From 0e23e491998eb8ad888cc04218c6fd3ab040f05d Mon Sep 17 00:00:00 2001
From: Quentin Dufour <quentin@deuxfleurs.fr>
Date: Fri, 16 Jun 2023 10:50:37 +0200
Subject: [PATCH] rescue logic

---
 src/header.rs | 40 +++++++++++++++++++++++++++++++++++-----
 src/model.rs  |  3 +--
 src/parse.rs  |  3 +++
 3 files changed, 39 insertions(+), 7 deletions(-)

diff --git a/src/header.rs b/src/header.rs
index 960881d..f5a2ac2 100644
--- a/src/header.rs
+++ b/src/header.rs
@@ -1,10 +1,9 @@
 use nom::{
     IResult,
     branch::alt,
-    bytes::complete::take_while1,
-    bytes::complete::tag,
+    bytes::complete::{is_not, take_while1, take_while, tag},
     character::complete::space0,
-    combinator::{map, opt},
+    combinator::{map, opt, recognize},
     multi::{many0, many1, fold_many0, separated_list1},
     sequence::{terminated, preceded, pair, tuple},
 };
@@ -25,7 +24,7 @@ use crate::{datetime, trace, model};
 /// See: https://www.rfc-editor.org/rfc/rfc5322.html#section-2.2
 pub fn section(input: &str) -> IResult<&str, HeaderSection> {
     let (input, headers) = fold_many0(
-        alt((header_field, unknown_field)),
+        alt((header_field, unknown_field, rescue)),
         HeaderSection::default,
         |mut section, head| {
             match head {
@@ -109,6 +108,11 @@ pub fn section(input: &str) -> IResult<&str, HeaderSection> {
                 HeaderField::Optional(name, body) => {
                     section.optional.insert(name, body);
                 }
+
+                // Rescue
+                HeaderField::Rescue(x) => {
+                    section.unparsed.push(x);
+                }
             };
             section
         }
@@ -149,7 +153,10 @@ pub enum HeaderField<'a> {
     ReturnPath(Option<model::MailboxRef>),
 
     // 3.6.8.  Optional Fields
-    Optional(&'a str, String)
+    Optional(&'a str, String),
+
+    // None
+    Rescue(&'a str),
 }
 
 /// Parse one known header field
@@ -285,6 +292,21 @@ fn field_name(input: &str) -> IResult<&str, &str> {
     )(input)
 }
 
+/// Rescue rule
+///
+/// Something went wrong while parsing headers,
+/// trying to fix parsing by consuming
+/// one unfolded header line.
+///
+/// ```abnf
+/// rescue = *(*any FWS) *any CRLF
+fn rescue(input: &str) -> IResult<&str, HeaderField> {
+    map(recognize(pair(
+        many0(pair(is_not("\r\n"), fws)),
+        pair(is_not("\r\n"), perm_crlf,
+    ))), |x| HeaderField::Rescue(x))(input)
+}
+
 #[cfg(test)]
 mod tests {
     use super::*;
@@ -447,6 +469,14 @@ mod tests {
     fn test_invalid_field_name() {
         assert!(header_field("Unknown: unknown\r\n").is_err());
     }
+
+    #[test]
+    fn test_rescue() {
+        assert_eq!(
+            rescue("Héron: élan\r\n\tnoël: test\r\n"),
+            Ok(("", HeaderField::Rescue("Héron: élan\r\n\tnoël: test\r\n"))),
+        );
+    }
 }
 
 
diff --git a/src/model.rs b/src/model.rs
index a18c434..6bce998 100644
--- a/src/model.rs
+++ b/src/model.rs
@@ -101,6 +101,5 @@ pub struct HeaderSection<'a> {
 
     // 3.6.8.  Optional Fields
     pub optional: HashMap<&'a str, String>,
-    //pub failed: HashMap<&'a str, String>,
-    //pub garbage: &'a str,
+    pub unparsed: Vec<&'a str>,
 }
diff --git a/src/parse.rs b/src/parse.rs
index 72a6278..adc3880 100644
--- a/src/parse.rs
+++ b/src/parse.rs
@@ -20,6 +20,9 @@ Subject: Re: Saying Hello
 Comments: A simple message
 Comments: Not that complicated
 Keywords: hello, world
+Héron: Raté
+ Raté raté
+raté raté
 Keywords: salut, le, monde
 Message-ID: <3456@example.net>
 In-Reply-To: <1234@local.machine.example>