From 797deb6457695d79c57e96930dd37f2672e7a94f Mon Sep 17 00:00:00 2001
From: Nicolas Delsaux <nicolas.delsaux@zenika.com>
Date: Sun, 8 Sep 2019 17:46:16 +0200
Subject: [PATCH] Fixes some troubles regarding rfc2822 date parsing

---
 Cargo.toml  |  1 +
 src/feed.rs | 31 ++++++++++++++++++-------------
 src/main.rs |  1 +
 3 files changed, 20 insertions(+), 13 deletions(-)

diff --git a/Cargo.toml b/Cargo.toml
index 5c9ca33..88366d3 100644
--- a/Cargo.toml
+++ b/Cargo.toml
@@ -34,6 +34,7 @@ treexml = "0.7"
 atom_syndication = "0.6"
 rss = "1.8"
 chrono = { version = "0.4", features = ["serde"] }
+rfc822_sanitizer = "0.3.2"
 # A good html parser to allow generation of valid mail messages
 # kuchiki is in fact a layer over html5ever, which adds easier tree traversal/transformation
 kuchiki = "0.7"
diff --git a/src/feed.rs b/src/feed.rs
index 49d075c..13bf5ef 100644
--- a/src/feed.rs
+++ b/src/feed.rs
@@ -179,12 +179,7 @@ fn extract_authors_from_rss(entry: &RssItem, feed: &RssChannel) -> Vec<String> {
         Some(l) => message_authors = vec![l.to_owned()],
         _ => message_authors = vec![feed.title().to_owned()],
     }
-    message_authors = message_authors
-        .iter()
-        .map(|author| (author, author.replace(" ", "_")))
-        .map(|tuple| format!("{} <{}@{}>", tuple.0, tuple.1, domain))
-        .collect();
-    message_authors
+    sanitize_message_authors(message_authors, domain)
 }
 fn find_rss_domain(feed: &RssChannel) -> String {
     return Some(feed.link())
@@ -230,12 +225,11 @@ fn extract_from_rss(entry: &RssItem, feed: &RssChannel) -> Message {
 
 fn extract_date_from_rss(entry: &RssItem) -> NaiveDateTime {
     if entry.pub_date().is_some() {
-        let mut pub_date = str::replace(entry.pub_date().unwrap(), "-0000", "+0000");
-        pub_date = str::replace(&pub_date, "+00:00", "+0000");
-        return DateTime::parse_from_rfc2822(&pub_date)
+        let pub_date = entry.pub_date().unwrap().to_owned();
+        return rfc822_sanitizer::parse_from_rfc2822_with_fallback(&pub_date)
             .unwrap_or_else(|e| {
                 panic!(
-                    "pub_date for item {:?} (value is {:?}) can't be parsed. {:?}",
+                    "pub_date for item {:?} (value is {:?}) can't be parsed as rfc2822. {:?}",
                     &entry, pub_date, e
                 )
             })
@@ -272,12 +266,23 @@ fn extract_authors_from_atom(entry: &AtomEntry, feed: &AtomFeed) -> Vec<String>
     if message_authors.is_empty() {
         message_authors = vec![feed.title().to_owned()]
     }
-    message_authors = message_authors
+    sanitize_message_authors(message_authors, domain)
+}
+
+fn sanitize_message_authors(message_authors:Vec<String>, domain:String)->Vec<String> {
+    let fixed = message_authors
         .iter()
-        .map(|author| (author, author.replace(" ", "_")))
+        // ni next line, we create a tuple to be used to generate the email address
+        .map(|author| (author, // first element of tuple is email displayed name
+            author.to_lowercase() // second element of tuple is generated user address
+                .replace(" ", "_")
+                .replace("&", "and")
+                .replace(",;:!", "")
+                .replace("ï", "i")
+            ))
         .map(|tuple| format!("{} <{}@{}>", tuple.0, tuple.1, domain))
         .collect();
-    message_authors
+    return fixed;
 }
 
 fn find_atom_domain(feed: &AtomFeed) -> String {
diff --git a/src/main.rs b/src/main.rs
index 35f478e..22bd630 100644
--- a/src/main.rs
+++ b/src/main.rs
@@ -123,6 +123,7 @@ extern crate flexi_logger;
 extern crate treexml;
 
 extern crate chrono;
+extern crate rfc822_sanitizer;
 
 #[macro_use]
 extern crate tera;