
Date: Fri, 17 Nov 2023 16:04:08 -0600
Subject: [PATCH] Add a custom sfeedrc for getting only new posts

---
 README.md       | 25 +++++++++------------
 sfeedrc.example | 60 +++++++++++++++++++++++++++++++++++++++++++++++++
 2 files changed, 71 insertions(+), 14 deletions(-)
 create mode 100644 sfeedrc.example

diff --git a/README.md b/README.md
index 35e5a2f..8d1149b 100644
--- a/README.md
+++ b/README.md
@@ -20,20 +20,15 @@ You’ve done it!
 
 ## Configuration
 ### sfeed
-We need to create a config file and feed directory for sfeed — they can be anywhere, your choice.
+We need to create a config file and feed directory for sfeed_update.
+You can use the sfeedrc.example file in this repo as a base for your own config file.
 ```
 $ mkdir ~/.config/sfeed/
-$ cat > ~/.config/sfeed/config <<EOF
-sfeedpath="$HOME/.config/sfeed/"
-
-feeds() {
-	feed "Planet GNU" "https://planet.gnu.org/rss20.xml" "https://planet.gnu.org" "UTF-8"
-	feed "Tiriftjo" "https://tirifto.xwx.moe/en/news.atom" "https://tirifto.xwx.moe" "UTF-8"
-}
-EOF
+$ cp sfeedrc.example ~/.config/sfeedrc
 ```
 
-You can read up more on sfeed’s configuration in its documentation¸ sfeedrc(5).
+You need to edit the example sfeedrc to add in your own Atom/RSS feeds, or to change the feed path.
+You can read up more on sfeed’s configuration in its man-page¸ sfeedrc(5).
 
 
 ### Mastodon
@@ -55,12 +50,14 @@ Whenever you use sfeed_mastodon, make sure that this token is stored in the envi
 ## Usage
 ```
 $ FEDI_AUTH="yourAuthorizationTokenHere"
-$ sfeed_update ~/.config/sfeed/config | sfeed_mastodon https://yourServer.here
+$ sfeed_update ~/.config/sfeedrc
+$ cat ~/.config/sfeed/* | sfeed_mastodon https://yourServer.here
 ```
 
-It’s that simple. It’s safe to run this command several times in a row  — feed entries that have
-already been posted won’t be reposted. You can even add this to your crontab to mirror an Atom/RSS
-feed automatically.
+It’s that simple. It’s safe to run these commands several times in a row  — feed entries that have
+already been posted won’t be reposted, if you use our example sfeedrc.
+
+To automatically mirror an Atom/RSS feed, you can put these commands into a script and put it in your crontab.
 
 
 ### Templates
diff --git a/sfeedrc.example b/sfeedrc.example
new file mode 100644
index 0000000..a20fab5
--- /dev/null
+++ b/sfeedrc.example
@@ -0,0 +1,60 @@
+# This is an sfeedrc(5) configuration file for sfeed_update(1).
+
+# The key difference is that it truncates your feed files at every update,
+# keeping only new posts. This saves you the work of filtering out old posts.
+
+
+# You probably want to EDIT this.
+sfeedpath="$HOME/.config/sfeed_mastodon/"
+
+
+# You probably want to EDIT this.
+# This contains a list of all your feeds, in the format:
+#     feed NAME URL DOMAIN ENCODING
+feeds() {
+	feed "Planet GNU" "https://planet.gnu.org/rss20.xml"     "https://planet.gnu.org"  "UTF-8"
+	feed "Tiriftjo"   "https://tirifto.xwx.moe/en/news.atom" "https://tirifto.xwx.moe" "UTF-8"
+}
+
+
+# This overrides sfeed_update’s default merge() function.
+# This makes it so that only new and unseen posts are put in the feed file.
+# This is done by storing the date of the latest post in an extended attribute,
+# for comparison during the next update..
+merge() {
+	local oldfile="$2"
+	local newfile="$3"
+
+	local previous_max_date="$(attr -q -g sfeed_latest "$oldfile" 2> /dev/null)"
+	if test -z "$previous_max_date"; then
+		previous_max_date=0
+	fi
+
+	# Update the date of the last-processed post.
+	local latest_date="$(latest_date "$newfile")"
+	attr -qs sfeed_latest -V "$latest_date" "$oldfile" 2> /dev/null
+
+	# Output only new and unprocessed posts.
+	after_date "$newfile" "$previous_max_date"
+}
+
+
+# Given an sfeed file, this returns the date of the latest post (in seconds
+# since the UNIX epoch).
+latest_date() {
+	local file="$1"
+	awk -F '\t' \
+		'$1 > latest { latest = $1 } END { print latest }' \
+		"$file"
+}
+
+
+# This outputs only lines of an sfeed file with a date after the given min_date
+# (in seconds since UNIX epoch).
+after_date() {
+	local file="$1"
+	local min_date="$2"
+	awk -F '\t' -v min_date="$min_date" \
+		'$1 > min_date { print $0 }' \
+		"$file"
+}